diff --git a/example.bib b/example.bib deleted file mode 100644 index 4230c37..0000000 --- a/example.bib +++ /dev/null @@ -1,1602 +0,0 @@ -@article{ZHANG2021100025, -title = {AI technologies for education: Recent research and future directions}, -journal = {Computers and Education: Artificial Intelligence}, -volume = {2}, -pages = {100025}, -language = {English}, -year = {2021}, -issn = {2666-920X}, -type = {article}, -doi = {https://doi.org/10.1016/j.caeai.2021.100025}, -url = {https://www.sciencedirect.com/science/article/pii/S2666920X21000199}, -author = {Ke Zhang. and Ayse Begum Aslan}, -address={USA}, -affiliation={Wayne State University; Eastern Michigan University}, -keywords = {Artificial intelligence, AI, AI in Education}, -abstract = {From unique educational perspectives, this article reports a comprehensive review of selected empirical studies on artificial intelligence in education (AIEd) published in 1993–2020, as collected in the Web of Sciences database and selected AIEd-specialized journals. A total of 40 empirical studies met all selection criteria, and were fully reviewed using multiple methods, including selected bibliometrics, content analysis and categorical meta-trends analysis. This article reports the current state of AIEd research, highlights selected AIEd technologies and applications, reviews their proven and potential benefits for education, bridges the gaps between AI technological innovations and their educational applications, and generates practical examples and inspirations for both technological experts that create AIEd technologies and educators who spearhead AI innovations in education. It also provides rich discussions on practical implications and future research directions from multiple perspectives. The advancement of AIEd calls for critical initiatives to address AI ethics and privacy concerns, and requires interdisciplinary and transdisciplinary collaborations in large-scaled, longitudinal research and development efforts.} -} - -@article{PETROVIC201617, -title = {Knowledge-light adaptation approaches in case-based reasoning for radiotherapy treatment planning}, -journal = {Artificial Intelligence in Medicine}, -volume = {68}, -pages = {17-28}, -year = {2016}, -language = {English}, -issn = {0933-3657}, -type = {article}, -doi = {https://doi.org/10.1016/j.artmed.2016.01.006}, -url = {https://www.sciencedirect.com/science/article/pii/S093336571630015X}, -author = {Sanja Petrovic and Gulmira Khussainova and Rupa Jagannathan}, -affiliation={Nottingham University}, -address={UK}, -keywords = {Case-based reasoning, Adaptation-guided retrieval, Machine-learning tools, Radiotherapy treatment planning}, -abstract = {Objective -Radiotherapy treatment planning aims at delivering a sufficient radiation dose to cancerous tumour cells while sparing healthy organs in the tumour-surrounding area. It is a time-consuming trial-and-error process that requires the expertise of a group of medical experts including oncologists and medical physicists and can take from 2 to 3h to a few days. Our objective is to improve the performance of our previously built case-based reasoning (CBR) system for brain tumour radiotherapy treatment planning. In this system, a treatment plan for a new patient is retrieved from a case base containing patient cases treated in the past and their treatment plans. However, this system does not perform any adaptation, which is needed to account for any difference between the new and retrieved cases. Generally, the adaptation phase is considered to be intrinsically knowledge-intensive and domain-dependent. Therefore, an adaptation often requires a large amount of domain-specific knowledge, which can be difficult to acquire and often is not readily available. In this study, we investigate approaches to adaptation that do not require much domain knowledge, referred to as knowledge-light adaptation. -Methodology -We developed two adaptation approaches: adaptation based on machine-learning tools and adaptation-guided retrieval. They were used to adapt the beam number and beam angles suggested in the retrieved case. Two machine-learning tools, neural networks and naive Bayes classifier, were used in the adaptation to learn how the difference in attribute values between the retrieved and new cases affects the output of these two cases. The adaptation-guided retrieval takes into consideration not only the similarity between the new and retrieved cases, but also how to adapt the retrieved case. -Results -The research was carried out in collaboration with medical physicists at the Nottingham University Hospitals NHS Trust, City Hospital Campus, UK. All experiments were performed using real-world brain cancer patient cases treated with three-dimensional (3D)-conformal radiotherapy. Neural networks-based adaptation improved the success rate of the CBR system with no adaptation by 12%. However, naive Bayes classifier did not improve the current retrieval results as it did not consider the interplay among attributes. The adaptation-guided retrieval of the case for beam number improved the success rate of the CBR system by 29%. However, it did not demonstrate good performance for the beam angle adaptation. Its success rate was 29% versus 39% when no adaptation was performed. -Conclusions -The obtained empirical results demonstrate that the proposed adaptation methods improve the performance of the existing CBR system in recommending the number of beams to use. However, we also conclude that to be effective, the proposed adaptation of beam angles requires a large number of relevant cases in the case base.} -} - -@article{ROLDANREYES20151, -title = {Improvement of online adaptation knowledge acquisition and reuse in case-based reasoning: Application to process engineering design}, -journal = {Engineering Applications of Artificial Intelligence}, -volume = {41}, -pages = {1-16}, -affiliation={Université de Toulouse; Instituto Tecnologico de Orizaba}, -country={France}, -language = {English}, -year = {2015}, -type = {article}, -issn = {0952-1976}, -doi = {https://doi.org/10.1016/j.engappai.2015.01.015}, -url = {https://www.sciencedirect.com/science/article/pii/S0952197615000263}, -author = {E. {Roldan Reyes} and S. Negny and G. {Cortes Robles} and J.M. {Le Lann}}, -keywords = {Case based reasoning, Constraint satisfaction problems, Interactive adaptation method, Online knowledge acquisition, Failure diagnosis and repair}, -abstract = {Despite various publications in the area during the last few years, the adaptation step is still a crucial phase for a relevant and reasonable Case Based Reasoning system. Furthermore, the online acquisition of the new adaptation knowledge is of particular interest as it enables the progressive improvement of the system while reducing the knowledge engineering effort without constraints for the expert. Therefore this paper presents a new interactive method for adaptation knowledge elicitation, acquisition and reuse, thanks to a modification of the traditional CBR cycle. Moreover to improve adaptation knowledge reuse, a test procedure is also implemented to help the user in the adaptation step and its diagnosis during adaptation failure. A study on the quality and usefulness of the new knowledge acquired is also driven. As our Knowledge Based Systems (KBS) is more focused on preliminary design, and more particularly in the field of process engineering, we need to unify in the same method two types of knowledge: contextual and general. To realize this, this article proposes the integration of the Constraint Satisfaction Problem (based on general knowledge) approach into the Case Based Reasoning (based on contextual knowledge) process to improve the case representation and the adaptation of past experiences. To highlight its capability, the proposed approach is illustrated through a case study dedicated to the design of an industrial mixing device.} -} - -@article{JUNG20095695, -title = {Integrating radial basis function networks with case-based reasoning for product design}, -journal = {Expert Systems with Applications}, -volume = {36}, -number = {3, Part 1}, -language = {English}, -pages = {5695-5701}, -year = {2009}, -type = {article}, -issn = {0957-4174}, -doi = {https://doi.org/10.1016/j.eswa.2008.06.099}, -url = {https://www.sciencedirect.com/science/article/pii/S0957417408003667}, -author = {Sabum Jung and Taesoo Lim and Dongsoo Kim}, -affiliation={LG Production Engineering Research Institute; Sungkyul University; Soongsil University}, -keywords = {Case-based reasoning (CBR), Radial basis function network (RBFN), Design expert system, Product design}, -abstract = {This paper presents a case-based design expert system that automatically determines the design values of a product. We focus on the design problem of a shadow mask which is a core component of monitors in the electronics industry. In case-based reasoning (CBR), it is important to retrieve similar cases and adapt them to meet design specifications exactly. Notably, difficulties in automating the adaptation process have prevented designers from being able to use design expert systems easily and efficiently. In this paper, we present a hybrid approach combining CBR and artificial neural networks in order to solve the problems occurring during the adaptation process. We first constructed a radial basis function network (RBFN) composed of representative cases created by K-means clustering. Then, the representative case most similar to the current problem was adjusted using the network. The rationale behind the proposed approach is discussed, and experimental results acquired from real shadow mask design are presented. Using the design expert system, designers can reduce design time and errors and enhance the total quality of design. Furthermore, the expert system facilitates effective sharing of design knowledge among designers.} -} - -@article{CHIU2023100118, -title = {Systematic literature review on opportunities, challenges, and future research recommendations of artificial intelligence in education}, -journal = {Computers and Education: Artificial Intelligence}, -volume = {4}, -language = {English}, -type = {article}, -pages = {100118}, -year = {2023}, -issn = {2666-920X}, -doi = {https://doi.org/10.1016/j.caeai.2022.100118}, -url = {https://www.sciencedirect.com/science/article/pii/S2666920X2200073X}, -author = {Thomas K.F. Chiu and Qi Xia and Xinyan Zhou and Ching Sing Chai and Miaoting Cheng}, -keywords = {Artificial intelligence, Artificial intelligence in education, Systematic review, Learning, Teaching, Assessment}, -abstract = {Applications of artificial intelligence in education (AIEd) are emerging and are new to researchers and practitioners alike. Reviews of the relevant literature have not examined how AI technologies have been integrated into each of the four key educational domains of learning, teaching, assessment, and administration. The relationships between the technologies and learning outcomes for students and teachers have also been neglected. This systematic review study aims to understand the opportunities and challenges of AIEd by examining the literature from the last 10 years (2012–2021) using matrix coding and content analysis approaches. The results present the current focus of AIEd research by identifying 13 roles of AI technologies in the key educational domains, 7 learning outcomes of AIEd, and 10 major challenges. The review also provides suggestions for future directions of AIEd research.} -} - -@article{Robertson2014ARO, -title = {A Review of Real-Time Strategy Game AI}, -author = {Glen Robertson and Ian D. Watson}, -affiliation = {University of Auckland }, -keywords = {Game, IA, Real-time strategy}, -type={article}, -language={English}, -abstract = {This literature review covers AI techniques used for real-time strategy video games, focusing specifically on StarCraft. It finds that the main areas of current academic research are in tactical and strategic decision making, plan recognition, and learning, and it outlines the research contributions in each of these areas. The paper then contrasts the use of game AI in academe and industry, finding the academic research heavily focused on creating game-winning agents, while the industry aims to maximize player enjoyment. It finds that industry adoption of academic research is low because it is either inapplicable or too time-consuming and risky to implement in a new game, which highlights an area for potential investigation: bridging the gap between academe and industry. Finally, the areas of spatial reasoning, multiscale AI, and cooperation are found to require future work, and standardized evaluation methods are proposed to produce comparable results between studies.}, -journal = {AI Mag.}, -year = {2014}, -volume = {35}, -pages = {75-104} -} - -@Inproceedings{10.1007/978-3-642-15973-2_50, -author={Butdee, S. -and Tichkiewitch, S.}, -affiliation={University of Technology North Bangkok; Grenoble Institute of Technology}, -editor={Bernard, Alain}, -title={Case-Based Reasoning for Adaptive Aluminum Extrusion Die Design Together with Parameters by Neural Networks}, -keywords={Adaptive die design and parameters, Optimal aluminum extrusion, Case-based reasoning, Neural networks}, -booktitle={Global Product Development}, -year={2011}, -type = {article; proceedings paper}, -language = {English}, -publisher = {Springer Berlin Heidelberg}, -address = {Berlin, Heidelberg}, -pages = {491--496}, -abstract = {Nowadays Aluminum extrusion die design is a critical task for improving productivity which involves with quality, time and cost. Case-Based Reasoning (CBR) method has been successfully applied to support the die design process in order to design a new die by tackling previous problems together with their solutions to match with a new similar problem. Such solutions are selected and modified to solve the present problem. However, the applications of the CBR are useful only retrieving previous features whereas the critical parameters are missing. In additions, the experience learning to such parameters are limited. This chapter proposes Artificial Neural Network (ANN) to associate the CBR in order to learning previous parameters and predict to the new die design according to the primitive die modification. The most satisfactory is to accommodate the optimal parameters of extrusion processes.}, -isbn = {978-3-642-15973-2} -} - -@Inproceedings{10.1007/978-3-319-47096-2_11, -author={Grace, Kazjon -and Maher, Mary Lou -and Wilson, David C. -and Najjar, Nadia A.}, -affiliation={University of North Carolina at Charlotte}, -editor={Goel, Ashok -and D{\'i}az-Agudo, M Bel{\'e}n -and Roth-Berghofer, Thomas}, -title={Combining CBR and Deep Learning to Generate Surprising Recipe Designs}, -keywords={Case-based reasoning, deep learning, recipe design}, -type = {article; proceedings paper}, -booktitle={Case-Based Reasoning Research and Development}, -year={2016}, -publisher={Springer International Publishing}, -address={Cham}, -language = {English}, -pages={154--169}, -abstract={This paper presents a dual-cycle CBR model in the domain of recipe generation. The model combines the strengths of deep learning and similarity-based retrieval to generate recipes that are novel and valuable (i.e. they are creative). The first cycle generates abstract descriptions which we call ``design concepts'' by synthesizing expectations from the entire case base, while the second cycle uses those concepts to retrieve and adapt objects. We define these conceptual object representations as an abstraction over complete cases on which expectations can be formed, allowing objects to be evaluated for surprisingness (the peak level of unexpectedness in the object, given the case base) and plausibility (the overall similarity of the object to those in the case base). The paper presents a prototype implementation of the model, and demonstrates its ability to generate objects that are simultaneously plausible and surprising, in addition to fitting a user query. This prototype is then compared to a traditional single-cycle CBR system.}, -isbn={978-3-319-47096-2} -} - -@Inproceedings{10.1007/978-3-319-61030-6_1, -author={Maher, Mary Lou -and Grace, Kazjon}, -editor={Aha, David W. -and Lieber, Jean}, -affiliation={University of North Carolina at Charlotte}, -title={Encouraging Curiosity in Case-Based Reasoning and Recommender Systems}, -keywords={Curiosity, Case-based reasoning, Recommender systems}, -booktitle={Case-Based Reasoning Research and Development}, -year={2017}, -publisher={Springer International Publishing}, -address={Cham}, -pages={3--15}, -language = {English}, -type = {article; proceedings paper}, -abstract={A key benefit of case-based reasoning (CBR) and recommender systems is the use of past experience to guide the synthesis or selection of the best solution for a specific context or user. Typically, the solution presented to the user is based on a value system that privileges the closest match in a query and the solution that performs best when evaluated according to predefined requirements. In domains in which creativity is desirable or the user is engaged in a learning activity, there is a benefit to moving beyond the expected or ``best match'' and include results based on computational models of novelty and surprise. In this paper, models of novelty and surprise are integrated with both CBR and Recommender Systems to encourage user curiosity.}, -isbn={978-3-319-61030-6} -} - -@Inproceedings{Muller, -author = {Müller, G. and Bergmann, R.}, -affiliation={University of Trier}, -year = {2015}, -month = {01}, -language = {English}, -type = {article; proceedings paper}, -abstract = {This paper presents CookingCAKE,a framework for the adaptation of cooking recipes represented as workflows. CookingCAKE integrates and combines several workflow adaptation approaches applied in process-oriented case based reasoning (POCBR) in a single adaptation framework, thus providing a capable tool for the adaptation of cooking recipes. The available case base of cooking workflows is analyzed to generate adaptation knowledge which is used to adapt a recipe regarding restrictions and resources, which the user may define for the preparation of a dish.}, -booktitle = {International Conference on Case-Based Reasoning}, -title = {CookingCAKE: A Framework for the adaptation of cooking recipes represented as workflows}, -keywords={recipe adaptation, workflow adaptation, workflows, process-oriented, case based reasoning} -} - -@Inproceedings{10.1007/978-3-319-24586-7_20, -author={Onta{\~{n}}{\'o}n, S. -and Plaza, E. -and Zhu, J.}, -editor={H{\"u}llermeier, Eyke -and Minor, Mirjam}, -affiliation={Drexel University; Artificial Intelligence Research Institute CSIC}, -title={Argument-Based Case Revision in CBR for Story Generation}, -keywords={CBR, Case-based reasoning, Story generation}, -booktitle={Case-Based Reasoning Research and Development}, -year={2015}, -publisher={Springer International Publishing}, -address={Cham}, -language = {English}, -pages={290--305}, -type = {article; proceedings paper}, -abstract={This paper presents a new approach to case revision in case-based reasoning based on the idea of argumentation. Previous work on case reuse has proposed the use of operations such as case amalgamation (or merging), which generate solutions by combining information coming from different cases. Such approaches are often based on exploring the search space of possible combinations looking for a solution that maximizes a certain criteria. We show how Revise can be performed by arguments attacking specific parts of a case produced by Reuse, and how they can guide and prevent repeating pitfalls in future cases. The proposed approach is evaluated in the task of automatic story generation.}, -isbn={978-3-319-24586-7} -} - -@Inproceedings{10.1007/978-3-030-58342-2_20, -author={Lepage, Yves -and Lieber, Jean -and Mornard, Isabelle -and Nauer, Emmanuel -and Romary, Julien -and Sies, Reynault}, -editor={Watson, Ian -and Weber, Rosina}, -title={The French Correction: When Retrieval Is Harder to Specify than Adaptation}, -affiliation={Waseda University; Université de Lorraine}, -keywords={case-based reasoning, retrieval, analogy, sentence correction}, -booktitle={Case-Based Reasoning Research and Development}, -year={2020}, -language = {English}, -type = {article; proceedings paper}, -publisher={Springer International Publishing}, -address={Cham}, -pages={309--324}, -abstract={A common idea in the field of case-based reasoning is that the retrieval step can be specified by the use of some similarity measure: the retrieved cases maximize the similarity to the target problem and, then, the adaptation step has to take into account the mismatches between the retrieved cases and the target problem in order to this latter. The use of this methodological schema for the application described in this paper has proven to be non efficient. Indeed, designing a retrieval procedure without the precise knowledge of the adaptation procedure has not been possible. The domain of this application is the correction of French sentences: a problem is an incorrect sentence and a valid solution is a correction of this problem. Adaptation consists in solving an analogical equation that enables to execute the correction of the retrieved case on the target problem. Thus, retrieval has to ensure that this application is feasible. The first version of such a retrieval procedure is described and evaluated: it is a knowledge-light procedure that does not use linguistic knowledge about French.}, -isbn={978-3-030-58342-2} -} - -@Inproceedings{10.1007/978-3-030-01081-2_25, -author={Smyth, Barry -and Cunningham, P{\'a}draig}, -editor={Cox, Michael T. -and Funk, Peter -and Begum, Shahina}, -affiliation={University College Dublin}, -title={An Analysis of Case Representations for Marathon Race Prediction and Planning}, -keywords={Marathon planning, Case representation, Case-based reasoning}, -booktitle={Case-Based Reasoning Research and Development}, -year={2018}, -language = {English}, -publisher={Springer International Publishing}, -address={Cham}, -pages={369--384}, -type = {article; proceedings paper}, -abstract={We use case-based reasoning to help marathoners achieve a personal best for an upcoming race, by helping them to select an achievable goal-time and a suitable pacing plan. We evaluate several case representations and, using real-world race data, highlight their performance implications. Richer representations do not always deliver better prediction performance, but certain representational configurations do offer very significant practical benefits for runners, when it comes to predicting, and planning for, challenging goal-times during an upcoming race.}, -isbn={978-3-030-01081-2} -} - -@Inproceedings{10.1007/978-3-030-58342-2_8, -author={Smyth, Barry -and Willemsen, Martijn C.}, -editor={Watson, Ian -and Weber, Rosina}, -affiliation={University College Dublin; Eindhoven University of Technology}, -title={Predicting the Personal-Best Times of Speed Skaters Using Case-Based Reasoning}, -keywords={CBR for health and exercise, speed skating, race-time prediction, case representation}, -booktitle={Case-Based Reasoning Research and Development}, -year={2020}, -type = {article; proceedings paper}, -language = {English}, -publisher={Springer International Publishing}, -address={Cham}, -pages={112--126}, -abstract={Speed skating is a form of ice skating in which the skaters race each other over a variety of standardised distances. Races take place on specialised ice-rinks and the type of track and ice conditions can have a significant impact on race-times. As race distances increase, pacing also plays an important role. In this paper we seek to extend recent work on the application of case-based reasoning to marathon-time prediction by predicting race-times for speed skaters. In particular, we propose and evaluate a number of case-based reasoning variants based on different case and feature representations to generate track-specific race predictions. We show it is possible to improve upon state-of-the-art prediction accuracy by harnessing richer case representations using shorter races and track-adjusted finish and lap-times.}, -isbn={978-3-030-58342-2} -} - -@Inproceedings{10.1007/978-3-030-58342-2_5, -author={Feely, Ciara -and Caulfield, Brian -and Lawlor, Aonghus -and Smyth, Barry}, -editor={Watson, Ian -and Weber, Rosina}, -affiliation={University College Dublin}, -title={Using Case-Based Reasoning to Predict Marathon Performance and Recommend Tailored Training Plans}, -keywords={CBR for health and exercise, marathon running, race-time prediction, plan recommendation}, -booktitle={Case-Based Reasoning Research and Development}, -year={2020}, -language = {English}, -publisher={Springer International Publishing}, -address={Cham}, -pages={67--81}, -type = {article; proceedings paper}, -abstract={Training for the marathon, especially a first marathon, is always a challenge. Many runners struggle to find the right balance between their workouts and their recovery, often leading to sub-optimal performance on race-day or even injury during training. We describe and evaluate a novel case-based reasoning system to help marathon runners as they train in two ways. First, it uses a case-base of training/workouts and race histories to predict future marathon times for a target runner, throughout their training program, helping runners to calibrate their progress and, ultimately, plan their race-day pacing. Second, the system recommends tailored training plans to runners, adapted for their current goal-time target, and based on the training plans of similar runners who have achieved this time. We evaluate the system using a dataset of more than 21,000 unique runners and 1.5 million training/workout sessions.}, -isbn={978-3-030-58342-2} -} - -@article{LALITHA2020583, -title = {Personalised Self-Directed Learning Recommendation System}, -journal = {Procedia Computer Science}, -volume = {171}, -pages = {583-592}, -year = {2020}, -type = {article}, -language = {English}, -note = {Third International Conference on Computing and Network Communications (CoCoNet'19)}, -issn = {1877-0509}, -doi = {https://doi.org/10.1016/j.procs.2020.04.063}, -url = {https://www.sciencedirect.com/science/article/pii/S1877050920310309}, -author = {T B Lalitha and P S Sreeja}, -affiliation={Hindustan Institute of Technology and Science}, -keywords = {e-Learning, PSDLR, Recommendation System, SDL, Self-Directed Learning}, -abstract = {Modern educational systems have changed drastically bringing in knowledge anywhere as needed by the learner with the evolution of Internet. Availability of knowledge in public domain, capability of exchanging large amount of information and filtering relevant information quickly has enabled disruption to conventional educational system. Thus, future trends are looking towards E-Learning (Electronic Learning) and M-Learning (Mobile Learning) technologies over the Internet for their vast knowledge acquisition. In this paper, the work gives an elaborate context of learning strategies prevailing and emerging with the classification of e-learning Techniques. It majorly focuses on the features and variety of aspects with the e-learning and the choice of learning method involved and facilitate the adoption of new ways for personalized selection on learning resources for SDL (Self-Directed Learning) from the unstructured, large web-based environment. Thereby, proposes a Personalised Self-Directed Learning Recommendation System (PSDLR) based on the personal specifications of the SDL learner. The result offers insight into the perspectives and challenges of Self-Directed Learning based on cognitive and constructive characteristics which majorly incorporates web-based learning and gives path in finding appropriate solutions using machine learning techniques and ontology for the open problems in the respective fields with personalised recommendations and guidelines for future research.} -} - -@article{Zhou2021, -author={Zhou, Lina -and Wang, Chunxia}, -affiliation={Baotou Medical College}, -title={Research on Recommendation of Personalized Exercises in English Learning Based on Data Mining}, -journal={Scientific Programming}, -year={2021}, -month={Dec}, -type = {article}, -language = {English}, -day={21}, -publisher={Hindawi}, -keywords={Recommender systems, Learning}, -volume={2021}, -pages={5042286}, -abstract={Aiming at the problems of traditional method of exercise recommendation precision, recall rate, long recommendation time, and poor recommendation comprehensiveness, this study proposes a personalized exercise recommendation method for English learning based on data mining. Firstly, a personalized recommendation model is designed, based on the model to preprocess the data in the Web access log, and cleaning the noise data to avoid its impact on the accuracy of the recommendation results is focused; secondly, the DINA model to diagnose the degree of mastery of students{\&}{\#}x2019; knowledge points is used and the students{\&}{\#}x2019; browsing patterns through fuzzy similar relationships are clustered; and finally, according to the clustering results, the similarity between students and the similarity between exercises are measured, and the collaborative filtering recommendation of personalized exercises for English learning is realized. The experimental results show that the exercise recommendation precision and recall rate of this method are higher, the recommendation time is shorter, and the recommendation results are comprehensive.}, -issn={1058-9244}, -doi={10.1155/2021/5042286}, -url={https://doi.org/10.1155/2021/5042286} -} - -@article{INGKAVARA2022100086, -title = {The use of a personalized learning approach to implementing self-regulated online learning}, -journal = {Computers and Education: Artificial Intelligence}, -volume = {3}, -pages = {100086}, -type = {article}, -language = {English}, -year = {2022}, -issn = {2666-920X}, -doi = {https://doi.org/10.1016/j.caeai.2022.100086}, -url = {https://www.sciencedirect.com/science/article/pii/S2666920X22000418}, -author = {Thanyaluck Ingkavara and Patcharin Panjaburee and Niwat Srisawasdi and Suthiporn Sajjapanroj}, -keywords = {Intelligent tutoring system, Personalization, Adaptive learning, E-learning, TAM, Artificial intelligence}, -abstract = {Nowadays, students are encouraged to learn via online learning systems to promote students' autonomy. Scholars have found that students' self-regulated actions impact their academic success in an online learning environment. However, because traditional online learning systems cannot personalize feedback to the student's personality, most students have less chance to obtain helpful suggestions for enhancing their knowledge linked to their learning problems. This paper incorporated self-regulated online learning in the Physics classroom and used a personalized learning approach to help students receive proper learning paths and material corresponding to their learning preferences. This study conducted a quasi-experimental design using a quantitative approach to evaluate the effectiveness of the proposed learning environment in secondary schools. The experimental group of students participated in self-regulated online learning with a personalized learning approach, while the control group participated in conventional self-regulated online learning. The experimental results showed that the experimental group's post-test and the learning-gain score of the experimental group were significantly higher than those of the control group. Moreover, the results also suggested that the student's perceptions about the usefulness of learning suggestions, ease of use, goal setting, learning environmental structuring, task strategies, time management, self-evaluation, impact on learning, and attitude toward the learning environment are important predictors of behavioral intention to learn with the self-regulated online learning that integrated with the personalized learning approach.} -} - -@article{HUANG2023104684, -title = {Effects of artificial Intelligence–Enabled personalized recommendations on learners’ learning engagement, motivation, and outcomes in a flipped classroom}, -journal = {Computers and Education}, -volume = {194}, -pages = {104684}, -year = {2023}, -language = {English}, -type = {article}, -issn = {0360-1315}, -doi = {https://doi.org/10.1016/j.compedu.2022.104684}, -url = {https://www.sciencedirect.com/science/article/pii/S036013152200255X}, -author = {Anna Y.Q. Huang and Owen H.T. Lu and Stephen J.H. Yang}, -keywords = {Data science applications in education, Distance education and online learning, Improving classroom teaching}, -abstract = {The flipped classroom approach is aimed at improving learning outcomes by promoting learning motivation and engagement. Recommendation systems can also be used to improve learning outcomes. With the rapid development of artificial intelligence (AI) technology, various systems have been developed to facilitate student learning. Accordingly, we applied AI-enabled personalized video recommendations to stimulate students' learning motivation and engagement during a systems programming course in a flipped classroom setting. We assigned students to control and experimental groups comprising 59 and 43 college students, respectively. The students in both groups received flipped classroom instruction, but only those in the experimental group received AI-enabled personalized video recommendations. We quantitatively measured students’ engagement based on their learning profiles in a learning management system. The results revealed that the AI-enabled personalized video recommendations could significantly improve the learning performance and engagement of students with a moderate motivation level.} -} - -@article{ZHAO2023118535, -title = {A recommendation system for effective learning strategies: An integrated approach using context-dependent DEA}, -journal = {Expert Systems with Applications}, -volume = {211}, -pages = {118535}, -year = {2023}, -language = {English}, -type = {article}, -issn = {0957-4174}, -doi = {https://doi.org/10.1016/j.eswa.2022.118535}, -url = {https://www.sciencedirect.com/science/article/pii/S0957417422016104}, -author = {Lu-Tao Zhao and Dai-Song Wang and Feng-Yun Liang and Jian Chen}, -keywords = {Recommendation system, Learning strategies, Context-dependent DEA, Efficiency analysis}, -abstract = {Universities have been focusing on increasing individualized training and providing appropriate education for students. The individual differences and learning needs of college students should be given enough attention. From the perspective of learning efficiency, we establish a clustering hierarchical progressive improvement model (CHPI), which is based on cluster analysis and context-dependent data envelopment analysis (DEA) methods. The CHPI clusters students' ontological features, employs the context-dependent DEA method to stratify students of different classes, and calculates measures, such as obstacles, to determine the reference path for individuals with inefficient learning processes. The learning strategies are determined according to the gap between the inefficient individual to be improved and the individuals on the reference path. By the study of college English courses as an example, it is found that the CHPI can accurately recommend targeted learning strategies to satisfy the individual needs of college students so that the learning of individuals with inefficient learning processes in a certain stage can be effectively improved. In addition, CHPI can provide specific, efficient suggestions to improve learning efficiency comparing to existing recommendation systems, and has great potential in promoting the integration of education-related researches and expert systems.} -} - -@article{SU2022109547, -title = {Graph-based cognitive diagnosis for intelligent tutoring systems}, -journal = {Knowledge-Based Systems}, -volume = {253}, -pages = {109547}, -year = {2022}, -language = {English}, -type = {article}, -issn = {0950-7051}, -doi = {https://doi.org/10.1016/j.knosys.2022.109547}, -url = {https://www.sciencedirect.com/science/article/pii/S095070512200778X}, -author = {Yu Su and Zeyu Cheng and Jinze Wu and Yanmin Dong and Zhenya Huang and Le Wu and Enhong Chen and Shijin Wang and Fei Xie}, -keywords = {Cognitive diagnosis, Graph neural networks, Interpretable machine learning}, -abstract = {For intelligent tutoring systems, Cognitive Diagnosis (CD) is a fundamental task that aims to estimate the mastery degree of a student on each skill according to the exercise record. The CD task is considered rather challenging since we need to model inner-relations and inter-relations among students, skills, and questions to obtain more abundant information. Most existing methods attempt to solve this problem through two-way interactions between students and questions (or between students and skills), ignoring potential high-order relations among entities. Furthermore, how to construct an end-to-end framework that can model the complex interactions among different types of entities at the same time remains unexplored. Therefore, in this paper, we propose a graph-based Cognitive Diagnosis model (GCDM) that directly discovers the interactions among students, skills, and questions through a heterogeneous cognitive graph. Specifically, we design two graph-based layers: a performance-relative propagator and an attentive knowledge aggregator. The former is applied to propagate a student’s cognitive state through different types of graph edges, while the latter selectively gathers messages from neighboring graph nodes. Extensive experimental results on two real-world datasets clearly show the effectiveness and extendibility of our proposed model.} -} - -@article{EZALDEEN2022100700, -title = {A hybrid E-learning recommendation integrating adaptive profiling and sentiment analysis}, -journal = {Journal of Web Semantics}, -volume = {72}, -pages = {100700}, -year = {2022}, -type = {article}, -language = {English}, -issn = {1570-8268}, -doi = {https://doi.org/10.1016/j.websem.2021.100700}, -url = {https://www.sciencedirect.com/science/article/pii/S1570826821000664}, -author = {Hadi Ezaldeen and Rachita Misra and Sukant Kishoro Bisoy and Rawaa Alatrash and Rojalina Priyadarshini}, -keywords = {Hybrid E-learning recommendation, Adaptive profiling, Semantic learner profile, Fine-grained sentiment analysis, Convolutional Neural Network, Word embeddings}, -abstract = {This research proposes a novel framework named Enhanced e-Learning Hybrid Recommender System (ELHRS) that provides an appropriate e-content with the highest predicted ratings corresponding to the learner’s particular needs. To accomplish this, a new model is developed to deduce the Semantic Learner Profile automatically. It adaptively associates the learning patterns and rules depending on the learner’s behavior and the semantic relations computed in the semantic matrix that mutually links e-learning materials and terms. Here, a semantic-based approach for term expansion is introduced using DBpedia and WordNet ontologies. Further, various sentiment analysis models are proposed and incorporated as a part of the recommender system to predict ratings of e-learning resources from posted text reviews utilizing fine-grained sentiment classification on five discrete classes. Qualitative Natural Language Processing (NLP) methods with tailored-made Convolutional Neural Network (CNN) are developed and evaluated on our customized dataset collected for a specific domain and a public dataset. Two improved language models are introduced depending on Skip-Gram (S-G) and Continuous Bag of Words (CBOW) techniques. In addition, a robust language model based on hybridization of these couple of methods is developed to derive better vocabulary representation, yielding better accuracy 89.1% for the CNN-Three-Channel-Concatenation model. The suggested recommendation methodology depends on the learner’s preferences, other similar learners’ experience and background, deriving their opinions from the reviews towards the best learning resources. This assists the learners in finding the desired e-content at the proper time.} -} - -@article{MUANGPRATHUB2020e05227, -title = {Learning recommendation with formal concept analysis for intelligent tutoring system}, -journal = {Heliyon}, -volume = {6}, -number = {10}, -pages = {e05227}, -language = {English}, -type = {article}, -year = {2020}, -issn = {2405-8440}, -doi = {https://doi.org/10.1016/j.heliyon.2020.e05227}, -url = {https://www.sciencedirect.com/science/article/pii/S2405844020320703}, -author = {Jirapond Muangprathub and Veera Boonjing and Kosin Chamnongthai}, -keywords = {Computer Science, Learning recommendation, Formal concept analysis, Intelligent tutoring system, Adaptive learning}, -abstract = {The aim of this research was to develop a learning recommendation component in an intelligent tutoring system (ITS) that dynamically predicts and adapts to a learner's style. In order to develop a proper ITS, we present an improved knowledge base supporting adaptive learning, which can be achieved by a suitable knowledge construction. This process is illustrated by implementing a web-based online tutor system. In addition, our knowledge structure provides adaptive presentation and personalized learning with the proposed adaptive algorithm, to retrieve content according to individual learner characteristics. To demonstrate the proposed adaptive algorithm, pre-test and post-test were used to evaluate suggestion accuracy of the course in a class for adapting to a learner's style. In addition, pre- and post-testing were also used with students in a real teaching/learning environment to evaluate the performance of the proposed model. The results show that the proposed system can be used to help students or learners achieve improved learning.} -} - -@article{min8100434, -author = {Leikola, Maria and Sauer, Christian and Rintala, Lotta and Aromaa, Jari and Lundström, Mari}, -title = {Assessing the Similarity of Cyanide-Free Gold Leaching Processes: A Case-Based Reasoning Application}, -journal = {Minerals}, -volume = {8}, -type = {article}, -language = {English}, -year = {2018}, -number = {10}, -url = {https://www.mdpi.com/2075-163X/8/10/434}, -issn = {2075-163X}, -keywords={hydrometallurgy, cyanide-free gold, knowledge modelling, case-based reasoning, information retrieval}, -abstract = {Hydrometallurgical researchers, and other professionals alike, invest significant amounts of time reading scientific articles, technical notes, and other scientific documents, while looking for the most relevant information for their particular research interest. In an attempt to save the researcher’s time, this study presents an information retrieval tool using case-based reasoning. The tool was built for comparing scientific articles concerning cyanide-free leaching of gold ores/concentrates/tailings. Altogether, 50 cases of experiments were gathered in a case base. 15 different attributes related to the treatment of the raw material and the leaching conditions were selected to compare the cases. The attributes were as follows: Pretreatment, Overall method, Complexant source, Oxidant source, Complexant concentration, Oxidant concentration, Temperature, pH, Redox-potential, Pressure, Materials of construction, Extraction, Extraction rate, Reagent consumption, and Solid-liquid ratio. The resulting retrieval tool (LeachSim) was able to rank the scientific articles according to their similarity with the user’s research interest. Such a tool could eventually aid the user in finding the most relevant information, but not replace thorough understanding and human expertise.}, -doi = {10.3390/min8100434} -} - -@article{10.1145/3459665, -author = {Cunningham, P\'{a}draig and Delany, Sarah Jane}, -title = {K-Nearest Neighbour Classifiers - A Tutorial}, -year = {2021}, -issue_date = {July 2022}, -publisher = {Association for Computing Machinery}, -address = {New York, NY, USA}, -type={article}, -language={English}, -volume = {54}, -number = {6}, -issn = {0360-0300}, -url = {https://doi.org/10.1145/3459665}, -doi = {10.1145/3459665}, -abstract = {Perhaps the most straightforward classifier in the arsenal or Machine Learning techniques is the Nearest Neighbour Classifier—classification is achieved by identifying the nearest neighbours to a query example and using those neighbours to determine the class of the query. This approach to classification is of particular importance, because issues of poor runtime performance is not such a problem these days with the computational power that is available. This article presents an overview of techniques for Nearest Neighbour classification focusing on: mechanisms for assessing similarity (distance), computational issues in identifying nearest neighbours, and mechanisms for reducing the dimension of the data.This article is the second edition of a paper previously published as a technical report [16]. Sections on similarity measures for time-series, retrieval speedup, and intrinsic dimensionality have been added. An Appendix is included, providing access to Python code for the key methods.}, -journal = {ACM Comput. Surv.}, -month = {jul}, -articleno = {128}, -numpages = {25}, -keywords = {k-Nearest neighbour classifiers} -} - -@article{9072123, -author={Sinaga, Kristina P. and Yang, Miin-Shen}, -journal={IEEE Access}, -type={article}, -language={English}, -title={Unsupervised K-Means Clustering Algorithm}, -year={2020}, -volume={8}, -number={}, -pages={80716-80727}, -doi={10.1109/ACCESS.2020.2988796} -} - -@article{WANG2021331, -title = {A new prediction strategy for dynamic multi-objective optimization using Gaussian Mixture Model}, -journal = {Information Sciences}, -volume = {580}, -type = {article}, -language = {English}, -pages = {331-351}, -year = {2021}, -issn = {0020-0255}, -doi = {https://doi.org/10.1016/j.ins.2021.08.065}, -url = {https://www.sciencedirect.com/science/article/pii/S0020025521008732}, -author = {Feng Wang and Fanshu Liao and Yixuan Li and Hui Wang}, -keywords = {Dynamic multi-objective optimization, Gaussian Mixture Model, Change type detection, Resampling}, -abstract = {Dynamic multi-objective optimization problems (DMOPs), in which the environments change over time, have attracted many researchers’ attention in recent years. Since the Pareto set (PS) or the Pareto front (PF) can change over time, how to track the movement of the PS or PF is a challenging problem in DMOPs. Over the past few years, lots of methods have been proposed, and the prediction based strategy has been considered the most effective way to track the new PS. However, the performance of most existing prediction strategies depends greatly on the quantity and quality of the historical information and will deteriorate due to non-linear changes, leading to poor results. In this paper, we propose a new prediction method, named MOEA/D-GMM, which incorporates the Gaussian Mixture Model (GMM) into the MOEA/D framework for the prediction of the new PS when changes occur. Since GMM is a powerful non-linear model to accurately fit various data distributions, it can effectively generate solutions with better quality according to the distributions. In the proposed algorithm, a change type detection strategy is first designed to estimate an approximate PS according to different change types. Then, GMM is employed to make a more accurate prediction by training it with the approximate PS. To overcome the shortcoming of a lack of training solutions for GMM, the Empirical Cumulative Distribution Function (ECDF) method is used to resample more training solutions before GMM training. Experimental results on various benchmark test problems and a classical real-world problem show that, compared with some state-of-the-art dynamic optimization algorithms, MOEA/D-GMM outperforms others in most cases.} -} - -@article{9627973, -author={Xu, Shengbing and Cai, Wei and Xia, Hongxi and Liu, Bo and Xu, Jie}, -journal={IEEE Access}, -title={Dynamic Metric Accelerated Method for Fuzzy Clustering}, -year={2021}, -type={article}, -language={English}, -volume={9}, -number={}, -pages={166838-166854}, -doi={10.1109/ACCESS.2021.3131368} -} - -@article{9434422, -author={Gupta, Samarth and Chaudhari, Shreyas and Joshi, Gauri and Yağan, Osman}, -journal={IEEE Transactions on Information Theory}, -title={Multi-Armed Bandits With Correlated Arms}, -year={2021}, -language={English}, -type={article}, -volume={67}, -number={10}, -pages={6711-6732}, -doi={10.1109/TIT.2021.3081508} -} - -@Inproceedings{8495930, -author={Supic, H.}, -booktitle={2018 IEEE 27th International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE)}, -title={Case-Based Reasoning Model for Personalized Learning Path Recommendation in Example-Based Learning Activities}, -year={2018}, -type={article}, -language={English}, -volume={}, -number={}, -pages={175-178}, -doi={10.1109/WETICE.2018.00040} -} - -@Inproceedings{9870279, -author={Lin, Baihan}, -booktitle={2022 IEEE Congress on Evolutionary Computation (CEC)}, -title={Evolutionary Multi-Armed Bandits with Genetic Thompson Sampling}, -year={2022}, -type={article}, -language={English}, -volume={}, -number={}, -pages={1-8}, -doi={10.1109/CEC55065.2022.9870279} -} - -@article{Obeid, -author={Obeid, C. and Lahoud, C. and Khoury, H. E. and Champin, P.}, -title={A Novel Hybrid Recommender System Approach for Student Academic Advising Named COHRS, Supported by Case-based Reasoning and Ontology}, -journal={Computer Science and Information Systems}, -type={article}, -language={English}, -volume={19}, -number={2}, -pages={979–1005}, -year={2022}, -doi={https://doi.org/10.2298/CSIS220215011O} -} - -@book{Nkambou, -author = {Nkambou, R. and Bourdeau, J. and Mizoguchi, R.}, -title = {Advances in Intelligent Tutoring Systems}, -year = {2010}, -type = {article}, -language = {English}, -publisher = {Springer Berlin, Heidelberg}, -edition = {1} -} - -@book{hajduk2019cognitive, -title={Cognitive Multi-agent Systems: Structures, Strategies and Applications to Mobile Robotics and Robosoccer}, -author={Hajduk, M. and Sukop, M. and Haun, M.}, -type={book}, -language={English}, -isbn={9783319936857}, -series={Studies in Systems, Decision and Control}, -year={2019}, -publisher={Springer International Publishing} -} - -@article{RICHTER20093, -title = {The search for knowledge, contexts, and Case-Based Reasoning}, -journal = {Engineering Applications of Artificial Intelligence}, -language = {English}, -type = {article}, -volume = {22}, -number = {1}, -pages = {3-9}, -year = {2009}, -issn = {0952-1976}, -doi = {https://doi.org/10.1016/j.engappai.2008.04.021}, -url = {https://www.sciencedirect.com/science/article/pii/S095219760800078X}, -author = {Michael M. Richter}, -keywords = {Case-Based Reasoning, Knowledge, Processes, Utility, Context}, -abstract = {A major goal of this paper is to compare Case-Based Reasoning with other methods searching for knowledge. We consider knowledge as a resource that can be traded. It has no value in itself; the value is measured by the usefulness of applying it in some process. Such a process has info-needs that have to be satisfied. The concept to measure this is the economical term utility. In general, utility depends on the user and its context, i.e., it is subjective. Here, we introduce levels of contexts from general to individual. We illustrate that Case-Based Reasoning on the lower, i.e., more personal levels CBR is quite useful, in particular in comparison with traditional informational retrieval methods.} -} - -@Thesis{Marie, -author={Marie, F.}, -title={COLISEUM-3D. Une plate-forme innovante pour la segmentation d’images médicales par Raisonnement à Partir de Cas (RàPC) et méthodes d’apprentissage de type Deep Learning}, -type={diplomathesis}, -language={French}, -institution={Université de Franche-Comte}, -year={2019} -} - -@book{Hoang, -title = {La formule du savoir. Une philosophie unifiée du savoir fondée sur le théorème de Bayes}, -author = {Hoang, L.N.}, -type = {book}, -language = {French}, -isbn = {9782759822607}, -year = {2018}, -publisher = {EDP Sciences} -} - -@book{Richter2013, -title={Case-Based Reasoning (A Textbook)}, -author={Richter, M. and Weber, R.}, -type={book}, -language={English}, -isbn={9783642401664}, -year={2013}, -publisher={Springer-Verlag GmbH} -} - -@book{kedia2020hands, -title={Hands-On Python Natural Language Processing: Explore tools and techniques to analyze and process text with a view to building real-world NLP applications}, -author={Kedia, A. and Rasu, M.}, -language={English}, -type={book}, -isbn={9781838982584}, -url={https://books.google.fr/books?id=1AbuDwAAQBAJ}, -year={2020}, -publisher={Packt Publishing} -} - -@book{ghosh2019natural, -title={Natural Language Processing Fundamentals: Build intelligent applications that can interpret the human language to deliver impactful results}, -author={Ghosh, S. and Gunning, D.}, -language={English}, -type={book}, -isbn={9781789955989}, -url={https://books.google.fr/books?id=i8-PDwAAQBAJ}, -year={2019}, -publisher={Packt Publishing} -} - -@article{Akerblom, -title={Online learning of network bottlenecks via minimax paths}, -author={kerblom, Niklas and Hoseini, Fazeleh Sadat and Haghir Chehreghani, Morteza}, -language={English}, -type={article}, -volume = {122}, -year = {2023}, -issn = {1573-0565}, -doi = {https://doi.org/10.1007/s10994-022-06270-0}, -url = {https://doi.org/10.1007/s10994-022-06270-0}, -abstract={In this paper, we study bottleneck identification in networks via extracting minimax paths. Many real-world networks have stochastic weights for which full knowledge is not available in advance. Therefore, we model this task as a combinatorial semi-bandit problem to which we apply a combinatorial version of Thompson Sampling and establish an upper bound on the corresponding Bayesian regret. Due to the computational intractability of the problem, we then devise an alternative problem formulation which approximates the original objective. Finally, we experimentally evaluate the performance of Thompson Sampling with the approximate formulation on real-world directed and undirected networks.} -} - -@article{Simen, -title={Dynamic slate recommendation with gated recurrent units and Thompson sampling}, -author={Eide, Simen and Leslie, David S. and Frigessi, Arnoldo}, -language={English}, -type={article}, -volume = {36}, -year = {2022}, -issn = {1573-756X}, -doi = {https://doi.org/10.1007/s10618-022-00849-w}, -url = {https://doi.org/10.1007/s10618-022-00849-w}, -abstract={We consider the problem of recommending relevant content to users of an internet platform in the form of lists of items, called slates. We introduce a variational Bayesian Recurrent Neural Net recommender system that acts on time series of interactions between the internet platform and the user, and which scales to real world industrial situations. The recommender system is tested both online on real users, and on an offline dataset collected from a Norwegian web-based marketplace, FINN.no, that is made public for research. This is one of the first publicly available datasets which includes all the slates that are presented to users as well as which items (if any) in the slates were clicked on. Such a data set allows us to move beyond the common assumption that implicitly assumes that users are considering all possible items at each interaction. Instead we build our likelihood using the items that are actually in the slate, and evaluate the strengths and weaknesses of both approaches theoretically and in experiments. We also introduce a hierarchical prior for the item parameters based on group memberships. Both item parameters and user preferences are learned probabilistically. Furthermore, we combine our model with bandit strategies to ensure learning, and introduce ‘in-slate Thompson sampling’ which makes use of the slates to maximise explorative opportunities. We show experimentally that explorative recommender strategies perform on par or above their greedy counterparts. Even without making use of exploration to learn more effectively, click rates increase simply because of improved diversity in the recommended slates.} -} - -@Inproceedings{Arthurs, -author={Arthurs, Noah and Stenhaug, Ben and Karayev, Sergey and Piech, Chris}, -booktitle={International Conference on Educational Data Mining (EDM)}, -title={Grades Are Not Normal: Improving Exam Score Models Using the Logit-Normal Distribution}, -year={2019}, -type={article}, -language={English}, -volume={}, -number={}, -pages={6}, -url={https://eric.ed.gov/?id=ED599204} -} - -@article{Bahramian, -title={A Cold Start Context-Aware Recommender System for Tour Planning Using Artificial Neural Network and Case Based Reasoning}, -author={Bahramian, Zahra and Ali Abbaspour, Rahim and Claramunt, Christophe}, -language={English}, -type={article}, -year = {2017}, -issn = {1574-017X}, -doi = {https://doi.org/10.1155/2017/9364903}, -url = {https://doi.org/10.1155/2017/9364903}, -abstract={Nowadays, large amounts of tourism information and services are available over the Web. This makes it difficult for the user to search for some specific information such as selecting a tour in a given city as an ordered set of points of interest. Moreover, the user rarely knows all his needs upfront and his preferences may change during a recommendation process. The user may also have a limited number of initial ratings and most often the recommender system is likely to face the well-known cold start problem. The objective of the research presented in this paper is to introduce a hybrid interactive context-aware tourism recommender system that takes into account user’s feedbacks and additional contextual information. It offers personalized tours to the user based on his preferences thanks to the combination of a case based reasoning framework and an artificial neural network. The proposed method has been tried in the city of Tehran in Iran. The results show that the proposed method outperforms current artificial neural network methods and combinations of case based reasoning with -nearest neighbor methods in terms of user effort, accuracy, and user satisfaction.} -} - -@Thesis{Daubias2011, -author={Sthéphanie Jean-Daubias}, -title={Ingénierie des profils d'apprenants}, -type={diplomathesis}, -language={French}, -institution={Université Claude Bernard Lyon 1}, -year={2011} -} - -@article{Tapalova, -author = {Olga Tapalova and Nadezhda Zhiyenbayeva}, -title ={Artificial Intelligence in Education: AIEd for Personalised Learning Pathways}, -journal = {Electronic Journal of e-Learning}, -volume = {}, -number = {}, -pages = {15}, -year = {2022}, -URL = {https://eric.ed.gov/?q=Artificial+Intelligence+in+Education%3a+AIEd+for+Personalised+Learning+Pathways&id=EJ1373006}, -language={English}, -type={article}, -abstract = {Artificial intelligence is the driving force of change focusing on the needs and demands of the student. The research explores Artificial Intelligence in Education (AIEd) for building personalised learning systems for students. The research investigates and proposes a framework for AIEd: social networking sites and chatbots, expert systems for education, intelligent mentors and agents, machine learning, personalised educational systems and virtual educational environments. These technologies help educators to develop and introduce personalised approaches to master new knowledge and develop professional competencies. The research presents a case study of AIEd implementation in education. The scholars conducted the experiment in educational establishments using artificial intelligence in the curriculum. The scholars surveyed 184 second-year students of the Institute of Pedagogy and Psychology at the Abay Kazakh National Pedagogical University and the Kuban State Technological University to collect the data. The scholars considered the collective group discussions regarding the application of artificial intelligence in education to improve the effectiveness of learning. The research identified key advantages to creating personalised learning pathways such as access to training in 24/7 mode, training in virtual contexts, adaptation of educational content to personal needs of students, real-time and regular feedback, improvements in the educational process and mental stimulations. The proposed education paradigm reflects the increasing role of artificial intelligence in socio-economic life, the social and ethical concerns artificial intelligence may pose to humanity and its role in the digitalisation of education. The current article may be used as a theoretical framework for many educational institutions planning to exploit the capabilities of artificial intelligence in their adaptation to personalized learning.} -} - -@article{Auer, -title = {From monolithic systems to Microservices: An assessment framework}, -journal = {Information and Software Technology}, -volume = {137}, -pages = {106600}, -year = {2021}, -issn = {0950-5849}, -doi = {https://doi.org/10.1016/j.infsof.2021.106600}, -url = {https://www.sciencedirect.com/science/article/pii/S0950584921000793}, -author = {Florian Auer and Valentina Lenarduzzi and Michael Felderer and Davide Taibi}, -keywords = {Microservices, Cloud migration, Software measurement}, -abstract = {Context: -Re-architecting monolithic systems with Microservices-based architecture is a common trend. Various companies are migrating to Microservices for different reasons. However, making such an important decision like re-architecting an entire system must be based on real facts and not only on gut feelings. -Objective: -The goal of this work is to propose an evidence-based decision support framework for companies that need to migrate to Microservices, based on the analysis of a set of characteristics and metrics they should collect before re-architecting their monolithic system. -Method: -We conducted a survey done in the form of interviews with professionals to derive the assessment framework based on Grounded Theory. -Results: -We identified a set consisting of information and metrics that companies can use to decide whether to migrate to Microservices or not. The proposed assessment framework, based on the aforementioned metrics, could be useful for companies if they need to migrate to Microservices and do not want to run the risk of failing to consider some important information.} -} - -@Article{jmse10040464, -AUTHOR = {Zuluaga, Carlos A. and Aristizábal, Luis M. and Rúa, Santiago and Franco, Diego A. and Osorio, Dorie A. and Vásquez, Rafael E.}, -TITLE = {Development of a Modular Software Architecture for Underwater Vehicles Using Systems Engineering}, -JOURNAL = {Journal of Marine Science and Engineering}, -VOLUME = {10}, -YEAR = {2022}, -NUMBER = {4}, -ARTICLE-NUMBER = {464}, -URL = {https://www.mdpi.com/2077-1312/10/4/464}, -ISSN = {2077-1312}, -ABSTRACT = {This paper addresses the development of a modular software architecture for the design/construction/operation of a remotely operated vehicle (ROV), based on systems engineering. First, systems engineering and the Vee model are presented with the objective of defining the interactions of the stakeholders with the software architecture development team and establishing the baselines that must be met in each development phase. In the development stage, the definition of the architecture and its connection with the hardware is presented, taking into account the use of the actor model, which represents the high-level software architecture used to solve concurrency problems. Subsequently, the structure of the classes is defined both at high and low levels in the instruments using the object-oriented programming paradigm. Finally, unit tests are developed for each component in the software architecture, quality assessment tests are implemented for system functions fulfillment, and a field sea trial for testing different modules of the vehicle is described. This approach is well suited for the development of complex systems such as marine vehicles and those systems which require scalability and modularity to add functionalities.}, -DOI = {10.3390/jmse10040464} -} - -@article{doi:10.1177/1754337116651013, -author = {Julien Henriet and Lang Christophe and Philippe Laurent}, -title ={Artificial Intelligence-Virtual Trainer: An educative system based on artificial intelligence and designed to produce varied and consistent training lessons}, -journal = {Proceedings of the Institution of Mechanical Engineers, Part P: Journal of Sports Engineering and Technology}, -volume = {231}, -number = {2}, -pages = {110-124}, -year = {2017}, -doi = {10.1177/1754337116651013}, -URL = {https://doi.org/10.1177/1754337116651013}, -eprint = {https://doi.org/10.1177/1754337116651013}, -abstract = { AI-Virtual Trainer is an educative system using Artificial Intelligence to propose varied lessons to trainers. The agents of this multi-agent system apply case-based reasoning to build solutions by analogy. However, as required by the field, Artificial Intelligence-Virtual Trainer never proposes the same lesson twice, whereas the same objective may be set many times consecutively. The adaptation process of Artificial Intelligence-Virtual Trainer delivers an ordered set of exercises adapted to the objectives and sub-objectives chosen by trainers. This process has been enriched by including the notion of distance between exercises: the proposed tasks are not only appropriate but are hierarchically ordered. With this new version of the system, students are guided towards their objectives via an underlying theme. Finally, the agents responsible for the different parts of lessons collaborate with each other according to a dedicated protocol and decision-making policy since no exercise must appear more than once in the same lesson. The results prove that Artificial Intelligence-Virtual Trainer, however perfectible, meets the requirements of this field. } -} - -@InProceedings{10.1007/978-3-030-01081-2_9, -author="Henriet, Julien -and Greffier, Fran{\c{c}}oise", -editor="Cox, Michael T. -and Funk, Peter -and Begum, Shahina", -title="AI-VT: An Example of CBR that Generates a Variety of Solutions to the Same Problem", -booktitle="Case-Based Reasoning Research and Development", -year="2018", -publisher="Springer International Publishing", -address="Cham", -pages="124--139", -abstract="AI-Virtual Trainer (AI-VT) is an intelligent tutoring system based on case-based reasoning. AI-VT has been designed to generate personalised, varied, and consistent training sessions for learners. The AI-VT training sessions propose different exercises in regard to a capacity associated with sub-capacities. For example, in the field of training for algorithms, a capacity could be ``Use a control structure alternative'' and an associated sub-capacity could be ``Write a boolean condition''. AI-VT can elaborate a personalised list of exercises for each learner. One of the main requirements and challenges studied in this work is its ability to propose varied training sessions to the same learner for many weeks, which constitutes the challenge studied in our work. Indeed, if the same set of exercises is proposed time after time to learners, they will stop paying attention and lose motivation. Thus, even if the generation of training sessions is based on analogy and must integrate the repetition of some exercises, it also must introduce some diversity and AI-VT must deal with this diversity. In this paper, we have highlighted the fact that the retaining (or capitalisation) phase of CBR is of the utmost importance for diversity, and we have also highlighted that the equilibrium between repetition and variety depends on the abilities learned. This balance has an important impact on the retaining phase of AI-VT.", -isbn="978-3-030-01081-2" -} - -@article{BAKUROV2021100913, -title = {Genetic programming for stacked generalization}, -journal = {Swarm and Evolutionary Computation}, -volume = {65}, -pages = {100913}, -year = {2021}, -issn = {2210-6502}, -doi = {https://doi.org/10.1016/j.swevo.2021.100913}, -url = {https://www.sciencedirect.com/science/article/pii/S2210650221000742}, -author = {Illya Bakurov and Mauro Castelli and Olivier Gau and Francesco Fontanella and Leonardo Vanneschi}, -keywords = {Genetic Programming, Stacking, Ensemble Learning, Stacked Generalization}, -abstract = {In machine learning, ensemble techniques are widely used to improve the performance of both classification and regression systems. They combine the models generated by different learning algorithms, typically trained on different data subsets or with different parameters, to obtain more accurate models. Ensemble strategies range from simple voting rules to more complex and effective stacked approaches. They are based on adopting a meta-learner, i.e. a further learning algorithm, and are trained on the predictions provided by the single algorithms making up the ensemble. The paper aims at exploiting some of the most recent genetic programming advances in the context of stacked generalization. In particular, we investigate how the evolutionary demes despeciation initialization technique, ϵ-lexicase selection, geometric-semantic operators, and semantic stopping criterion, can be effectively used to improve GP-based systems’ performance for stacked generalization (a.k.a. stacking). The experiments, performed on a broad set of synthetic and real-world regression problems, confirm the effectiveness of the proposed approach.} -} - -@article{Liang, -author={Liang Mang and Chang Tianpeng and An Bingxing and Duan Xinghai and Du Lili and Wang Xiaoqiao and Miao Jian and Xu Lingyang and Gao Xue and Zhang Lupei and Li Junya and Gao Huijiang}, -Title={A Stacking Ensemble Learning Framework for Genomic Prediction}, -Journal={Frontiers in Genetics}, -year={2021}, -doi ={10.3389/fgene.2021.600040}, -PMID={33747037}, -PMCID={PMC7969712} -} - -@Article{cmc.2023.033417, -AUTHOR = {Jeonghoon Choi and Dongjun Suh and Marc-Oliver Otto}, -TITLE = {Boosted Stacking Ensemble Machine Learning Method for Wafer Map Pattern Classification}, -JOURNAL = {Computers, Materials \& Continua}, -VOLUME = {74}, -YEAR = {2023}, -NUMBER = {2}, -PAGES = {2945--2966}, -URL = {http://www.techscience.com/cmc/v74n2/50296}, -ISSN = {1546-2226}, -ABSTRACT = {Recently, machine learning-based technologies have been developed to automate the classification of wafer map defect patterns during semiconductor manufacturing. The existing approaches used in the wafer map pattern classification include directly learning the image through a convolution neural network and applying the ensemble method after extracting image features. This study aims to classify wafer map defects more effectively and derive robust algorithms even for datasets with insufficient defect patterns. First, the number of defects during the actual process may be limited. Therefore, insufficient data are generated using convolutional auto-encoder (CAE), and the expanded data are verified using the evaluation technique of structural similarity index measure (SSIM). After extracting handcrafted features, a boosted stacking ensemble model that integrates the four base-level classifiers with the extreme gradient boosting classifier as a meta-level classifier is designed and built for training the model based on the expanded data for final prediction. Since the proposed algorithm shows better performance than those of existing ensemble classifiers even for insufficient defect patterns, the results of this study will contribute to improving the product quality and yield of the actual semiconductor manufacturing process.}, -DOI = {10.32604/cmc.2023.033417} -} - -@ARTICLE{10.3389/fgene.2021.600040, -AUTHOR={Liang, Mang and Chang, Tianpeng and An, Bingxing and Duan, Xinghai and Du, Lili and Wang, Xiaoqiao and Miao, Jian and Xu, Lingyang and Gao, Xue and Zhang, Lupei and Li, Junya and Gao, Huijiang}, -TITLE={A Stacking Ensemble Learning Framework for Genomic Prediction}, -JOURNAL={Frontiers in Genetics}, -VOLUME={12}, -YEAR={2021}, -URL={https://www.frontiersin.org/articles/10.3389/fgene.2021.600040}, -DOI={10.3389/fgene.2021.600040}, -ISSN={1664-8021}, -ABSTRACT={Machine learning (ML) is perhaps the most useful tool for the interpretation of large genomic datasets. However, the performance of a single machine learning method in genomic selection (GS) is currently unsatisfactory. To improve the genomic predictions, we constructed a stacking ensemble learning framework (SELF), integrating three machine learning methods, to predict genomic estimated breeding values (GEBVs). The present study evaluated the prediction ability of SELF by analyzing three real datasets, with different genetic architecture; comparing the prediction accuracy of SELF, base learners, genomic best linear unbiased prediction (GBLUP) and BayesB. For each trait, SELF performed better than base learners, which included support vector regression (SVR), kernel ridge regression (KRR) and elastic net (ENET). The prediction accuracy of SELF was, on average, 7.70% higher than GBLUP in three datasets. Except for the milk fat percentage (MFP) traits, of the German Holstein dairy cattle dataset, SELF was more robust than BayesB in all remaining traits. Therefore, we believed that SEFL has the potential to be promoted to estimate GEBVs in other animals and plants.} -} - -@article{DIDDEN2023338, -title = {Decentralized learning multi-agent system for online machine shop scheduling problem}, -journal = {Journal of Manufacturing Systems}, -volume = {67}, -pages = {338-360}, -year = {2023}, -issn = {0278-6125}, -doi = {https://doi.org/10.1016/j.jmsy.2023.02.004}, -url = {https://www.sciencedirect.com/science/article/pii/S0278612523000286}, -author = {Jeroen B.H.C. Didden and Quang-Vinh Dang and Ivo J.B.F. Adan}, -keywords = {Multi-agent system, Decentralized systems, Learning algorithm, Industry 4.0, Smart manufacturing}, -abstract = {Customer profiles have rapidly changed over the past few years, with products being requested with more customization and with lower demand. In addition to the advances in technologies owing to Industry 4.0, manufacturers explore autonomous and smart factories. This paper proposes a decentralized multi-agent system (MAS), including intelligent agents that can respond to their environment autonomously through learning capabilities, to cope with an online machine shop scheduling problem. In the proposed system, agents participate in auctions to receive jobs to process, learn how to bid for jobs correctly, and decide when to start processing a job. The objective is to minimize the mean weighted tardiness of all jobs. In contrast to the existing literature, the proposed MAS is assessed on its learning capabilities, producing novel insights concerning what is relevant for learning, when re-learning is needed, and system response to dynamic events (such as rush jobs, increase in processing time, and machine unavailability). Computational experiments also reveal the outperformance of the proposed MAS to other multi-agent systems by at least 25% and common dispatching rules in mean weighted tardiness, as well as other performance measures.} -} - -@article{REZAEI20221, -title = {A Biased Inferential Naivety learning model for a network of agents}, -journal = {Cognitive Systems Research}, -volume = {76}, -pages = {1-12}, -year = {2022}, -issn = {1389-0417}, -doi = {https://doi.org/10.1016/j.cogsys.2022.07.001}, -url = {https://www.sciencedirect.com/science/article/pii/S1389041722000298}, -author = {Zeinab Rezaei and Saeed Setayeshi and Ebrahim Mahdipour}, -keywords = {Bayesian decision making, Heuristic method, Inferential naivety assumption, Observational learning, Social learning}, -abstract = {We propose a Biased Inferential Naivety social learning model. In this model, a group of agents tries to determine the true state of the world and make the best possible decisions. The agents have limited computational abilities. They receive noisy private signals about the true state and observe the history of their neighbors' decisions. The proposed model is rooted in the Bayesian method but avoids the complexity of fully Bayesian inference. In our model, the role of knowledge obtained from social observations is separated from the knowledge obtained from private observations. Therefore, the Bayesian inferences on social observations are approximated using inferential naivety assumption, while purely Bayesian inferences are made on private observations. The reduction of herd behavior is another innovation of the proposed model. This advantage is achieved by reducing the effect of social observations on agents' beliefs over time. Therefore, all the agents learn the truth, and the correct consensus is achieved effectively. In this model, using two cognitive biases, there is heterogeneity in agents' behaviors. Therefore, the growth of beliefs and the learning speed can be improved in different situations. Several Monte Carlo simulations confirm the features of the proposed model. The conditions under which the proposed model leads to asymptotic learning are proved.} -} - -@article{KAMALI2023110242, -title = {An immune inspired multi-agent system for dynamic multi-objective optimization}, -journal = {Knowledge-Based Systems}, -volume = {262}, -pages = {110242}, -year = {2023}, -issn = {0950-7051}, -doi = {https://doi.org/10.1016/j.knosys.2022.110242}, -url = {https://www.sciencedirect.com/science/article/pii/S0950705122013387}, -author = {Seyed Ruhollah Kamali and Touraj Banirostam and Homayun Motameni and Mohammad Teshnehlab}, -keywords = {Immune inspired multi-agent system, Dynamic multi-objective optimization, Severe and frequent changes}, -abstract = {In this research, an immune inspired multi-agent system (IMAS) is proposed to solve optimization problems in dynamic and multi-objective environments. The proposed IMAS uses artificial immune system metaphors to shape the local behaviors of agents to detect environmental changes, generate Pareto optimal solutions, and react to the dynamics of the problem environment. Apart from that, agents enhance their adaptive capacity in dealing with environmental changes to find the global optimum, with a hierarchical structure without any central control. This study used a combination of diversity-, multi-population- and memory-based approaches to perform better in multi-objective environments with severe and frequent changes. The proposed IMAS is compared with six state-of-the-art algorithms on various benchmark problems. The results indicate its superiority in many of the experiments.} -} - -@article{ZHANG2023110564, -title = {A novel human learning optimization algorithm with Bayesian inference learning}, -journal = {Knowledge-Based Systems}, -volume = {271}, -pages = {110564}, -year = {2023}, -issn = {0950-7051}, -doi = {https://doi.org/10.1016/j.knosys.2023.110564}, -url = {https://www.sciencedirect.com/science/article/pii/S0950705123003143}, -author = {Pinggai Zhang and Ling Wang and Zixiang Fei and Lisheng Wei and Minrui Fei and Muhammad Ilyas Menhas}, -keywords = {Human learning optimization, Meta-heuristic, Bayesian inference, Bayesian inference learning, Individual learning, Social learning}, -abstract = {Humans perform Bayesian inference in a wide variety of tasks, which can help people make selection decisions effectively and therefore enhances learning efficiency and accuracy. Inspired by this fact, this paper presents a novel human learning optimization algorithm with Bayesian inference learning (HLOBIL), in which a Bayesian inference learning operator (BILO) is developed to utilize the inference strategy for enhancing learning efficiency. The in-depth analysis shows that the proposed BILO can efficiently improve the exploitation ability of the algorithm as it can achieve the optimal values and retrieve the optimal information with the accumulated search information. Besides, the exploration ability of HLOBIL is also strengthened by the inborn characteristics of Bayesian inference. The experimental results demonstrate that the developed HLOBIL is superior to previous HLO variants and other state-of-art algorithms with its improved exploitation and exploration abilities.} -} - -@article{HIPOLITO2023103510, -title = {Breaking boundaries: The Bayesian Brain Hypothesis for perception and prediction}, -journal = {Consciousness and Cognition}, -volume = {111}, -pages = {103510}, -year = {2023}, -issn = {1053-8100}, -doi = {https://doi.org/10.1016/j.concog.2023.103510}, -url = {https://www.sciencedirect.com/science/article/pii/S1053810023000478}, -author = {Inês Hipólito and Michael Kirchhoff}, -keywords = {Bayesian Brain Hypothesis, Modularity of the Mind, Cognitive processes, Informational boundaries}, -abstract = {This special issue aims to provide a comprehensive overview of the current state of the Bayesian Brain Hypothesis and its standing across neuroscience, cognitive science and the philosophy of cognitive science. By gathering cutting-edge research from leading experts, this issue seeks to showcase the latest advancements in our understanding of the Bayesian brain, as well as its potential implications for future research in perception, cognition, and motor control. A special focus to achieve this aim is adopted in this special issue, as it seeks to explore the relation between two seemingly incompatible frameworks for the understanding of cognitive structure and function: the Bayesian Brain Hypothesis and the Modularity Theory of the Mind. In assessing the compatibility between these theories, the contributors to this special issue open up new pathways of thinking and advance our understanding of cognitive processes.} -} - -@article{LI2023424, -title = {Multi-agent evolution reinforcement learning method for machining parameters optimization based on bootstrap aggregating graph attention network simulated environment}, -journal = {Journal of Manufacturing Systems}, -volume = {67}, -pages = {424-438}, -year = {2023}, -issn = {0278-6125}, -doi = {https://doi.org/10.1016/j.jmsy.2023.02.015}, -url = {https://www.sciencedirect.com/science/article/pii/S0278612523000390}, -author = {Weiye Li and Songping He and Xinyong Mao and Bin Li and Chaochao Qiu and Jinwen Yu and Fangyu Peng and Xin Tan}, -keywords = {Surface roughness, Cutting efficiency, Machining parameters optimization, Graph attention network, Multi-agent reinforcement learning, Evolutionary learning}, -abstract = {Improving machining quality and production efficiency is the focus of the manufacturing industry. How to obtain efficient machining parameters under multiple constraints such as machining quality is a severe challenge for manufacturing industry. In this paper, a multi-agent evolutionary reinforcement learning method (MAERL) is proposed to optimize the machining parameters for high quality and high efficiency machining by combining the graph neural network and reinforcement learning. Firstly, a bootstrap aggregating graph attention network (Bagging-GAT) based roughness estimation method for machined surface is proposed, which combines the structural knowledge between machining parameters and vibration features. Secondly, a mathematical model of machining parameters optimization problem is established, which is formalized into Markov decision process (MDP), and a multi-agent reinforcement learning method is proposed to solve the MDP problem, and evolutionary learning is introduced to improve the stability of multi-agent training. Finally, a series of experiments were carried out on the commutator production line, and the results show that the proposed Bagging-GAT-based method can improve the prediction effect by about 25% in the case of small samples, and the MAERL-based optimization method can better deal with the coupling problem of reward function in the optimization process. Compared with the classical optimization method, the optimization effect is improved by 13% and a lot of optimization time is saved.} -} - -@inproceedings{10.1145/3290605.3300912, -author = {Kim, Yea-Seul and Walls, Logan A. and Krafft, Peter and Hullman, Jessica}, -title = {A Bayesian Cognition Approach to Improve Data Visualization}, -year = {2019}, -isbn = {9781450359702}, -publisher = {Association for Computing Machinery}, -address = {New York, NY, USA}, -url = {https://doi.org/10.1145/3290605.3300912}, -doi = {10.1145/3290605.3300912}, -abstract = {People naturally bring their prior beliefs to bear on how they interpret the new information, yet few formal models exist for accounting for the influence of users' prior beliefs in interactions with data presentations like visualizations. We demonstrate a Bayesian cognitive model for understanding how people interpret visualizations in light of prior beliefs and show how this model provides a guide for improving visualization evaluation. In a first study, we show how applying a Bayesian cognition model to a simple visualization scenario indicates that people's judgments are consistent with a hypothesis that they are doing approximate Bayesian inference. In a second study, we evaluate how sensitive our observations of Bayesian behavior are to different techniques for eliciting people subjective distributions, and to different datasets. We find that people don't behave consistently with Bayesian predictions for large sample size datasets, and this difference cannot be explained by elicitation technique. In a final study, we show how normative Bayesian inference can be used as an evaluation framework for visualizations, including of uncertainty.}, -booktitle = {Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems}, -pages = {1–14}, -numpages = {14}, -keywords = {bayesian cognition, uncertainty elicitation, visualization}, -location = {Glasgow, Scotland Uk}, -series = {CHI '19} -} - -@article{DYER2024104827, -title = {Black-box Bayesian inference for agent-based models}, -journal = {Journal of Economic Dynamics and Control}, -volume = {161}, -pages = {104827}, -year = {2024}, -issn = {0165-1889}, -doi = {https://doi.org/10.1016/j.jedc.2024.104827}, -url = {https://www.sciencedirect.com/science/article/pii/S0165188924000198}, -author = {Joel Dyer and Patrick Cannon and J. Doyne Farmer and Sebastian M. Schmon}, -keywords = {Agent-based models, Bayesian inference, Neural networks, Parameter estimation, Simulation-based inference, Time series}, -abstract = {Simulation models, in particular agent-based models, are gaining popularity in economics and the social sciences. The considerable flexibility they offer, as well as their capacity to reproduce a variety of empirically observed behaviours of complex systems, give them broad appeal, and the increasing availability of cheap computing power has made their use feasible. Yet a widespread adoption in real-world modelling and decision-making scenarios has been hindered by the difficulty of performing parameter estimation for such models. In general, simulation models lack a tractable likelihood function, which precludes a straightforward application of standard statistical inference techniques. A number of recent works have sought to address this problem through the application of likelihood-free inference techniques, in which parameter estimates are determined by performing some form of comparison between the observed data and simulation output. However, these approaches are (a) founded on restrictive assumptions, and/or (b) typically require many hundreds of thousands of simulations. These qualities make them unsuitable for large-scale simulations in economics and the social sciences, and can cast doubt on the validity of these inference methods in such scenarios. In this paper, we investigate the efficacy of two classes of simulation-efficient black-box approximate Bayesian inference methods that have recently drawn significant attention within the probabilistic machine learning community: neural posterior estimation and neural density ratio estimation. We present a number of benchmarking experiments in which we demonstrate that neural network-based black-box methods provide state of the art parameter inference for economic simulation models, and crucially are compatible with generic multivariate or even non-Euclidean time-series data. In addition, we suggest appropriate assessment criteria for use in future benchmarking of approximate Bayesian inference procedures for simulation models in economics and the social sciences.} -} - -@Article{Nikpour2021, -author={Nikpour, Hoda -and Aamodt, Agnar}, -title={Inference and reasoning in a Bayesian knowledge-intensive CBR system}, -journal={Progress in Artificial Intelligence}, -year={2021}, -month={Mar}, -day={01}, -volume={10}, -number={1}, -pages={49-63}, -abstract={This paper presents the inference and reasoning methods in a Bayesian supported knowledge-intensive case-based reasoning (CBR) system called BNCreek. The inference and reasoning process in this system is a combination of three methods. The semantic network inference methods and the CBR method are employed to handle the difficulties of inferencing and reasoning in uncertain domains. The Bayesian network inference methods are employed to make the process more accurate. An experiment from oil well drilling as a complex and uncertain application domain is conducted. The system is evaluated against expert estimations and compared with seven other corresponding systems. The normalized discounted cumulative gain (NDCG) as a rank-based metric, the weighted error (WE), and root-square error (RSE) as the statistical metrics are employed to evaluate different aspects of the system capabilities. The results show the efficiency of the developed inference and reasoning methods.}, -issn={2192-6360}, -doi={10.1007/s13748-020-00223-1}, -url={https://doi.org/10.1007/s13748-020-00223-1} -} - -@article{PRESCOTT2024112577, -title = {Efficient multifidelity likelihood-free Bayesian inference with adaptive computational resource allocation}, -journal = {Journal of Computational Physics}, -volume = {496}, -pages = {112577}, -year = {2024}, -issn = {0021-9991}, -doi = {https://doi.org/10.1016/j.jcp.2023.112577}, -url = {https://www.sciencedirect.com/science/article/pii/S0021999123006721}, -author = {Thomas P. Prescott and David J. Warne and Ruth E. Baker}, -keywords = {Likelihood-free Bayesian inference, Multifidelity approaches}, -abstract = {Likelihood-free Bayesian inference algorithms are popular methods for inferring the parameters of complex stochastic models with intractable likelihoods. These algorithms characteristically rely heavily on repeated model simulations. However, whenever the computational cost of simulation is even moderately expensive, the significant burden incurred by likelihood-free algorithms leaves them infeasible for many practical applications. The multifidelity approach has been introduced in the context of approximate Bayesian computation to reduce the simulation burden of likelihood-free inference without loss of accuracy, by using the information provided by simulating computationally cheap, approximate models in place of the model of interest. In this work we demonstrate that multifidelity techniques can be applied in the general likelihood-free Bayesian inference setting. Analytical results on the optimal allocation of computational resources to simulations at different levels of fidelity are derived, and subsequently implemented practically. We provide an adaptive multifidelity likelihood-free inference algorithm that learns the relationships between models at different fidelities and adapts resource allocation accordingly, and demonstrate that this algorithm produces posterior estimates with near-optimal efficiency.} -} - -@article{RISTIC202030, -title = {A tutorial on uncertainty modeling for machine reasoning}, -journal = {Information Fusion}, -volume = {55}, -pages = {30-44}, -year = {2020}, -issn = {1566-2535}, -doi = {https://doi.org/10.1016/j.inffus.2019.08.001}, -url = {https://www.sciencedirect.com/science/article/pii/S1566253519301976}, -author = {Branko Ristic and Christopher Gilliam and Marion Byrne and Alessio Benavoli}, -keywords = {Information fusion, Uncertainty, Imprecision, Model based classification, Bayesian, Random sets, Belief function theory, Possibility functions, Imprecise probability}, -abstract = {Increasingly we rely on machine intelligence for reasoning and decision making under uncertainty. This tutorial reviews the prevalent methods for model-based autonomous decision making based on observations and prior knowledge, primarily in the context of classification. Both observations and the knowledge-base available for reasoning are treated as being uncertain. Accordingly, the central themes of this tutorial are quantitative modeling of uncertainty, the rules required to combine such uncertain information, and the task of decision making under uncertainty. The paper covers the main approaches to uncertain knowledge representation and reasoning, in particular, Bayesian probability theory, possibility theory, reasoning based on belief functions and finally imprecise probability theory. The main feature of the tutorial is that it illustrates various approaches with several testing scenarios, and provides MATLAB solutions for them as a supplementary material for an interested reader.} -} - -@article{CICIRELLO2022108619, -title = {Machine learning based optimization for interval uncertainty propagation}, -journal = {Mechanical Systems and Signal Processing}, -volume = {170}, -pages = {108619}, -year = {2022}, -issn = {0888-3270}, -doi = {https://doi.org/10.1016/j.ymssp.2021.108619}, -url = {https://www.sciencedirect.com/science/article/pii/S0888327021009493}, -author = {Alice Cicirello and Filippo Giunta}, -keywords = {Bounded uncertainty, Bayesian optimization, Expensive-to-evaluate deterministic computer models, Gaussian process, Communicating uncertainty}, -abstract = {Two non-intrusive uncertainty propagation approaches are proposed for the performance analysis of engineering systems described by expensive-to-evaluate deterministic computer models with parameters defined as interval variables. These approaches employ a machine learning based optimization strategy, the so-called Bayesian optimization, for evaluating the upper and lower bounds of a generic response variable over the set of possible responses obtained when each interval variable varies independently over its range. The lack of knowledge caused by not evaluating the response function for all the possible combinations of the interval variables is accounted for by developing a probabilistic description of the response variable itself by using a Gaussian Process regression model. An iterative procedure is developed for selecting a small number of simulations to be evaluated for updating this statistical model by using well-established acquisition functions and to assess the response bounds. In both approaches, an initial training dataset is defined. While one approach builds iteratively two distinct training datasets for evaluating separately the upper and lower bounds of the response variable, the other one builds iteratively a single training dataset. Consequently, the two approaches will produce different bound estimates at each iteration. The upper and lower response bounds are expressed as point estimates obtained from the mean function of the posterior distribution. Moreover, a confidence interval on each estimate is provided for effectively communicating to engineers when these estimates are obtained at a combination of the interval variables for which no deterministic simulation has been run. Finally, two metrics are proposed to define conditions for assessing if the predicted bound estimates can be considered satisfactory. The applicability of these two approaches is illustrated with two numerical applications, one focusing on vibration and the other on vibro-acoustics.} -} - -@INPROCEEDINGS{9278071, - author={Petit, Maxime and Dellandrea, Emmanuel and Chen, Liming}, - booktitle={2020 Joint IEEE 10th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)}, - title={Bayesian Optimization for Developmental Robotics with Meta-Learning by Parameters Bounds Reduction}, - year={2020}, - volume={}, - number={}, - pages={1-8}, - keywords={Optimization;Robots;Task analysis;Bayes methods;Visualization;Service robots;Cognition;developmental robotics;long-term memory;meta learning;hyperparmeters automatic optimization;case-based reasoning}, - doi={10.1109/ICDL-EpiRob48136.2020.9278071} -} - -@article{LI2023477, -title = {Hierarchical and partitioned planning strategy for closed-loop devices in low-voltage distribution network based on improved KMeans partition method}, -journal = {Energy Reports}, -volume = {9}, -pages = {477-485}, -year = {2023}, -note = {2022 The 3rd International Conference on Power and Electrical Engineering}, -issn = {2352-4847}, -doi = {https://doi.org/10.1016/j.egyr.2023.05.161}, -url = {https://www.sciencedirect.com/science/article/pii/S2352484723009137}, -author = {Jingqi Li and Junlin Li and Dan Wang and Chengxiong Mao and Zhitao Guan and Zhichao Liu and Miaomiao Du and Yuanzhuo Qi and Lexiang Wang and Wenge Liu and Pengfei Tang}, -keywords = {Closed-loop device, Distribution network partition, Device planning, Hierarchical planning, Improved KMeans partition method}, -abstract = {To improve the reliability of power supply, this paper proposes a hierarchical and partitioned planning strategy for closed-loop devices in low-voltage distribution network. Based on the geographic location and load situation of the distribution network area, an improved KMeans partition method is used to partition the area in the upper layer. In the lower layer, an intelligent algorithm is adopted to decide the numbers and placement locations of mobile low-voltage contact boxes and mobile seamless closed-loop load transfer devices in each partition with the goal of the highest closed-loop safety, the greatest improvement in annual power outage amount and the lowest cost. Finally, the feasibility and effectiveness of the proposed strategy are proved by an example.} -} - -@article{SAXENA2024100838, -title = {Hybrid KNN-SVM machine learning approach for solar power forecasting}, -journal = {Environmental Challenges}, -volume = {14}, -pages = {100838}, -year = {2024}, -issn = {2667-0100}, -doi = {https://doi.org/10.1016/j.envc.2024.100838}, -url = {https://www.sciencedirect.com/science/article/pii/S2667010024000040}, -author = {Nishant Saxena and Rahul Kumar and Yarrapragada K S S Rao and Dilbag Singh Mondloe and Nishikant Kishor Dhapekar and Abhishek Sharma and Anil Singh Yadav}, -keywords = {Solar power forecasting, Hybrid model, KNN, Optimization, Solar energy, SVM}, -abstract = {Predictions about solar power will have a significant impact on large-scale renewable energy plants. Photovoltaic (PV) power generation forecasting is particularly sensitive to measuring the uncertainty in weather conditions. Although several conventional techniques like long short-term memory (LSTM), support vector machine (SVM), etc. are available, but due to some restrictions, their application is limited. To enhance the precision of forecasting solar power from solar farms, a hybrid machine learning model that includes blends of the K-Nearest Neighbor (KNN) machine learning technique with the SVM to increase reliability for power system operators is proposed in this investigation. The conventional LSTM technique is also implemented to compare the performance of the proposed hybrid technique. The suggested hybrid model is improved by the use of structural diversity and data diversity in KNN and SVM, respectively. For the solar power predictions, the suggested method was tested on the Jodhpur real-time series dataset obtained from the data centers of weather stations using Meteonorm. The data set includes metrics such as Hourly Average Temperature (HAT), Hourly Total Sunlight Duration (HTSD), Hourly Total Global Solar Radiation (HTGSR), and Hourly Total Photovoltaic Energy Generation (HTPEG). The collated data has been segmented into training data, validation data, and testing data. Furthermore, the proposed technique performed better when evaluated on the three performance indices, viz., accuracy, sensitivity, and specificity. Compared with the conventional LSTM technique, the hybrid technique improved the prediction with 98\% accuracy.} -} - -@article{RAKESH2023100898, -title = {Moving object detection using modified GMM based background subtraction}, -journal = {Measurement: Sensors}, -volume = {30}, -pages = {100898}, -year = {2023}, -issn = {2665-9174}, -doi = {https://doi.org/10.1016/j.measen.2023.100898}, -url = {https://www.sciencedirect.com/science/article/pii/S2665917423002349}, -author = {S. Rakesh and Nagaratna P. Hegde and M. {Venu Gopalachari} and D. Jayaram and Bhukya Madhu and Mohd Abdul Hameed and Ramdas Vankdothu and L.K. {Suresh Kumar}}, -keywords = {Background subtraction, Gaussian mixture models, Intelligent video surveillance, Object detection}, -abstract = {Academics have become increasingly interested in creating cutting-edge technologies to enhance Intelligent Video Surveillance (IVS) performance in terms of accuracy, speed, complexity, and deployment. It has been noted that precise object detection is the only way for IVS to function well in higher level applications including event interpretation, tracking, classification, and activity recognition. Through the use of cutting-edge techniques, the current study seeks to improve the performance accuracy of object detection techniques based on Gaussian Mixture Models (GMM). It is achieved by developing crucial phases in the object detecting process. In this study, it is discussed how to model each pixel as a mixture of Gaussians and how to update the model using an online k-means approximation. The adaptive mixture model's Gaussian distributions are then analyzed to identify which ones are more likely to be the product of a background process. Each pixel is categorized according to whether the background model is thought to include the Gaussian distribution that best depicts it.} -} - -@article{JIAO2022540, -title = {Interpretable fuzzy clustering using unsupervised fuzzy decision trees}, -journal = {Information Sciences}, -volume = {611}, -pages = {540-563}, -year = {2022}, -issn = {0020-0255}, -doi = {https://doi.org/10.1016/j.ins.2022.08.077}, -url = {https://www.sciencedirect.com/science/article/pii/S0020025522009872}, -author = {Lianmeng Jiao and Haoyu Yang and Zhun-ga Liu and Quan Pan}, -keywords = {Fuzzy clustering, Interpretable clustering, Unsupervised decision tree, Cluster merging}, -abstract = {In clustering process, fuzzy partition performs better than hard partition when the boundaries between clusters are vague. Whereas, traditional fuzzy clustering algorithms produce less interpretable results, limiting their application in security, privacy, and ethics fields. To that end, this paper proposes an interpretable fuzzy clustering algorithm—fuzzy decision tree-based clustering which combines the flexibility of fuzzy partition with the interpretability of the decision tree. We constructed an unsupervised multi-way fuzzy decision tree to achieve the interpretability of clustering, in which each cluster is determined by one or several paths from the root to leaf nodes. The proposed algorithm comprises three main modules: feature and cutting point-selection, node fuzzy splitting, and cluster merging. The first two modules are repeated to generate an initial unsupervised decision tree, and the final module is designed to combine similar leaf nodes to form the final compact clustering model. Our algorithm optimizes an internal clustering validation metric to automatically determine the number of clusters without their initial positions. The synthetic and benchmark datasets were used to test the performance of the proposed algorithm. Furthermore, we provided two examples demonstrating its interest in solving practical problems.} -} - -@article{ARNAUGONZALEZ2023101516, -title = {A methodological approach to enable natural language interaction in an Intelligent Tutoring System}, -journal = {Computer Speech and Language}, -volume = {81}, -pages = {101516}, -year = {2023}, -issn = {0885-2308}, -doi = {https://doi.org/10.1016/j.csl.2023.101516}, -url = {https://www.sciencedirect.com/science/article/pii/S0885230823000359}, -author = {Pablo Arnau-González and Miguel Arevalillo-Herráez and Romina Albornoz-De Luise and David Arnau}, -keywords = {Intelligent tutoring systems (ITS), Interactive learning environments (ILE), Conversational agents, Rasa, Natural language understanding (NLU), Natural language processing (NLP)}, -abstract = {In this paper, we present and evaluate the recent incorporation of a conversational agent into an Intelligent Tutoring System (ITS), using the open-source machine learning framework Rasa. Once it has been appropriately trained, this tool is capable of identifying the intention of a given text input and extracting the relevant entities related to the message content. We describe both the generation of a realistic training set in Spanish language that enables the creation of the required Natural Language Understanding (NLU) models and the evaluation of the resulting system. For the generation of the training set, we have followed a methodology that can be easily exported to other ITS. The model evaluation shows that the conversational agent can correctly identify the majority of the user intents, reporting an f1-score above 95%, and cooperate with the ITS to produce a consistent dialogue flow that makes interaction more natural.} -} - -@article{MAO20224065, -title = {An Exploratory Approach to Intelligent Quiz Question Recommendation}, -journal = {Procedia Computer Science}, -volume = {207}, -pages = {4065-4074}, -year = {2022}, -note = {Knowledge-Based and Intelligent Information and Engineering Systems: Proceedings of the 26th International Conference KES2022}, -issn = {1877-0509}, -doi = {https://doi.org/10.1016/j.procs.2022.09.469}, -url = {https://www.sciencedirect.com/science/article/pii/S1877050922013631}, -author = {Kejie Mao and Qiwen Dong and Ye Wang and Daocheng Honga}, -keywords = {question recommendation, two-sided recommender systems, reinforcement learning, intelligent tutoring}, -abstract = {With the rapid advancement of ICT, the digital transformation on education is greatly accelerating in various applications. As a particularly prominent application of digital education, quiz question recommendation is playing a vital role in precision teaching, smart tutoring, and personalized learning. However, the looming challenge of quiz question recommender for students is to satisfy the question diversity demands for students ZPD (the zone of proximal development) stage dynamically online. Therefore, we propose to formalize quiz question recommendation with a novel approach of reinforcement learning based two-sided recommender system. We develop a recommendation framework RTR (Reinforcement-Learning based Two-sided Recommender Systems) for taking into account the interests of both sides of the system, learning and adapting to those interests in real time, and resulting in more satisfactory recommended content. This established recommendation framework captures question characters and student dynamic preferences by considering the emergence of both sides of the system, and it yields a better learning experience in the context of practical quiz question generation.} -} - -@article{CLEMENTE2022118171, -title = {A proposal for an adaptive Recommender System based on competences and ontologies}, -journal = {Expert Systems with Applications}, -volume = {208}, -pages = {118171}, -year = {2022}, -issn = {0957-4174}, -doi = {https://doi.org/10.1016/j.eswa.2022.118171}, -url = {https://www.sciencedirect.com/science/article/pii/S0957417422013392}, -author = {Julia Clemente and Héctor Yago and Javier {de Pedro-Carracedo} and Javier Bueno}, -keywords = {Recommender system, , Ontology network, Methodological development, Student modeling}, -abstract = {Context: -Competences represent an interesting pedagogical support in many processes like diagnosis or recommendation. From these, it is possible to infer information about the progress of the student to provide help targeted both, trainers who must make adaptive tutoring decisions for each learner, and students to detect and correct their learning weaknesses. For the correct development of any of these tasks, it is important to have a suitable student model that allows the representation of the most significant information possible about the student. Additionally, it would be very advantageous for this modeling to incorporate mechanisms from which it would be possible to infer more information about the student’s state of knowledge. -Objective: -To facilitate this goal, in this paper a new approach to develop an adaptive competence-based recommender system is proposed. -Method: -We present a methodological development guide as well as a set of ontological and non-ontological resources to develop and adapt the prototype of the proposed recommender system. -Results: -A modular flexible ontology network previously built for this purpose has been extended, which is responsible for recording the instructional design and student information. Furthermore, we describe a case study based on a first aid learning experience to assess the prototype with the proposed methodology. -Conclusions: -We highlight the relevance of flexibility and adaptability in learning modeling and recommendation processes. In order to promote improvement in the personalized learning of students, we present a Recommender System prototype taking advantages of ontologies, with a methodological guide, a broad taxonomy of recommendation criteria and the nature of competences. Future lines of research lines, including a more comprehensive evaluation of the system, will allow us to demonstrate in depth its adaptability according to the characteristics of the student, flexibility and extensibility for its integration in various environments and domains.} -} - -@article{https://doi.org/10.1155/2023/2578286, -author = {Li, Linqing and Wang, Zhifeng}, -title = {Knowledge Graph-Enhanced Intelligent Tutoring System Based on Exercise Representativeness and Informativeness}, -journal = {International Journal of Intelligent Systems}, -volume = {2023}, -number = {1}, -pages = {2578286}, -doi = {https://doi.org/10.1155/2023/2578286}, -url = {https://onlinelibrary.wiley.com/doi/abs/10.1155/2023/2578286}, -eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1155/2023/2578286}, -abstract = {In the realm of online tutoring intelligent systems, e-learners are exposed to a substantial volume of learning content. The extraction and organization of exercises and skills hold significant importance in establishing clear learning objectives and providing appropriate exercise recommendations. Presently, knowledge graph-based recommendation algorithms have garnered considerable attention among researchers. However, these algorithms solely consider knowledge graphs with single relationships and do not effectively model exercise-rich features, such as exercise representativeness and informativeness. Consequently, this paper proposes a framework, namely, the Knowledge Graph Importance-Exercise Representativeness and Informativeness Framework, to address these two issues. The framework consists of four intricate components and a novel cognitive diagnosis model called the Neural Attentive Cognitive Diagnosis model to recommend the proper exercises. These components encompass the informativeness component, exercise representation component, knowledge importance component, and exercise representativeness component. The informativeness component evaluates the informational value of each exercise and identifies the candidate exercise set (EC) that exhibits the highest exercise informativeness. Moreover, the exercise representation component utilizes a graph neural network to process student records. The output of the graph neural network serves as the input for exercise-level attention and skill-level attention, ultimately generating exercise embeddings and skill embeddings. Furthermore, the skill embeddings are employed as input for the knowledge importance component. This component transforms a one-dimensional knowledge graph into a multidimensional one through four class relations and calculates skill importance weights based on novelty and popularity. Subsequently, the exercise representativeness component incorporates exercise weight knowledge coverage to select exercises from the candidate exercise set for the tested exercise set. Lastly, the cognitive diagnosis model leverages exercise representation and skill importance weights to predict student performance on the test set and estimate their knowledge state. To evaluate the effectiveness of our selection strategy, extensive experiments were conducted on two types of publicly available educational datasets. The experimental results demonstrate that our framework can recommend appropriate exercises to students, leading to improved student performance.}, -year = {2023} -} - -@inproceedings{badier:hal-04092828, - TITLE = {{Comprendre les usages et effets d'un syst{\`e}me de recommandations p{\'e}dagogiques en contexte d'apprentissage non-formel}}, - AUTHOR = {Badier, Ana{\"e}lle and Lefort, Mathieu and Lefevre, Marie}, - URL = {https://hal.science/hal-04092828}, - BOOKTITLE = {{EIAH'23}}, - ADDRESS = {Brest, France}, - YEAR = {2023}, - MONTH = Jun, - HAL_ID = {hal-04092828}, - HAL_VERSION = {v1}, -} - -@article{BADRA2023108920, -title = {Case-based prediction – A survey}, -journal = {International Journal of Approximate Reasoning}, -volume = {158}, -pages = {108920}, -year = {2023}, -issn = {0888-613X}, -doi = {https://doi.org/10.1016/j.ijar.2023.108920}, -url = {https://www.sciencedirect.com/science/article/pii/S0888613X23000440}, -author = {Fadi Badra and Marie-Jeanne Lesot}, -keywords = {Case-based prediction, Analogical transfer, Similarity}, -abstract = {This paper clarifies the relation between case-based prediction and analogical transfer. Case-based prediction consists in predicting the outcome associated with a new case directly from its comparison with a set of cases retrieved from a case base, by relying solely on a structured memory and some similarity measures. Analogical transfer is a cognitive process that allows to derive some new information about a target situation by applying a plausible inference principle, according to which if two situations are similar with respect to some criteria, then it is plausible that they are also similar with respect to other criteria. Case-based prediction algorithms are known to apply analogical transfer to make predictions, but the existing approaches are diverse, and developing a unified theory of case-based prediction remains a challenge. In this paper, we show that a common principle underlying case-based prediction methods is that they interpret the plausible inference as a transfer of similarity knowledge from a situation space to an outcome space. Among all potential outcomes, the predicted outcome is the one that optimizes this transfer, i.e., that makes the similarities in the outcome space most compatible with the observed similarities in the situation space. Based on this observation, a systematic analysis of the different theories of case-based prediction is presented, where the approaches are distinguished according to the type of knowledge used to measure the compatibility between the two sets of similarity relations.} -} - - -@Article{jmse11050890 , -AUTHOR = {Louvros, Panagiotis and Stefanidis, Fotios and Boulougouris, Evangelos and Komianos, Alexandros and Vassalos, Dracos}, -TITLE = {Machine Learning and Case-Based Reasoning for Real-Time Onboard Prediction of the Survivability of Ships}, -JOURNAL = {Journal of Marine Science and Engineering}, -VOLUME = {11}, -YEAR = {2023}, -NUMBER = {5}, -ARTICLE-NUMBER = {890}, -URL = {https://www.mdpi.com/2077-1312/11/5/890}, -ISSN = {2077-1312}, -ABSTRACT = {The subject of damaged stability has greatly profited from the development of new tools and techniques in recent history. Specifically, the increased computational power and the probabilistic approach have transformed the subject, increasing accuracy and fidelity, hence allowing for a universal application and the inclusion of the most probable scenarios. Currently, all ships are evaluated for their stability and are expected to survive the dangers they will most likely face. However, further advancements in simulations have made it possible to further increase the fidelity and accuracy of simulated casualties. Multiple time domain and, to a lesser extent, Computational Fluid dynamics (CFD) solutions have been suggested as the next “evolutionary” step for damage stability. However, while those techniques are demonstrably more accurate, the computational power to utilize them for the task of probabilistic evaluation is not there yet. In this paper, the authors present a novel approach that aims to serve as a stopgap measure for introducing the time domain simulations in the existing framework. Specifically, the methodology presented serves the purpose of a fast decision support tool which is able to provide information regarding the ongoing casualty utilizing prior knowledge gained from simulations. This work was needed and developed for the purposes of the EU-funded project SafePASS.}, -DOI = {10.3390/jmse11050890} -} - - -@Article{su14031366, -AUTHOR = {Chun, Se-Hak and Jang, Jae-Won}, -TITLE = {A New Trend Pattern-Matching Method of Interactive Case-Based Reasoning for Stock Price Predictions}, -JOURNAL = {Sustainability}, -VOLUME = {14}, -YEAR = {2022}, -NUMBER = {3}, -ARTICLE-NUMBER = {1366}, -URL = {https://www.mdpi.com/2071-1050/14/3/1366}, -ISSN = {2071-1050}, -ABSTRACT = {In this paper, we suggest a new case-based reasoning method for stock price predictions using the knowledge of traders to select similar past patterns among nearest neighbors obtained from a traditional case-based reasoning machine. Thus, this method overcomes the limitation of conventional case-based reasoning, which does not consider how to retrieve similar neighbors from previous patterns in terms of a graphical pattern. In this paper, we show how the proposed method can be used when traders find similar time series patterns among nearest cases. For this, we suggest an interactive prediction system where traders can select similar patterns with individual knowledge among automatically recommended neighbors by case-based reasoning. In this paper, we demonstrate how traders can use their knowledge to select similar patterns using a graphical interface, serving as an exemplar for the target. These concepts are investigated against the backdrop of a practical application involving the prediction of three individual stock prices, i.e., Zoom, Airbnb, and Twitter, as well as the prediction of the Dow Jones Industrial Average (DJIA). The verification of the prediction results is compared with a random walk model based on the RMSE and Hit ratio. The results show that the proposed technique is more effective than the random walk model but it does not statistically surpass the random walk model.}, -DOI = {10.3390/su14031366} -} - -@Article{fire7040107, -AUTHOR = {Pei, Qiuyan and Jia, Zhichao and Liu, Jia and Wang, Yi and Wang, Junhui and Zhang, Yanqi}, -TITLE = {Prediction of Coal Spontaneous Combustion Hazard Grades Based on Fuzzy Clustered Case-Based Reasoning}, -JOURNAL = {Fire}, -VOLUME = {7}, -YEAR = {2024}, -NUMBER = {4}, -ARTICLE-NUMBER = {107}, -URL = {https://www.mdpi.com/2571-6255/7/4/107}, -ISSN = {2571-6255}, -ABSTRACT = {Accurate prediction of the coal spontaneous combustion hazard grades is of great significance to ensure the safe production of coal mines. However, traditional coal temperature prediction models have low accuracy and do not predict the coal spontaneous combustion hazard grades. In order to accurately predict coal spontaneous combustion hazard grades, a prediction model of coal spontaneous combustion based on principal component analysis (PCA), case-based reasoning (CBR), fuzzy clustering (FM), and the snake optimization (SO) algorithm was proposed in this manuscript. Firstly, based on the change rule of the concentration of signature gases in the process of coal warming, a new method of classifying the risk of spontaneous combustion of coal was established. Secondly, MeanRadius-SMOTE was adopted to balance the data structure. The weights of the prediction indicators were calculated through PCA to enhance the prediction precision of the CBR model. Then, by employing FM in the case base, the computational cost of CBR was reduced and its computational efficiency was improved. The SO algorithm was used to determine the hyperparameters in the PCA-FM-CBR model. In addition, multiple comparative experiments were conducted to verify the superiority of the model proposed in this manuscript. The results indicated that SO-PCA-FM-CBR possesses good prediction performance and also improves computational efficiency. Finally, the authors of this manuscript adopted the Random Balance Designs—Fourier Amplitude Sensitivity Test (RBD-FAST) to explain the output of the model and analyzed the global importance of input variables. The results demonstrated that CO is the most important variable affecting the coal spontaneous combustion hazard grades.}, -DOI = {10.3390/fire7040107} -} - -@Article{Desmarais2012, -author={Desmarais, Michel C. -and Baker, Ryan S. J. d.}, -title={A review of recent advances in learner and skill modeling in intelligent learning environments}, -journal={User Modeling and User-Adapted Interaction}, -year={2012}, -month={Apr}, -day={01}, -volume={22}, -number={1}, -pages={9-38}, -abstract={In recent years, learner models have emerged from the research laboratory and research classrooms into the wider world. Learner models are now embedded in real world applications which can claim to have thousands, or even hundreds of thousands, of users. Probabilistic models for skill assessment are playing a key role in these advanced learning environments. In this paper, we review the learner models that have played the largest roles in the success of these learning environments, and also the latest advances in the modeling and assessment of learner skills. We conclude by discussing related advancements in modeling other key constructs such as learner motivation, emotional and attentional state, meta-cognition and self-regulated learning, group learning, and the recent movement towards open and shared learner models.}, -issn={1573-1391}, -doi={10.1007/s11257-011-9106-8}, -url={https://doi.org/10.1007/s11257-011-9106-8} -} - -@article{Eide, -title={Dynamic slate recommendation with gated recurrent units and Thompson sampling}, -author={Eide, Simen and Leslie, David S. and Frigessi, Arnoldo}, -language={English}, -type={article}, -volume = {36}, -year = {2022}, -issn = {1573-756X}, -doi = {https://doi.org/10.1007/s10618-022-00849-w}, -url = {https://doi.org/10.1007/s10618-022-00849-w}, -abstract={We consider the problem of recommending relevant content to users of an internet platform in the form of lists of items, called slates. We introduce a variational Bayesian Recurrent Neural Net recommender system that acts on time series of interactions between the internet platform and the user, and which scales to real world industrial situations. The recommender system is tested both online on real users, and on an offline dataset collected from a Norwegian web-based marketplace, FINN.no, that is made public for research. This is one of the first publicly available datasets which includes all the slates that are presented to users as well as which items (if any) in the slates were clicked on. Such a data set allows us to move beyond the common assumption that implicitly assumes that users are considering all possible items at each interaction. Instead we build our likelihood using the items that are actually in the slate, and evaluate the strengths and weaknesses of both approaches theoretically and in experiments. We also introduce a hierarchical prior for the item parameters based on group memberships. Both item parameters and user preferences are learned probabilistically. Furthermore, we combine our model with bandit strategies to ensure learning, and introduce ‘in-slate Thompson sampling’ which makes use of the slates to maximise explorative opportunities. We show experimentally that explorative recommender strategies perform on par or above their greedy counterparts. Even without making use of exploration to learn more effectively, click rates increase simply because of improved diversity in the recommended slates.} -} - -@InProceedings{10.1007/978-3-031-09680-8_14, -author={Sablayrolles, Louis -and Lefevre, Marie -and Guin, Nathalie -and Broisin, Julien}, -editor={Crossley, Scott -and Popescu, Elvira}, -title={Design and Evaluation of a Competency-Based Recommendation Process}, -booktitle={Intelligent Tutoring Systems}, -year={2022}, -publisher={Springer International Publishing}, -address={Cham}, -pages={148--160}, -abstract={The purpose of recommending activities to learners is to provide them with resources adapted to their needs, to facilitate the learning process. However, when teachers face a large number of students, it is difficult for them to recommend a personalized list of resources to each learner. In this paper, we are interested in the design of a system that automatically recommends resources to learners using their cognitive profile expressed in terms of competencies, but also according to a specific strategy defined by teachers. Our contributions relate to (1) a competency-based pedagogical strategy allowing to express the teacher's expertise, and (2) a recommendation process based on this strategy. This process has been experimented and assessed with students learning Shell programming in a first-year computer science degree. The first results show that (i) the items selected by our system from the set of possible items were relevant according to the experts; (ii) our system provided recommendations in a reasonable time; (iii) the recommendations were consulted by the learners but lacked usability.}, -isbn={978-3-031-09680-8} -} - -@inproceedings{10.1145/3578337.3605122, -author = {Xu, Shuyuan and Ge, Yingqiang and Li, Yunqi and Fu, Zuohui and Chen, Xu and Zhang, Yongfeng}, -title = {Causal Collaborative Filtering}, -year = {2023}, -isbn = {9798400700736}, -publisher = {Association for Computing Machinery}, -address = {New York, NY, USA}, -url = {https://doi.org/10.1145/3578337.3605122}, -doi = {10.1145/3578337.3605122}, -abstract = {Many of the traditional recommendation algorithms are designed based on the fundamental idea of mining or learning correlative patterns from data to estimate the user-item correlative preference. However, pure correlative learning may lead to Simpson's paradox in predictions, and thus results in sacrificed recommendation performance. Simpson's paradox is a well-known statistical phenomenon, which causes confusions in statistical conclusions and ignoring the paradox may result in inaccurate decisions. Fortunately, causal and counterfactual modeling can help us to think outside of the observational data for user modeling and personalization so as to tackle such issues. In this paper, we propose Causal Collaborative Filtering (CCF) --- a general framework for modeling causality in collaborative filtering and recommendation. We provide a unified causal view of CF and mathematically show that many of the traditional CF algorithms are actually special cases of CCF under simplified causal graphs. We then propose a conditional intervention approach for do-operations so that we can estimate the user-item causal preference based on the observational data. Finally, we further propose a general counterfactual constrained learning framework for estimating the user-item preferences. Experiments are conducted on two types of real-world datasets---traditional and randomized trial data---and results show that our framework can improve the recommendation performance and reduce the Simpson's paradox problem of many CF algorithms.}, -booktitle = {Proceedings of the 2023 ACM SIGIR International Conference on Theory of Information Retrieval}, -pages = {235–245}, -numpages = {11}, -keywords = {recommender systems, counterfactual reasoning, collaborative filtering, causal analysis, Simpson's paradox}, -location = {Taipei, Taiwan}, -series = {ICTIR '23} -} - -@inproceedings{10.1145/3583780.3615048, -author = {Zhu, Zheqing and Van Roy, Benjamin}, -title = {Scalable Neural Contextual Bandit for Recommender Systems}, -year = {2023}, -isbn = {9798400701245}, -publisher = {Association for Computing Machinery}, -address = {New York, NY, USA}, -url = {https://doi.org/10.1145/3583780.3615048}, -doi = {10.1145/3583780.3615048}, -abstract = {High-quality recommender systems ought to deliver both innovative and relevant content through effective and exploratory interactions with users. Yet, supervised learning-based neural networks, which form the backbone of many existing recommender systems, only leverage recognized user interests, falling short when it comes to efficiently uncovering unknown user preferences. While there has been some progress with neural contextual bandit algorithms towards enabling online exploration through neural networks, their onerous computational demands hinder widespread adoption in real-world recommender systems. In this work, we propose a scalable sample-efficient neural contextual bandit algorithm for recommender systems. To do this, we design an epistemic neural network architecture, Epistemic Neural Recommendation (ENR), that enables Thompson sampling at a large scale. In two distinct large-scale experiments with real-world tasks, ENR significantly boosts click-through rates and user ratings by at least 9\% and 6\% respectively compared to state-of-the-art neural contextual bandit algorithms. Furthermore, it achieves equivalent performance with at least 29\% fewer user interactions compared to the best-performing baseline algorithm. Remarkably, while accomplishing these improvements, ENR demands orders of magnitude fewer computational resources than neural contextual bandit baseline algorithms.}, -booktitle = {Proceedings of the 32nd ACM International Conference on Information and Knowledge Management}, -pages = {3636–3646}, -numpages = {11}, -keywords = {contextual bandits, decision making under uncertainty, exploration vs exploitation, recommender systems, reinforcement learning}, -location = {Birmingham, United Kingdom}, -series = {CIKM '23} -} - -@ARTICLE{10494875, - author={Ghoorchian, Saeed and Kortukov, Evgenii and Maghsudi, Setareh}, - journal={IEEE Open Journal of Signal Processing}, - title={Non-Stationary Linear Bandits With Dimensionality Reduction for Large-Scale Recommender Systems}, - year={2024}, - volume={5}, - number={}, - pages={548-558}, - keywords={Vectors;Recommender systems;Decision making;Runtime;Signal processing algorithms;Covariance matrices;Robustness;Decision-making;multi-armed bandit;non-stationary environment;online learning;recommender systems}, - doi={10.1109/OJSP.2024.3386490} -} - -@article{GIANNIKIS2024111752, -title = {Reinforcement learning for addressing the cold-user problem in recommender systems}, -journal = {Knowledge-Based Systems}, -volume = {294}, -pages = {111752}, -year = {2024}, -issn = {0950-7051}, -doi = {https://doi.org/10.1016/j.knosys.2024.111752}, -url = {https://www.sciencedirect.com/science/article/pii/S0950705124003873}, -author = {Stelios Giannikis and Flavius Frasincar and David Boekestijn}, -keywords = {Recommender systems, Reinforcement learning, Active learning, Cold-user problem}, -abstract = {Recommender systems are widely used in webshops because of their ability to provide users with personalized recommendations. However, the cold-user problem (i.e., recommending items to new users) is an important issue many webshops face. With the recent General Data Protection Regulation in Europe, the use of additional user information such as demographics is not possible without the user’s explicit consent. Several techniques have been proposed to solve the cold-user problem. Many of these techniques utilize Active Learning (AL) methods, which let cold users rate items to provide better recommendations for them. In this research, we propose two novel approaches that combine reinforcement learning with AL to elicit the users’ preferences and provide them with personalized recommendations. We compare reinforcement learning approaches that are either AL-based or item-based, where the latter predicts users’ ratings of an item by using their ratings of similar items. Differently than many of the existing approaches, this comparison is made based on implicit user information. Using a large real-world dataset, we show that the item-based strategy is more accurate than the AL-based strategy as well as several existing AL strategies.} -} - -@article{IFTIKHAR2024121541, -title = {A reinforcement learning recommender system using bi-clustering and Markov Decision Process}, -journal = {Expert Systems with Applications}, -volume = {237}, -pages = {121541}, -year = {2024}, -issn = {0957-4174}, -doi = {https://doi.org/10.1016/j.eswa.2023.121541}, -url = {https://www.sciencedirect.com/science/article/pii/S0957417423020432}, -author = {Arta Iftikhar and Mustansar Ali Ghazanfar and Mubbashir Ayub and Saad {Ali Alahmari} and Nadeem Qazi and Julie Wall}, -keywords = {Reinforcement learning, Markov Decision Process, Bi-clustering, Q-learning, Policy}, -abstract = {Collaborative filtering (CF) recommender systems are static in nature and does not adapt well with changing user preferences. User preferences may change after interaction with a system or after buying a product. Conventional CF clustering algorithms only identifies the distribution of patterns and hidden correlations globally. However, the impossibility of discovering local patterns by these algorithms, headed to the popularization of bi-clustering algorithms. Bi-clustering algorithms can analyze all dataset dimensions simultaneously and consequently, discover local patterns that deliver a better understanding of the underlying hidden correlations. In this paper, we modelled the recommendation problem as a sequential decision-making problem using Markov Decision Processes (MDP). To perform state representation for MDP, we first converted user-item votings matrix to a binary matrix. Then we performed bi-clustering on this binary matrix to determine a subset of similar rows and columns. A bi-cluster merging algorithm is designed to merge similar and overlapping bi-clusters. These bi-clusters are then mapped to a squared grid (SG). RL is applied on this SG to determine best policy to give recommendation to users. Start state is determined using Improved Triangle Similarity (ITR similarity measure. Reward function is computed as grid state overlapping in terms of users and items in current and prospective next state. A thorough comparative analysis was conducted, encompassing a diverse array of methodologies, including RL-based, pure Collaborative Filtering (CF), and clustering methods. The results demonstrate that our proposed method outperforms its competitors in terms of precision, recall, and optimal policy learning.} -} - -@article{Soto2, -author={Soto-Forero, Daniel and Ackermann, Simha and Betbeder, Marie-Laure and Henriet, Julien}, -title={Automatic Real-Time Adaptation of Training Session Difficulty Using Rules and Reinforcement Learning in the AI-VT ITS}, -journal = {International Journal of Modern Education and Computer Science(IJMECS)}, -volume = {16}, -pages = {56-71}, -year = {2024}, -issn = {2075-0161}, -doi = { https://doi.org/10.5815/ijmecs.2024.03.05}, -url = {https://www.mecs-press.org/ijmecs/ijmecs-v16-n3/v16n3-5.html}, -keywords={Real Time Adaptation, Intelligent Training System, Thompson Sampling, Case-Based Reasoning, Automatic Adaptation}, -abstract={Some of the most common and typical issues in the field of intelligent tutoring systems (ITS) are (i) the correct identification of learners’ difficulties in the learning process, (ii) the adaptation of content or presentation of the system according to the difficulties encountered, and (iii) the ability to adapt without initial data (cold-start). In some cases, the system tolerates modifications after the realization and assessment of competences. Other systems require complicated real-time adaptation since only a limited number of data can be captured. In that case, it must be analyzed properly and with a certain precision in order to obtain the appropriate adaptations. Generally, for the adaptation step, the ITS gathers common learners together and adapts their training similarly. Another type of adaptation is more personalized, but requires acquired or estimated information about each learner (previous grades, probability of success, etc.). Some of these parameters may be difficult to obtain, and others are imprecise and can lead to misleading adaptations. The adaptation using machine learning requires prior training with a lot of data. This article presents a model for the real time automatic adaptation of a predetermined session inside an ITS called AI-VT. This adaptation process is part of a case-based reasoning global model. The characteristics of the model proposed in this paper (i) require a limited number of data in order to generate a personalized adaptation, (ii) do not require training, (iii) are based on the correlation to complexity levels, and (iv) are able to adapt even at the cold-start stage. The proposed model is presented with two different configurations, deterministic and stochastic. The model has been tested with a database of 1000 learners, corresponding to different knowledge levels in three different scenarios. The results show the dynamic adaptation of the proposed model in both versions, with the adaptations obtained helping the system to evolve more rapidly and identify learner weaknesses in the different levels of complexity as well as the generation of pertinent recommendations in specific cases for each learner capacity.} -} - -@InProceedings{10.1007/978-3-031-63646-2_11 , -author={Soto-Forero, Daniel and Betbeder, Marie-Laure and Henriet, Julien}, -editor={Recio-Garcia, Juan A. and Orozco-del-Castillo, Mauricio G. and Bridge, Derek}, -title={Ensemble Stacking Case-Based Reasoning for Regression}, -booktitle={Case-Based Reasoning Research and Development}, -year={2024}, -publisher={Springer Nature Switzerland}, -address={Cham}, -pages={159--174}, -abstract={This paper presents a case-based reasoning algorithm with a two-stage iterative double stacking to find approximate solutions to one and multidimensional regression problems. This approach does not require training, so it can work with dynamic data at run time. The solutions are generated using stochastic algorithms in order to allow exploration of the solution space. The evaluation is performed by transforming the regression problem into an optimization problem with an associated objective function. The algorithm has been tested in comparison with nine classical regression algorithms on ten different regression databases extracted from the UCI site. The results show that the proposed algorithm generates solutions in most cases quite close to the real solutions. According to the RMSE, the proposed algorithm globally among the four best algorithms, according to MAE, to the fourth best algorithms of the ten evaluated, suggesting that the results are reasonably good.}, -isbn={978-3-031-63646-2} -} - -@article{ZHANG2018189, -title = {A three learning states Bayesian knowledge tracing model}, -journal = {Knowledge-Based Systems}, -volume = {148}, -pages = {189-201}, -year = {2018}, -issn = {0950-7051}, -doi = {https://doi.org/10.1016/j.knosys.2018.03.001}, -url = {https://www.sciencedirect.com/science/article/pii/S0950705118301199}, -author = {Kai Zhang and Yiyu Yao}, -keywords = {Bayesian knowledge tracing, Three-way decisions}, -abstract = {This paper proposes a Bayesian knowledge tracing model with three learning states by extending the original two learning states. We divide a learning process into three sections by using an evaluation function for three-way decisions. Advantages of such a trisection over traditional bisection are demonstrated by comparative experiments. We develop a three learning states model based on the trisection of the learning process. We apply the model to a series of comparative experiments with the original model. Qualitative and quantitative analyses of the experimental results indicate the superior performance of the proposed model over the original model in terms of prediction accuracies and related statistical measures.} -} - -@article{Li_2024, -doi = {10.3847/1538-4357/ad3215}, -url = {https://dx.doi.org/10.3847/1538-4357/ad3215}, -year = {2024}, -month = {apr}, -publisher = {The American Astronomical Society}, -volume = {965}, -number = {2}, -pages = {125}, -author = {Zhigang Li and Zhejie Ding and Yu Yu and Pengjie Zhang}, -title = {The Kullback–Leibler Divergence and the Convergence Rate of Fast Covariance Matrix Estimators in Galaxy Clustering Analysis}, -journal = {The Astrophysical Journal}, -abstract = {We present a method to quantify the convergence rate of the fast estimators of the covariance matrices in the large-scale structure analysis. Our method is based on the Kullback–Leibler (KL) divergence, which describes the relative entropy of two probability distributions. As a case study, we analyze the delete-d jackknife estimator for the covariance matrix of the galaxy correlation function. We introduce the information factor or the normalized KL divergence with the help of a set of baseline covariance matrices to diagnose the information contained in the jackknife covariance matrix. Using a set of quick particle mesh mock catalogs designed for the Baryon Oscillation Spectroscopic Survey DR11 CMASS galaxy survey, we find that the jackknife resampling method succeeds in recovering the covariance matrix with 10 times fewer simulation mocks than that of the baseline method at small scales (s ≤ 40 h −1 Mpc). However, the ability to reduce the number of mock catalogs is degraded at larger scales due to the increasing bias on the jackknife covariance matrix. Note that the analysis in this paper can be applied to any fast estimator of the covariance matrix for galaxy clustering measurements.} -} - -@Article{Kim2024, -author={Kim, Wonjik}, -title={A Random Focusing Method with Jensen--Shannon Divergence for Improving Deep Neural Network Performance Ensuring Architecture Consistency}, -journal={Neural Processing Letters}, -year={2024}, -month={Jun}, -day={17}, -volume={56}, -number={4}, -pages={199}, -abstract={Multiple hidden layers in deep neural networks perform non-linear transformations, enabling the extraction of meaningful features and the identification of relationships between input and output data. However, the gap between the training and real-world data can result in network overfitting, prompting the exploration of various preventive methods. The regularization technique called 'dropout' is widely used for deep learning models to improve the training of robust and generalized features. During the training phase with dropout, neurons in a particular layer are randomly selected to be ignored for each input. This random exclusion of neurons encourages the network to depend on different subsets of neurons at different times, fostering robustness and reducing sensitivity to specific neurons. This study introduces a novel approach called random focusing, departing from complete neuron exclusion in dropout. The proposed random focusing selectively highlights random neurons during training, aiming for a smoother transition between training and inference phases while keeping network architecture consistent. This study also incorporates Jensen--Shannon Divergence to enhance the stability and efficacy of the random focusing method. Experimental validation across tasks like image classification and semantic segmentation demonstrates the adaptability of the proposed methods across different network architectures, including convolutional neural networks and transformers.}, -issn={1573-773X}, -doi={10.1007/s11063-024-11668-z}, -url={https://doi.org/10.1007/s11063-024-11668-z} -} - -@InProceedings{pmlr-v238-ou24a, - title = {Thompson Sampling Itself is Differentially Private}, - author = {Ou, Tingting and Cummings, Rachel and Avella Medina, Marco}, - booktitle = {Proceedings of The 27th International Conference on Artificial Intelligence and Statistics}, - pages = {1576--1584}, - year = {2024}, - editor = {Dasgupta, Sanjoy and Mandt, Stephan and Li, Yingzhen}, - volume = {238}, - series = {Proceedings of Machine Learning Research}, - month = {02--04 May}, - publisher = {PMLR}, - pdf = {https://proceedings.mlr.press/v238/ou24a/ou24a.pdf}, - url = {https://proceedings.mlr.press/v238/ou24a.html}, - abstract = {In this work we first show that the classical Thompson sampling algorithm for multi-arm bandits is differentially private as-is, without any modification. We provide per-round privacy guarantees as a function of problem parameters and show composition over $T$ rounds; since the algorithm is unchanged, existing $O(\sqrt{NT\log N})$ regret bounds still hold and there is no loss in performance due to privacy. We then show that simple modifications – such as pre-pulling all arms a fixed number of times, increasing the sampling variance – can provide tighter privacy guarantees. We again provide privacy guarantees that now depend on the new parameters introduced in the modification, which allows the analyst to tune the privacy guarantee as desired. We also provide a novel regret analysis for this new algorithm, and show how the new parameters also impact expected regret. Finally, we empirically validate and illustrate our theoretical findings in two parameter regimes and demonstrate that tuning the new parameters substantially improve the privacy-regret tradeoff.} -} - -@Article{math12111758, -AUTHOR = {Uguina, Antonio R. and Gomez, Juan F. and Panadero, Javier and Martínez-Gavara, Anna and Juan, Angel A.}, -TITLE = {A Learnheuristic Algorithm Based on Thompson Sampling for the Heterogeneous and Dynamic Team Orienteering Problem}, -JOURNAL = {Mathematics}, -VOLUME = {12}, -YEAR = {2024}, -NUMBER = {11}, -ARTICLE-NUMBER = {1758}, -URL = {https://www.mdpi.com/2227-7390/12/11/1758}, -ISSN = {2227-7390}, -ABSTRACT = {The team orienteering problem (TOP) is a well-studied optimization challenge in the field of Operations Research, where multiple vehicles aim to maximize the total collected rewards within a given time limit by visiting a subset of nodes in a network. With the goal of including dynamic and uncertain conditions inherent in real-world transportation scenarios, we introduce a novel dynamic variant of the TOP that considers real-time changes in environmental conditions affecting reward acquisition at each node. Specifically, we model the dynamic nature of environmental factors—such as traffic congestion, weather conditions, and battery level of each vehicle—to reflect their impact on the probability of obtaining the reward when visiting each type of node in a heterogeneous network. To address this problem, a learnheuristic optimization framework is proposed. It combines a metaheuristic algorithm with Thompson sampling to make informed decisions in dynamic environments. Furthermore, we conduct empirical experiments to assess the impact of varying reward probabilities on resource allocation and route planning within the context of this dynamic TOP, where nodes might offer a different reward behavior depending upon the environmental conditions. Our numerical results indicate that the proposed learnheuristic algorithm outperforms static approaches, achieving up to 25% better performance in highly dynamic scenarios. Our findings highlight the effectiveness of our approach in adapting to dynamic conditions and optimizing decision-making processes in transportation systems.}, -DOI = {10.3390/math12111758} -} - -@inproceedings{NEURIPS2023_9d8cf124, - author = {Abel, David and Barreto, Andre and Van Roy, Benjamin and Precup, Doina and van Hasselt, Hado P and Singh, Satinder}, - booktitle = {Advances in Neural Information Processing Systems}, - editor = {A. Oh and T. Naumann and A. Globerson and K. Saenko and M. Hardt and S. Levine}, - pages = {50377--50407}, - publisher = {Curran Associates, Inc.}, - title = {A Definition of Continual Reinforcement Learning}, - url = {https://proceedings.neurips.cc/paper_files/paper/2023/file/9d8cf1247786d6dfeefeeb53b8b5f6d7-Paper-Conference.pdf}, - volume = {36}, - year = {2023} -} - -@article{NGUYEN2024111566, -title = {Dynamic metaheuristic selection via Thompson Sampling for online optimization}, -journal = {Applied Soft Computing}, -volume = {158}, -pages = {111566}, -year = {2024}, -issn = {1568-4946}, -doi = {https://doi.org/10.1016/j.asoc.2024.111566}, -url = {https://www.sciencedirect.com/science/article/pii/S1568494624003405}, -author = {Alain Nguyen}, -keywords = {Selection hyper-heuristic, Multi-armed-bandit, Thompson Sampling, Online optimization}, -abstract = {It is acknowledged that no single heuristic can outperform all the others in every optimization problem. This has given rise to hyper-heuristic methods for providing solutions to a wider range of problems. In this work, a set of five non-competing low-level heuristics is proposed in a hyper-heuristic framework. The multi-armed bandit problem analogy is efficiently leveraged and Thompson Sampling is used to actively select the best heuristic for online optimization. The proposed method is compared against ten population-based metaheuristic algorithms on the well-known CEC’05 optimizing benchmark consisting of 23 functions of various landscapes. The results show that the proposed algorithm is the only one able to find the global minimum of all functions with remarkable consistency.} -} - -@Article{Malladi2024, -author={Malladi, Rama K.}, -title={Application of Supervised Machine Learning Techniques to Forecast the COVID-19 U.S. Recession and Stock Market Crash}, -journal={Computational Economics}, -year={2024}, -month={Mar}, -day={01}, -volume={63}, -number={3}, -pages={1021-1045}, -abstract={Machine learning (ML), a transformational technology, has been successfully applied to forecasting events down the road. This paper demonstrates that supervised ML techniques can be used in recession and stock market crash (more than 20{\%} drawdown) forecasting. After learning from strictly past monthly data, ML algorithms detected the Covid-19 recession by December 2019, six months before the official NBER announcement. Moreover, ML algorithms foresaw the March 2020 S{\&}P500 crash two months before it happened. The current labor market and housing are harbingers of a future U.S. recession (in 3 months). Financial factors have a bigger role to play in stock market crashes than economic factors. The labor market appears as a top-two feature in predicting both recessions and crashes. ML algorithms detect that the U.S. exited recession before December 2020, even though the official NBER announcement has not yet been made. They also do not anticipate a U.S. stock market crash before March 2021. ML methods have three times higher false discovery rates of recessions compared to crashes.}, -issn={1572-9974}, -doi={10.1007/s10614-022-10333-8}, -url={https://doi.org/10.1007/s10614-022-10333-8} -} - -@INPROCEEDINGS{10493943, - author={Raaa Subha and Naaa Gayathri and Saaa Sasireka and Raaa Sathiyabanu and Baaa Santhiyaa and Baaa Varshini}, - booktitle={2024 5th International Conference on Mobile Computing and Sustainable Informatics (ICMCSI)}, - title={Intelligent Tutoring Systems using Long Short-Term Memory Networks and Bayesian Knowledge Tracing}, - year={2024}, - volume={0}, - number={0}, - pages={24-29}, - keywords={Knowledge engineering;Filtering;Estimation;Transforms;Real-time systems;Bayes methods;Problem-solving;Intelligent Tutoring System (ITS);Long Short-Term Memory (LSTM);Bayesian Knowledge Tracing (BKT);Reinforcement Learning}, - doi={10.1109/ICMCSI61536.2024.00010} -} - -@article{https://doi.org/10.1155/2024/4067721, -author = {Ahmed, Esmael}, -title = {Student Performance Prediction Using Machine Learning Algorithms}, -journal = {Applied Computational Intelligence and Soft Computing}, -volume = {2024}, -number = {1}, -pages = {4067721}, -doi = {https://doi.org/10.1155/2024/4067721}, -url = {https://onlinelibrary.wiley.com/doi/abs/10.1155/2024/4067721}, -eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1155/2024/4067721}, -abstract = {Education is crucial for a productive life and providing necessary resources. With the advent of technology like artificial intelligence, higher education institutions are incorporating technology into traditional teaching methods. Predicting academic success has gained interest in education as a strong academic record improves a university’s ranking and increases student employment opportunities. Modern learning institutions face challenges in analyzing performance, providing high-quality education, formulating strategies for evaluating students’ performance, and identifying future needs. E-learning is a rapidly growing and advanced form of education, where students enroll in online courses. Platforms like Intelligent Tutoring Systems (ITS), learning management systems (LMS), and massive open online courses (MOOC) use educational data mining (EDM) to develop automatic grading systems, recommenders, and adaptative systems. However, e-learning is still considered a challenging learning environment due to the lack of direct interaction between students and course instructors. Machine learning (ML) is used in developing adaptive intelligent systems that can perform complex tasks beyond human abilities. Some areas of applications of ML algorithms include cluster analysis, pattern recognition, image processing, natural language processing, and medical diagnostics. In this research work, K-means, a clustering data mining technique using Davies’ Bouldin method, obtains clusters to find important features affecting students’ performance. The study found that the SVM algorithm had the best prediction results after parameter adjustment, with a 96\% accuracy rate. In this paper, the researchers have examined the functions of the Support Vector Machine, Decision Tree, naive Bayes, and KNN classifiers. The outcomes of parameter adjustment greatly increased the accuracy of the four prediction models. Naïve Bayes model’s prediction accuracy is the lowest when compared to other prediction methods, as it assumes a strong independent relationship between features.}, -year = {2024} -} - -@Inproceedings{Arthurs, -author={Arthurs, Noah and Stenhaug, Ben and Karayev, Sergey and Piech, Chris}, -booktitle={International Conference on Educational Data Mining (EDM)}, -title={Grades Are Not Normal: Improving Exam Score Models Using the Logit-Normal Distribution}, -year={2019}, -type={article}, -language={English}, -volume={}, -number={}, -pages={6}, -url={https://eric.ed.gov/?id=ED599204} -} - -@article{HAZEM, -author = {Hazem A. Alrakhawi and Nurullizam Jamiat and Samy S. Abu-Naser}, -title = {Intelligent Tutoring Systems in education: A systematic review of usage, tools, effects and evaluation}, -journal = {Journal of Theoretical and Applied Information Technology}, -volume = {2023}, -number = {4}, -pages = {4067721}, -doi = {}, -url = {}, -abstract = {}, -year = {2023} -} - -@Article{Liu2023, -author={Liu, Mengchi -and Yu, Dongmei}, -title={Towards intelligent E-learning systems}, -journal={Education and Information Technologies}, -year={2023}, -month={Jul}, -day={01}, -volume={28}, -number={7}, -pages={7845-7876}, -abstract={The prevalence of e-learning systems has made educational resources more accessible, interactive and effective to learners without the geographic and temporal boundaries. However, as the number of users increases and the volume of data grows, current e-learning systems face some technical and pedagogical challenges. This paper provides a comprehensive review on the efforts of applying new information and communication technologies to improve e-learning services. We first systematically investigate current e-learning systems in terms of their classification, architecture, functions, challenges, and current trends. We then present a general architecture for big data based e-learning systems to meet the ever-growing demand for e-learning. We also describe how to use data generated in big data based e-learning systems to support more flexible and customized course delivery and personalized learning.}, -issn={1573-7608}, -doi={10.1007/s10639-022-11479-6}, -url={https://doi.org/10.1007/s10639-022-11479-6} -} - -@InProceedings{10.1007/978-3-031-63646-2_13, -author="Soto-Forero, Daniel -and Ackermann, Simha -and Betbeder, Marie-Laure -and Henriet, Julien", -editor="Recio-Garcia, Juan A. -and Orozco-del-Castillo, Mauricio G. -and Bridge, Derek", -title="The Intelligent Tutoring System AI-VT with Case-Based Reasoning and Real Time Recommender Models", -booktitle="Case-Based Reasoning Research and Development", -year="2024", -publisher="Springer Nature Switzerland", -address="Cham", -pages="191--205", -abstract="This paper presents a recommendation model coupled on an existing CBR system model through a new modular architecture designed to integrate multiple services in a learning system called AI-VT (Artificial Intelligence Training System). The recommendation model provides a semi-automatic review of the CBR, two variants of the recommendation model have been implemented: deterministic and stochastic. The model has been tested with 1000 simulated learners, and compared with an original CBR system and BKT (Bayesian Knowledge Tracing) recommender system. The results show that the proposed model identifies learners' weaknesses correctly and revises the content of the ITS (Intelligent Tutoring System) better than the original ITS with CBR. Compared to BKT, the results at each level of complexity are variable, but overall the proposed stochastic model obtains better results.", -isbn="978-3-031-63646-2" -} - -@article{doi:10.1137/23M1592420, -author = {Minsker, Stanislav and Strawn, Nate}, -title = {The Geometric Median and Applications to Robust Mean Estimation}, -journal = {SIAM Journal on Mathematics of Data Science}, -volume = {6}, -number = {2}, -pages = {504-533}, -year = {2024}, -doi = {10.1137/23M1592420}, -URL = { https://doi.org/10.1137/23M1592420}, -eprint = {https://doi.org/10.1137/23M1592420}, -abstract = { Abstract.This paper is devoted to the statistical and numerical properties of the geometric median and its applications to the problem of robust mean estimation via the median of means principle. Our main theoretical results include (a) an upper bound for the distance between the mean and the median for general absolutely continuous distributions in \(\mathbb R^d\), and examples of specific classes of distributions for which these bounds do not depend on the ambient dimension \(d\); (b) exponential deviation inequalities for the distance between the sample and the population versions of the geometric median, which again depend only on the trace-type quantities and not on the ambient dimension. As a corollary, we deduce improved bounds for the (geometric) median of means estimator that hold for large classes of heavy-tailed distributions. Finally, we address the error of numerical approximation, which is an important practical aspect of any statistical estimation procedure. We demonstrate that the objective function minimized by the geometric median satisfies a “local quadratic growth” condition that allows one to translate suboptimality bounds for the objective function to the corresponding bounds for the numerical approximation to the median itself and propose a simple stopping rule applicable to any optimization method which yields explicit error guarantees. We conclude with the numerical experiments, including the application to estimation of mean values of log-returns for S\&P 500 data. } -} - -@article{lei2024analysis, - title={Analysis of Simpson’s Paradox and Its Applications}, - author={Lei, Zhihao}, - journal={Highlights in Science, Engineering and Technology}, - volume={88}, - pages={357--362}, - year={2024} -} - -@InProceedings{pmlr-v108-seznec20a, - title = {A single algorithm for both restless and rested rotting bandits}, - author = {Seznec, Julien and Menard, Pierre and Lazaric, Alessandro and Valko, Michal}, - booktitle = {Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics}, - pages = {3784--3794}, - year = {2020}, - editor = {Chiappa, Silvia and Calandra, Roberto}, - volume = {108}, - series = {Proceedings of Machine Learning Research}, - month = {26--28 Aug}, - publisher = {PMLR}, - pdf = {http://proceedings.mlr.press/v108/seznec20a/seznec20a.pdf}, - url = {https://proceedings.mlr.press/v108/seznec20a.html}, - abstract = {In many application domains (e.g., recommender systems, intelligent tutoring systems), the rewards associated to the available actions tend to decrease over time. This decay is either caused by the actions executed in the past (e.g., a user may get bored when songs of the same genre are recommended over and over) or by an external factor (e.g., content becomes outdated). These two situations can be modeled as specific instances of the rested and restless bandit settings, where arms are rotting (i.e., their value decrease over time). These problems were thought to be significantly different, since Levine et al. (2017) showed that state-of-the-art algorithms for restless bandit perform poorly in the rested rotting setting. In this paper, we introduce a novel algorithm, Rotting Adaptive Window UCB (RAW-UCB), that achieves near-optimal regret in both rotting rested and restless bandit, without any prior knowledge of the setting (rested or restless) and the type of non-stationarity (e.g., piece-wise constant, bounded variation). This is in striking contrast with previous negative results showing that no algorithm can achieve similar results as soon as rewards are allowed to increase. We confirm our theoretical findings on a number of synthetic and dataset-based experiments.} -} diff --git a/main.aux b/main.aux new file mode 100644 index 0000000..36762cb --- /dev/null +++ b/main.aux @@ -0,0 +1,178 @@ +\relax +\providecommand\babel@aux[2]{} +\@nameuse{bbl@beforestart} +\catcode `:\active +\catcode `;\active +\catcode `!\active +\catcode `?\active +\babel@aux{french}{} +\providecommand\hyper@newdestlabel[2]{} +\providecommand\HyperFirstAtBeginDocument{\AtBeginDocument} +\HyperFirstAtBeginDocument{\ifx\hyper@anchor\@undefined +\global\let\oldnewlabel\newlabel +\gdef\newlabel#1#2{\newlabelxx{#1}#2} +\gdef\newlabelxx#1#2#3#4#5#6{\oldnewlabel{#1}{{#2}{#3}}} +\AtEndDocument{\ifx\hyper@anchor\@undefined +\let\newlabel\oldnewlabel +\fi} +\fi} +\global\let\hyper@last\relax +\gdef\HyperFirstAtBeginDocument#1{#1} +\providecommand\HyField@AuxAddToFields[1]{} +\providecommand\HyField@AuxAddToCoFields[2]{} +\@writefile{toc}{\contentsline {part}{I\hspace {1em}Contexte et Problématiques}{1}{part.1}\protected@file@percent } +\citation{Nkambou} +\citation{doi:10.1177/1754337116651013} +\@writefile{toc}{\contentsline {chapter}{\numberline {1}Introduction}{3}{chapter.1}\protected@file@percent } +\@writefile{lof}{\addvspace {10\p@ }} +\@writefile{lot}{\addvspace {10\p@ }} +\@writefile{toc}{\contentsline {section}{\numberline {1.1}Contributions Principales}{4}{section.1.1}\protected@file@percent } +\@writefile{toc}{\contentsline {section}{\numberline {1.2}Plan de la thèse}{5}{section.1.2}\protected@file@percent } +\@writefile{toc}{\contentsline {chapter}{\numberline {2}Contexte}{7}{chapter.2}\protected@file@percent } +\@writefile{lof}{\addvspace {10\p@ }} +\@writefile{lot}{\addvspace {10\p@ }} +\@writefile{toc}{\contentsline {section}{\numberline {2.1}Notions et Algorithmes}{7}{section.2.1}\protected@file@percent } +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.1}Environnements Informatiques pour l'Apprentissage Humain (EIAH)}{7}{subsection.2.1.1}\protected@file@percent } +\citation{Nkambou} +\citation{Nkambou} +\citation{LALITHA2020583} +\citation{LALITHA2020583} +\citation{LALITHA2020583} +\@writefile{lof}{\contentsline {figure}{\numberline {2.1}{\ignorespaces L'architecture générale des EIAH, les composantes et leurs interactions (\cite {Nkambou})\relax }}{8}{figure.caption.3}\protected@file@percent } +\providecommand*\caption@xref[2]{\@setref\relax\@undefined{#1}} +\newlabel{fig:figEIAH}{{2.1}{8}{L'architecture générale des EIAH, les composantes et leurs interactions (\cite {Nkambou})\relax }{figure.caption.3}{}} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.2}Stratégies d'Apprentissage}{8}{subsection.2.1.2}\protected@file@percent } +\citation{ROLDANREYES20151} +\@writefile{lof}{\contentsline {figure}{\numberline {2.2}{\ignorespaces Stratégies d'apprentissage (\cite {LALITHA2020583})\relax }}{9}{figure.caption.4}\protected@file@percent } +\newlabel{fig:figStra}{{2.2}{9}{Stratégies d'apprentissage (\cite {LALITHA2020583})\relax }{figure.caption.4}{}} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.3}Raisonnement à partir de cas (RàPC)}{9}{subsection.2.1.3}\protected@file@percent } +\citation{10.1007/978-3-030-58342-2_20} +\citation{min8100434} +\citation{min8100434} +\citation{min8100434} +\citation{Richter2013} +\@writefile{lof}{\contentsline {figure}{\numberline {2.3}{\ignorespaces Cycle fondamental du raisonnement à partir de cas (\cite {min8100434})\relax }}{10}{figure.caption.5}\protected@file@percent } +\newlabel{fig:figCycle}{{2.3}{10}{Cycle fondamental du raisonnement à partir de cas (\cite {min8100434})\relax }{figure.caption.5}{}} +\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.1.3.1}Rechercher}{11}{subsubsection.2.1.3.1}\protected@file@percent } +\newlabel{eqSim}{{2.1}{11}{Rechercher}{equation.2.1.1}{}} +\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.1.3.2}Adapter (Réutiliser)}{11}{subsubsection.2.1.3.2}\protected@file@percent } +\citation{Richter2013} +\citation{Richter2013} +\@writefile{lof}{\contentsline {figure}{\numberline {2.4}{\ignorespaces Principe de réutilisation dans le RàPC (\cite {Richter2013})\relax }}{12}{figure.caption.6}\protected@file@percent } +\newlabel{fig:figReut}{{2.4}{12}{Principe de réutilisation dans le RàPC (\cite {Richter2013})\relax }{figure.caption.6}{}} +\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.1.3.3}Réviser et Réparer}{12}{subsubsection.2.1.3.3}\protected@file@percent } +\citation{RICHTER20093} +\citation{Marie} +\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.1.3.4}Stocker (Retenir)}{13}{subsubsection.2.1.3.4}\protected@file@percent } +\@writefile{toc}{\contentsline {subsubsection}{\numberline {2.1.3.5}Conteneurs de Connaissance}{13}{subsubsection.2.1.3.5}\protected@file@percent } +\citation{hajduk2019cognitive} +\@writefile{lof}{\contentsline {figure}{\numberline {2.5}{\ignorespaces Cycle du RàPC, les étapes, les conteneurs et leurs liens\relax }}{14}{figure.caption.7}\protected@file@percent } +\newlabel{fig:figCycle}{{2.5}{14}{Cycle du RàPC, les étapes, les conteneurs et leurs liens\relax }{figure.caption.7}{}} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.4}Systèmes Multi-Agents}{14}{subsection.2.1.4}\protected@file@percent } +\citation{Hoang} +\citation{Hoang} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.5}Pensée Bayesienne}{15}{subsection.2.1.5}\protected@file@percent } +\newlabel{eqBayes}{{2.2}{15}{Pensée Bayesienne}{equation.2.1.2}{}} +\citation{10.1145/3459665} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.6}Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{16}{subsection.2.1.6}\protected@file@percent } +\newlabel{eqDist}{{2.3}{16}{Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{equation.2.1.3}{}} +\newlabel{eqMet1}{{2.4}{16}{Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{equation.2.1.4}{}} +\newlabel{eqMet2}{{2.5}{16}{Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{equation.2.1.5}{}} +\citation{9072123} +\newlabel{eqV1}{{2.6}{17}{Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{equation.2.1.6}{}} +\newlabel{eqV2}{{2.7}{17}{Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{equation.2.1.7}{}} +\newlabel{eqV3}{{2.8}{17}{Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{equation.2.1.8}{}} +\newlabel{eqCond}{{2.9}{17}{Méthode des k plus proches voisins (K-Nearest Neighborhood - KNN)}{equation.2.1.9}{}} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.7}K-Moyennes}{17}{subsection.2.1.7}\protected@file@percent } +\newlabel{eqKM}{{2.10}{17}{K-Moyennes}{equation.2.1.10}{}} +\citation{WANG2021331} +\newlabel{eqCentres}{{2.11}{18}{K-Moyennes}{equation.2.1.11}{}} +\newlabel{eqApp}{{2.12}{18}{K-Moyennes}{equation.2.1.12}{}} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.8}Modèle de Mélange Gaussien GMM (\textit {Gaussian Mixture Model})}{18}{subsection.2.1.8}\protected@file@percent } +\newlabel{eqGMM}{{2.13}{18}{Modèle de Mélange Gaussien GMM (\textit {Gaussian Mixture Model})}{equation.2.1.13}{}} +\newlabel{eqK}{{2.14}{18}{Modèle de Mélange Gaussien GMM (\textit {Gaussian Mixture Model})}{equation.2.1.14}{}} +\newlabel{eqdProb}{{2.15}{18}{Modèle de Mélange Gaussien GMM (\textit {Gaussian Mixture Model})}{equation.2.1.15}{}} +\citation{9627973} +\citation{9434422} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.9}Fuzzy-C}{19}{subsection.2.1.9}\protected@file@percent } +\newlabel{eqFuzzy}{{2.16}{19}{Fuzzy-C}{equation.2.1.16}{}} +\newlabel{eqFCond}{{2.17}{19}{Fuzzy-C}{equation.2.1.17}{}} +\newlabel{eqFV}{{2.18}{19}{Fuzzy-C}{equation.2.1.18}{}} +\newlabel{eqFU}{{2.19}{19}{Fuzzy-C}{equation.2.1.19}{}} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.10}Bandit Manchot MAB (\textit {Multi-Armed Bandits})}{19}{subsection.2.1.10}\protected@file@percent } +\citation{9870279} +\@writefile{toc}{\contentsline {subsection}{\numberline {2.1.11}Échantillonnage de Thompson TS (\textit {Thompson Sampling})}{20}{subsection.2.1.11}\protected@file@percent } +\newlabel{eqBeta}{{2.20}{20}{Échantillonnage de Thompson TS (\textit {Thompson Sampling})}{equation.2.1.20}{}} +\@writefile{lof}{\contentsline {figure}{\numberline {2.6}{\ignorespaces Comportement de la distribution Beta avec différents valeurs dans ses paramètres\relax }}{21}{figure.caption.8}\protected@file@percent } +\newlabel{fig:figBeta}{{2.6}{21}{Comportement de la distribution Beta avec différents valeurs dans ses paramètres\relax }{figure.caption.8}{}} +\@writefile{toc}{\contentsline {part}{II\hspace {1em}État de l'art}{23}{part.2}\protected@file@percent } +\citation{ZHANG2021100025} +\citation{CHIU2023100118} +\@writefile{toc}{\contentsline {chapter}{\numberline {3}Environnements Informatiques d'Apprentissage Humain}{25}{chapter.3}\protected@file@percent } +\@writefile{lof}{\addvspace {10\p@ }} +\@writefile{lot}{\addvspace {10\p@ }} +\@writefile{toc}{\contentsline {section}{\numberline {3.1}L'Intelligence Artificielle}{25}{section.3.1}\protected@file@percent } +\citation{Robertson2014ARO} +\citation{MUANGPRATHUB2020e05227} +\citation{HUANG2023104684} +\@writefile{toc}{\contentsline {section}{\numberline {3.2}Systèmes de Recommandation dans les EIAH}{26}{section.3.2}\protected@file@percent } +\citation{HUANG2023104684} +\citation{HUANG2023104684} +\citation{pmlr-v108-seznec20a} +\citation{INGKAVARA2022100086} +\@writefile{lof}{\contentsline {figure}{\numberline {3.1}{\ignorespaces Architecture du système de recommandation proposé dans \cite {HUANG2023104684}\relax }}{27}{figure.caption.9}\protected@file@percent } +\newlabel{figArch}{{3.1}{27}{Architecture du système de recommandation proposé dans \cite {HUANG2023104684}\relax }{figure.caption.9}{}} +\citation{LALITHA2020583} +\citation{ZHAO2023118535} +\citation{SU2022109547} +\citation{MUANGPRATHUB2020e05227} +\citation{Zhou2021} +\citation{EZALDEEN2022100700} +\citation{EZALDEEN2022100700} +\citation{EZALDEEN2022100700} +\citation{ZHANG2021100025} +\citation{CHIU2023100118} +\citation{Robertson2014ARO} +\citation{HUANG2023104684} +\citation{pmlr-v108-seznec20a} +\citation{INGKAVARA2022100086} +\citation{LALITHA2020583} +\citation{SU2022109547} +\citation{MUANGPRATHUB2020e05227} +\citation{Zhou2021} +\citation{EZALDEEN2022100700} +\bibstyle{apalike} +\bibdata{main.bib} +\@writefile{lof}{\contentsline {figure}{\numberline {3.2}{\ignorespaces Niveaux du système de recommandation dans \cite {EZALDEEN2022100700}\relax }}{30}{figure.caption.10}\protected@file@percent } +\newlabel{fig:figLevels}{{3.2}{30}{Niveaux du système de recommandation dans \cite {EZALDEEN2022100700}\relax }{figure.caption.10}{}} +\@writefile{lot}{\contentsline {table}{\numberline {3.1}{\ignorespaces Tableau de synthèse des articles analysés dans l'état de l'art des EIAH\relax }}{30}{table.caption.11}\protected@file@percent } +\newlabel{tabArts}{{3.1}{30}{Tableau de synthèse des articles analysés dans l'état de l'art des EIAH\relax }{table.caption.11}{}} +\bibcite{CHIU2023100118}{Chiu et~al., 2023} +\bibcite{10.1145/3459665}{Cunningham and Delany, 2021} +\bibcite{EZALDEEN2022100700}{Ezaldeen et~al., 2022} +\bibcite{9434422}{Gupta et~al., 2021} +\bibcite{hajduk2019cognitive}{Hajduk et~al., 2019} +\bibcite{doi:10.1177/1754337116651013}{Henriet et~al., 2017} +\bibcite{Hoang}{Hoang, 2018} +\bibcite{HUANG2023104684}{Huang et~al., 2023} +\bibcite{INGKAVARA2022100086}{Ingkavara et~al., 2022} +\bibcite{LALITHA2020583}{Lalitha and Sreeja, 2020} +\bibcite{min8100434}{Leikola et~al., 2018} +\bibcite{10.1007/978-3-030-58342-2_20}{Lepage et~al., 2020} +\bibcite{9870279}{Lin, 2022} +\bibcite{Marie}{Marie, 2019} +\bibcite{MUANGPRATHUB2020e05227}{Muangprathub et~al., 2020} +\bibcite{Nkambou}{Nkambou et~al., 2010} +\bibcite{Richter2013}{Richter and Weber, 2013} +\bibcite{RICHTER20093}{Richter, 2009} +\bibcite{Robertson2014ARO}{Robertson and Watson, 2014} +\bibcite{ROLDANREYES20151}{{Roldan Reyes} et~al., 2015} +\bibcite{pmlr-v108-seznec20a}{Seznec et~al., 2020} +\bibcite{9072123}{Sinaga and Yang, 2020} +\bibcite{SU2022109547}{Su et~al., 2022} +\bibcite{WANG2021331}{Wang et~al., 2021} +\bibcite{9627973}{Xu et~al., 2021} +\bibcite{ZHANG2021100025}{Zhang. and Aslan, 2021} +\bibcite{ZHAO2023118535}{Zhao et~al., 2023} +\bibcite{Zhou2021}{Zhou and Wang, 2021} +\gdef \@abspage@last{42} diff --git a/main.bbl b/main.bbl new file mode 100644 index 0000000..dea65ef --- /dev/null +++ b/main.bbl @@ -0,0 +1,174 @@ +\begin{thebibliography}{} + +\bibitem[Chiu et~al., 2023]{CHIU2023100118} +Chiu, T.~K., Xia, Q., Zhou, X., Chai, C.~S., and Cheng, M. (2023). +\newblock Systematic literature review on opportunities, challenges, and future + research recommendations of artificial intelligence in education. +\newblock {\em Computers and Education: Artificial Intelligence}, 4:100118. + +\bibitem[Cunningham and Delany, 2021]{10.1145/3459665} +Cunningham, P. and Delany, S.~J. (2021). +\newblock K-nearest neighbour classifiers - a tutorial. +\newblock {\em ACM Comput. Surv.}, 54(6). + +\bibitem[Ezaldeen et~al., 2022]{EZALDEEN2022100700} +Ezaldeen, H., Misra, R., Bisoy, S.~K., Alatrash, R., and Priyadarshini, R. + (2022). +\newblock A hybrid e-learning recommendation integrating adaptive profiling and + sentiment analysis. +\newblock {\em Journal of Web Semantics}, 72:100700. + +\bibitem[Gupta et~al., 2021]{9434422} +Gupta, S., Chaudhari, S., Joshi, G., and Yağan, O. (2021). +\newblock Multi-armed bandits with correlated arms. +\newblock {\em IEEE Transactions on Information Theory}, 67(10):6711--6732. + +\bibitem[Hajduk et~al., 2019]{hajduk2019cognitive} +Hajduk, M., Sukop, M., and Haun, M. (2019). +\newblock {\em Cognitive Multi-agent Systems: Structures, Strategies and + Applications to Mobile Robotics and Robosoccer}. +\newblock Studies in Systems, Decision and Control. Springer International + Publishing. + +\bibitem[Henriet et~al., 2017]{doi:10.1177/1754337116651013} +Henriet, J., Christophe, L., and Laurent, P. (2017). +\newblock Artificial intelligence-virtual trainer: An educative system based on + artificial intelligence and designed to produce varied and consistent + training lessons. +\newblock {\em Proceedings of the Institution of Mechanical Engineers, Part P: + Journal of Sports Engineering and Technology}, 231(2):110--124. + +\bibitem[Hoang, 2018]{Hoang} +Hoang, L. (2018). +\newblock {\em La formule du savoir. Une philosophie unifiée du savoir fondée + sur le théorème de Bayes}. +\newblock EDP Sciences. + +\bibitem[Huang et~al., 2023]{HUANG2023104684} +Huang, A.~Y., Lu, O.~H., and Yang, S.~J. (2023). +\newblock Effects of artificial intelligence–enabled personalized + recommendations on learners’ learning engagement, motivation, and outcomes + in a flipped classroom. +\newblock {\em Computers and Education}, 194:104684. + +\bibitem[Ingkavara et~al., 2022]{INGKAVARA2022100086} +Ingkavara, T., Panjaburee, P., Srisawasdi, N., and Sajjapanroj, S. (2022). +\newblock The use of a personalized learning approach to implementing + self-regulated online learning. +\newblock {\em Computers and Education: Artificial Intelligence}, 3:100086. + +\bibitem[Lalitha and Sreeja, 2020]{LALITHA2020583} +Lalitha, T.~B. and Sreeja, P.~S. (2020). +\newblock Personalised self-directed learning recommendation system. +\newblock {\em Procedia Computer Science}, 171:583--592. +\newblock Third International Conference on Computing and Network + Communications (CoCoNet'19). + +\bibitem[Leikola et~al., 2018]{min8100434} +Leikola, M., Sauer, C., Rintala, L., Aromaa, J., and Lundström, M. (2018). +\newblock Assessing the similarity of cyanide-free gold leaching processes: A + case-based reasoning application. +\newblock {\em Minerals}, 8(10). + +\bibitem[Lepage et~al., 2020]{10.1007/978-3-030-58342-2_20} +Lepage, Y., Lieber, J., Mornard, I., Nauer, E., Romary, J., and Sies, R. + (2020). +\newblock The french correction: When retrieval is harder to specify than + adaptation. +\newblock In Watson, I. and Weber, R., editors, {\em Case-Based Reasoning + Research and Development}, pages 309--324, Cham. Springer International + Publishing. + +\bibitem[Lin, 2022]{9870279} +Lin, B. (2022). +\newblock Evolutionary multi-armed bandits with genetic thompson sampling. +\newblock In {\em 2022 IEEE Congress on Evolutionary Computation (CEC)}, pages + 1--8. + +\bibitem[Marie, 2019]{Marie} +Marie, F. (2019). +\newblock Coliseum-3d. une plate-forme innovante pour la segmentation + d’images médicales par raisonnement à partir de cas (ràpc) et méthodes + d’apprentissage de type deep learning. + +\bibitem[Muangprathub et~al., 2020]{MUANGPRATHUB2020e05227} +Muangprathub, J., Boonjing, V., and Chamnongthai, K. (2020). +\newblock Learning recommendation with formal concept analysis for intelligent + tutoring system. +\newblock {\em Heliyon}, 6(10):e05227. + +\bibitem[Nkambou et~al., 2010]{Nkambou} +Nkambou, R., Bourdeau, J., and Mizoguchi, R. (2010). +\newblock {\em Advances in Intelligent Tutoring Systems}. +\newblock Springer Berlin, Heidelberg, 1 edition. + +\bibitem[Richter and Weber, 2013]{Richter2013} +Richter, M. and Weber, R. (2013). +\newblock {\em Case-Based Reasoning (A Textbook)}. +\newblock Springer-Verlag GmbH. + +\bibitem[Richter, 2009]{RICHTER20093} +Richter, M.~M. (2009). +\newblock The search for knowledge, contexts, and case-based reasoning. +\newblock {\em Engineering Applications of Artificial Intelligence}, + 22(1):3--9. + +\bibitem[Robertson and Watson, 2014]{Robertson2014ARO} +Robertson, G. and Watson, I.~D. (2014). +\newblock A review of real-time strategy game ai. +\newblock {\em AI Mag.}, 35:75--104. + +\bibitem[{Roldan Reyes} et~al., 2015]{ROLDANREYES20151} +{Roldan Reyes}, E., Negny, S., {Cortes Robles}, G., and {Le Lann}, J. (2015). +\newblock Improvement of online adaptation knowledge acquisition and reuse in + case-based reasoning: Application to process engineering design. +\newblock {\em Engineering Applications of Artificial Intelligence}, 41:1--16. + +\bibitem[Seznec et~al., 2020]{pmlr-v108-seznec20a} +Seznec, J., Menard, P., Lazaric, A., and Valko, M. (2020). +\newblock A single algorithm for both restless and rested rotting bandits. +\newblock In Chiappa, S. and Calandra, R., editors, {\em Proceedings of the + Twenty Third International Conference on Artificial Intelligence and + Statistics}, volume 108 of {\em Proceedings of Machine Learning Research}, + pages 3784--3794. PMLR. + +\bibitem[Sinaga and Yang, 2020]{9072123} +Sinaga, K.~P. and Yang, M.-S. (2020). +\newblock Unsupervised k-means clustering algorithm. +\newblock {\em IEEE Access}, 8:80716--80727. + +\bibitem[Su et~al., 2022]{SU2022109547} +Su, Y., Cheng, Z., Wu, J., Dong, Y., Huang, Z., Wu, L., Chen, E., Wang, S., and + Xie, F. (2022). +\newblock Graph-based cognitive diagnosis for intelligent tutoring systems. +\newblock {\em Knowledge-Based Systems}, 253:109547. + +\bibitem[Wang et~al., 2021]{WANG2021331} +Wang, F., Liao, F., Li, Y., and Wang, H. (2021). +\newblock A new prediction strategy for dynamic multi-objective optimization + using gaussian mixture model. +\newblock {\em Information Sciences}, 580:331--351. + +\bibitem[Xu et~al., 2021]{9627973} +Xu, S., Cai, W., Xia, H., Liu, B., and Xu, J. (2021). +\newblock Dynamic metric accelerated method for fuzzy clustering. +\newblock {\em IEEE Access}, 9:166838--166854. + +\bibitem[Zhang. and Aslan, 2021]{ZHANG2021100025} +Zhang., K. and Aslan, A.~B. (2021). +\newblock Ai technologies for education: Recent research and future directions. +\newblock {\em Computers and Education: Artificial Intelligence}, 2:100025. + +\bibitem[Zhao et~al., 2023]{ZHAO2023118535} +Zhao, L.-T., Wang, D.-S., Liang, F.-Y., and Chen, J. (2023). +\newblock A recommendation system for effective learning strategies: An + integrated approach using context-dependent dea. +\newblock {\em Expert Systems with Applications}, 211:118535. + +\bibitem[Zhou and Wang, 2021]{Zhou2021} +Zhou, L. and Wang, C. (2021). +\newblock Research on recommendation of personalized exercises in english + learning based on data mining. +\newblock {\em Scientific Programming}, 2021:5042286. + +\end{thebibliography} diff --git a/main.bib b/main.bib new file mode 100644 index 0000000..f5d1b18 --- /dev/null +++ b/main.bib @@ -0,0 +1,1602 @@ +@article{ZHANG2021100025, +title = {AI technologies for education: Recent research and future directions}, +journal = {Computers and Education: Artificial Intelligence}, +volume = {2}, +pages = {100025}, +language = {English}, +year = {2021}, +issn = {2666-920X}, +type = {article}, +doi = {https://doi.org/10.1016/j.caeai.2021.100025}, +url = {https://www.sciencedirect.com/science/article/pii/S2666920X21000199}, +author = {Ke Zhang. and Ayse Begum Aslan}, +address={USA}, +affiliation={Wayne State University; Eastern Michigan University}, +keywords = {Artificial intelligence, AI, AI in Education}, +abstract = {From unique educational perspectives, this article reports a comprehensive review of selected empirical studies on artificial intelligence in education (AIEd) published in 1993–2020, as collected in the Web of Sciences database and selected AIEd-specialized journals. A total of 40 empirical studies met all selection criteria, and were fully reviewed using multiple methods, including selected bibliometrics, content analysis and categorical meta-trends analysis. This article reports the current state of AIEd research, highlights selected AIEd technologies and applications, reviews their proven and potential benefits for education, bridges the gaps between AI technological innovations and their educational applications, and generates practical examples and inspirations for both technological experts that create AIEd technologies and educators who spearhead AI innovations in education. It also provides rich discussions on practical implications and future research directions from multiple perspectives. The advancement of AIEd calls for critical initiatives to address AI ethics and privacy concerns, and requires interdisciplinary and transdisciplinary collaborations in large-scaled, longitudinal research and development efforts.} +} + +@article{PETROVIC201617, +title = {Knowledge-light adaptation approaches in case-based reasoning for radiotherapy treatment planning}, +journal = {Artificial Intelligence in Medicine}, +volume = {68}, +pages = {17-28}, +year = {2016}, +language = {English}, +issn = {0933-3657}, +type = {article}, +doi = {https://doi.org/10.1016/j.artmed.2016.01.006}, +url = {https://www.sciencedirect.com/science/article/pii/S093336571630015X}, +author = {Sanja Petrovic and Gulmira Khussainova and Rupa Jagannathan}, +affiliation={Nottingham University}, +address={UK}, +keywords = {Case-based reasoning, Adaptation-guided retrieval, Machine-learning tools, Radiotherapy treatment planning}, +abstract = {Objective +Radiotherapy treatment planning aims at delivering a sufficient radiation dose to cancerous tumour cells while sparing healthy organs in the tumour-surrounding area. It is a time-consuming trial-and-error process that requires the expertise of a group of medical experts including oncologists and medical physicists and can take from 2 to 3h to a few days. Our objective is to improve the performance of our previously built case-based reasoning (CBR) system for brain tumour radiotherapy treatment planning. In this system, a treatment plan for a new patient is retrieved from a case base containing patient cases treated in the past and their treatment plans. However, this system does not perform any adaptation, which is needed to account for any difference between the new and retrieved cases. Generally, the adaptation phase is considered to be intrinsically knowledge-intensive and domain-dependent. Therefore, an adaptation often requires a large amount of domain-specific knowledge, which can be difficult to acquire and often is not readily available. In this study, we investigate approaches to adaptation that do not require much domain knowledge, referred to as knowledge-light adaptation. +Methodology +We developed two adaptation approaches: adaptation based on machine-learning tools and adaptation-guided retrieval. They were used to adapt the beam number and beam angles suggested in the retrieved case. Two machine-learning tools, neural networks and naive Bayes classifier, were used in the adaptation to learn how the difference in attribute values between the retrieved and new cases affects the output of these two cases. The adaptation-guided retrieval takes into consideration not only the similarity between the new and retrieved cases, but also how to adapt the retrieved case. +Results +The research was carried out in collaboration with medical physicists at the Nottingham University Hospitals NHS Trust, City Hospital Campus, UK. All experiments were performed using real-world brain cancer patient cases treated with three-dimensional (3D)-conformal radiotherapy. Neural networks-based adaptation improved the success rate of the CBR system with no adaptation by 12%. However, naive Bayes classifier did not improve the current retrieval results as it did not consider the interplay among attributes. The adaptation-guided retrieval of the case for beam number improved the success rate of the CBR system by 29%. However, it did not demonstrate good performance for the beam angle adaptation. Its success rate was 29% versus 39% when no adaptation was performed. +Conclusions +The obtained empirical results demonstrate that the proposed adaptation methods improve the performance of the existing CBR system in recommending the number of beams to use. However, we also conclude that to be effective, the proposed adaptation of beam angles requires a large number of relevant cases in the case base.} +} + +@article{ROLDANREYES20151, +title = {Improvement of online adaptation knowledge acquisition and reuse in case-based reasoning: Application to process engineering design}, +journal = {Engineering Applications of Artificial Intelligence}, +volume = {41}, +pages = {1-16}, +affiliation={Université de Toulouse; Instituto Tecnologico de Orizaba}, +country={France}, +language = {English}, +year = {2015}, +type = {article}, +issn = {0952-1976}, +doi = {https://doi.org/10.1016/j.engappai.2015.01.015}, +url = {https://www.sciencedirect.com/science/article/pii/S0952197615000263}, +author = {E. {Roldan Reyes} and S. Negny and G. {Cortes Robles} and J.M. {Le Lann}}, +keywords = {Case based reasoning, Constraint satisfaction problems, Interactive adaptation method, Online knowledge acquisition, Failure diagnosis and repair}, +abstract = {Despite various publications in the area during the last few years, the adaptation step is still a crucial phase for a relevant and reasonable Case Based Reasoning system. Furthermore, the online acquisition of the new adaptation knowledge is of particular interest as it enables the progressive improvement of the system while reducing the knowledge engineering effort without constraints for the expert. Therefore this paper presents a new interactive method for adaptation knowledge elicitation, acquisition and reuse, thanks to a modification of the traditional CBR cycle. Moreover to improve adaptation knowledge reuse, a test procedure is also implemented to help the user in the adaptation step and its diagnosis during adaptation failure. A study on the quality and usefulness of the new knowledge acquired is also driven. As our Knowledge Based Systems (KBS) is more focused on preliminary design, and more particularly in the field of process engineering, we need to unify in the same method two types of knowledge: contextual and general. To realize this, this article proposes the integration of the Constraint Satisfaction Problem (based on general knowledge) approach into the Case Based Reasoning (based on contextual knowledge) process to improve the case representation and the adaptation of past experiences. To highlight its capability, the proposed approach is illustrated through a case study dedicated to the design of an industrial mixing device.} +} + +@article{JUNG20095695, +title = {Integrating radial basis function networks with case-based reasoning for product design}, +journal = {Expert Systems with Applications}, +volume = {36}, +number = {3, Part 1}, +language = {English}, +pages = {5695-5701}, +year = {2009}, +type = {article}, +issn = {0957-4174}, +doi = {https://doi.org/10.1016/j.eswa.2008.06.099}, +url = {https://www.sciencedirect.com/science/article/pii/S0957417408003667}, +author = {Sabum Jung and Taesoo Lim and Dongsoo Kim}, +affiliation={LG Production Engineering Research Institute; Sungkyul University; Soongsil University}, +keywords = {Case-based reasoning (CBR), Radial basis function network (RBFN), Design expert system, Product design}, +abstract = {This paper presents a case-based design expert system that automatically determines the design values of a product. We focus on the design problem of a shadow mask which is a core component of monitors in the electronics industry. In case-based reasoning (CBR), it is important to retrieve similar cases and adapt them to meet design specifications exactly. Notably, difficulties in automating the adaptation process have prevented designers from being able to use design expert systems easily and efficiently. In this paper, we present a hybrid approach combining CBR and artificial neural networks in order to solve the problems occurring during the adaptation process. We first constructed a radial basis function network (RBFN) composed of representative cases created by K-means clustering. Then, the representative case most similar to the current problem was adjusted using the network. The rationale behind the proposed approach is discussed, and experimental results acquired from real shadow mask design are presented. Using the design expert system, designers can reduce design time and errors and enhance the total quality of design. Furthermore, the expert system facilitates effective sharing of design knowledge among designers.} +} + +@article{CHIU2023100118, +title = {Systematic literature review on opportunities, challenges, and future research recommendations of artificial intelligence in education}, +journal = {Computers and Education: Artificial Intelligence}, +volume = {4}, +language = {English}, +type = {article}, +pages = {100118}, +year = {2023}, +issn = {2666-920X}, +doi = {https://doi.org/10.1016/j.caeai.2022.100118}, +url = {https://www.sciencedirect.com/science/article/pii/S2666920X2200073X}, +author = {Thomas K.F. Chiu and Qi Xia and Xinyan Zhou and Ching Sing Chai and Miaoting Cheng}, +keywords = {Artificial intelligence, Artificial intelligence in education, Systematic review, Learning, Teaching, Assessment}, +abstract = {Applications of artificial intelligence in education (AIEd) are emerging and are new to researchers and practitioners alike. Reviews of the relevant literature have not examined how AI technologies have been integrated into each of the four key educational domains of learning, teaching, assessment, and administration. The relationships between the technologies and learning outcomes for students and teachers have also been neglected. This systematic review study aims to understand the opportunities and challenges of AIEd by examining the literature from the last 10 years (2012–2021) using matrix coding and content analysis approaches. The results present the current focus of AIEd research by identifying 13 roles of AI technologies in the key educational domains, 7 learning outcomes of AIEd, and 10 major challenges. The review also provides suggestions for future directions of AIEd research.} +} + +@article{Robertson2014ARO, +title = {A Review of Real-Time Strategy Game AI}, +author = {Glen Robertson and Ian D. Watson}, +affiliation = {University of Auckland }, +keywords = {Game, IA, Real-time strategy}, +type={article}, +language={English}, +abstract = {This literature review covers AI techniques used for real-time strategy video games, focusing specifically on StarCraft. It finds that the main areas of current academic research are in tactical and strategic decision making, plan recognition, and learning, and it outlines the research contributions in each of these areas. The paper then contrasts the use of game AI in academe and industry, finding the academic research heavily focused on creating game-winning agents, while the industry aims to maximize player enjoyment. It finds that industry adoption of academic research is low because it is either inapplicable or too time-consuming and risky to implement in a new game, which highlights an area for potential investigation: bridging the gap between academe and industry. Finally, the areas of spatial reasoning, multiscale AI, and cooperation are found to require future work, and standardized evaluation methods are proposed to produce comparable results between studies.}, +journal = {AI Mag.}, +year = {2014}, +volume = {35}, +pages = {75-104} +} + +@Inproceedings{10.1007/978-3-642-15973-2_50, +author={Butdee, S. +and Tichkiewitch, S.}, +affiliation={University of Technology North Bangkok; Grenoble Institute of Technology}, +editor={Bernard, Alain}, +title={Case-Based Reasoning for Adaptive Aluminum Extrusion Die Design Together with Parameters by Neural Networks}, +keywords={Adaptive die design and parameters, Optimal aluminum extrusion, Case-based reasoning, Neural networks}, +booktitle={Global Product Development}, +year={2011}, +type = {article; proceedings paper}, +language = {English}, +publisher = {Springer Berlin Heidelberg}, +address = {Berlin, Heidelberg}, +pages = {491--496}, +abstract = {Nowadays Aluminum extrusion die design is a critical task for improving productivity which involves with quality, time and cost. Case-Based Reasoning (CBR) method has been successfully applied to support the die design process in order to design a new die by tackling previous problems together with their solutions to match with a new similar problem. Such solutions are selected and modified to solve the present problem. However, the applications of the CBR are useful only retrieving previous features whereas the critical parameters are missing. In additions, the experience learning to such parameters are limited. This chapter proposes Artificial Neural Network (ANN) to associate the CBR in order to learning previous parameters and predict to the new die design according to the primitive die modification. The most satisfactory is to accommodate the optimal parameters of extrusion processes.}, +isbn = {978-3-642-15973-2} +} + +@Inproceedings{10.1007/978-3-319-47096-2_11, +author={Grace, Kazjon +and Maher, Mary Lou +and Wilson, David C. +and Najjar, Nadia A.}, +affiliation={University of North Carolina at Charlotte}, +editor={Goel, Ashok +and D{\'i}az-Agudo, M Bel{\'e}n +and Roth-Berghofer, Thomas}, +title={Combining CBR and Deep Learning to Generate Surprising Recipe Designs}, +keywords={Case-based reasoning, deep learning, recipe design}, +type = {article; proceedings paper}, +booktitle={Case-Based Reasoning Research and Development}, +year={2016}, +publisher={Springer International Publishing}, +address={Cham}, +language = {English}, +pages={154--169}, +abstract={This paper presents a dual-cycle CBR model in the domain of recipe generation. The model combines the strengths of deep learning and similarity-based retrieval to generate recipes that are novel and valuable (i.e. they are creative). The first cycle generates abstract descriptions which we call ``design concepts'' by synthesizing expectations from the entire case base, while the second cycle uses those concepts to retrieve and adapt objects. We define these conceptual object representations as an abstraction over complete cases on which expectations can be formed, allowing objects to be evaluated for surprisingness (the peak level of unexpectedness in the object, given the case base) and plausibility (the overall similarity of the object to those in the case base). The paper presents a prototype implementation of the model, and demonstrates its ability to generate objects that are simultaneously plausible and surprising, in addition to fitting a user query. This prototype is then compared to a traditional single-cycle CBR system.}, +isbn={978-3-319-47096-2} +} + +@Inproceedings{10.1007/978-3-319-61030-6_1, +author={Maher, Mary Lou +and Grace, Kazjon}, +editor={Aha, David W. +and Lieber, Jean}, +affiliation={University of North Carolina at Charlotte}, +title={Encouraging Curiosity in Case-Based Reasoning and Recommender Systems}, +keywords={Curiosity, Case-based reasoning, Recommender systems}, +booktitle={Case-Based Reasoning Research and Development}, +year={2017}, +publisher={Springer International Publishing}, +address={Cham}, +pages={3--15}, +language = {English}, +type = {article; proceedings paper}, +abstract={A key benefit of case-based reasoning (CBR) and recommender systems is the use of past experience to guide the synthesis or selection of the best solution for a specific context or user. Typically, the solution presented to the user is based on a value system that privileges the closest match in a query and the solution that performs best when evaluated according to predefined requirements. In domains in which creativity is desirable or the user is engaged in a learning activity, there is a benefit to moving beyond the expected or ``best match'' and include results based on computational models of novelty and surprise. In this paper, models of novelty and surprise are integrated with both CBR and Recommender Systems to encourage user curiosity.}, +isbn={978-3-319-61030-6} +} + +@Inproceedings{Muller, +author = {Müller, G. and Bergmann, R.}, +affiliation={University of Trier}, +year = {2015}, +month = {01}, +language = {English}, +type = {article; proceedings paper}, +abstract = {This paper presents CookingCAKE,a framework for the adaptation of cooking recipes represented as workflows. CookingCAKE integrates and combines several workflow adaptation approaches applied in process-oriented case based reasoning (POCBR) in a single adaptation framework, thus providing a capable tool for the adaptation of cooking recipes. The available case base of cooking workflows is analyzed to generate adaptation knowledge which is used to adapt a recipe regarding restrictions and resources, which the user may define for the preparation of a dish.}, +booktitle = {International Conference on Case-Based Reasoning}, +title = {CookingCAKE: A Framework for the adaptation of cooking recipes represented as workflows}, +keywords={recipe adaptation, workflow adaptation, workflows, process-oriented, case based reasoning} +} + +@Inproceedings{10.1007/978-3-319-24586-7_20, +author={Onta{\~{n}}{\'o}n, S. +and Plaza, E. +and Zhu, J.}, +editor={H{\"u}llermeier, Eyke +and Minor, Mirjam}, +affiliation={Drexel University; Artificial Intelligence Research Institute CSIC}, +title={Argument-Based Case Revision in CBR for Story Generation}, +keywords={CBR, Case-based reasoning, Story generation}, +booktitle={Case-Based Reasoning Research and Development}, +year={2015}, +publisher={Springer International Publishing}, +address={Cham}, +language = {English}, +pages={290--305}, +type = {article; proceedings paper}, +abstract={This paper presents a new approach to case revision in case-based reasoning based on the idea of argumentation. Previous work on case reuse has proposed the use of operations such as case amalgamation (or merging), which generate solutions by combining information coming from different cases. Such approaches are often based on exploring the search space of possible combinations looking for a solution that maximizes a certain criteria. We show how Revise can be performed by arguments attacking specific parts of a case produced by Reuse, and how they can guide and prevent repeating pitfalls in future cases. The proposed approach is evaluated in the task of automatic story generation.}, +isbn={978-3-319-24586-7} +} + +@Inproceedings{10.1007/978-3-030-58342-2_20, +author={Lepage, Yves +and Lieber, Jean +and Mornard, Isabelle +and Nauer, Emmanuel +and Romary, Julien +and Sies, Reynault}, +editor={Watson, Ian +and Weber, Rosina}, +title={The French Correction: When Retrieval Is Harder to Specify than Adaptation}, +affiliation={Waseda University; Université de Lorraine}, +keywords={case-based reasoning, retrieval, analogy, sentence correction}, +booktitle={Case-Based Reasoning Research and Development}, +year={2020}, +language = {English}, +type = {article; proceedings paper}, +publisher={Springer International Publishing}, +address={Cham}, +pages={309--324}, +abstract={A common idea in the field of case-based reasoning is that the retrieval step can be specified by the use of some similarity measure: the retrieved cases maximize the similarity to the target problem and, then, the adaptation step has to take into account the mismatches between the retrieved cases and the target problem in order to this latter. The use of this methodological schema for the application described in this paper has proven to be non efficient. Indeed, designing a retrieval procedure without the precise knowledge of the adaptation procedure has not been possible. The domain of this application is the correction of French sentences: a problem is an incorrect sentence and a valid solution is a correction of this problem. Adaptation consists in solving an analogical equation that enables to execute the correction of the retrieved case on the target problem. Thus, retrieval has to ensure that this application is feasible. The first version of such a retrieval procedure is described and evaluated: it is a knowledge-light procedure that does not use linguistic knowledge about French.}, +isbn={978-3-030-58342-2} +} + +@Inproceedings{10.1007/978-3-030-01081-2_25, +author={Smyth, Barry +and Cunningham, P{\'a}draig}, +editor={Cox, Michael T. +and Funk, Peter +and Begum, Shahina}, +affiliation={University College Dublin}, +title={An Analysis of Case Representations for Marathon Race Prediction and Planning}, +keywords={Marathon planning, Case representation, Case-based reasoning}, +booktitle={Case-Based Reasoning Research and Development}, +year={2018}, +language = {English}, +publisher={Springer International Publishing}, +address={Cham}, +pages={369--384}, +type = {article; proceedings paper}, +abstract={We use case-based reasoning to help marathoners achieve a personal best for an upcoming race, by helping them to select an achievable goal-time and a suitable pacing plan. We evaluate several case representations and, using real-world race data, highlight their performance implications. Richer representations do not always deliver better prediction performance, but certain representational configurations do offer very significant practical benefits for runners, when it comes to predicting, and planning for, challenging goal-times during an upcoming race.}, +isbn={978-3-030-01081-2} +} + +@Inproceedings{10.1007/978-3-030-58342-2_8, +author={Smyth, Barry +and Willemsen, Martijn C.}, +editor={Watson, Ian +and Weber, Rosina}, +affiliation={University College Dublin; Eindhoven University of Technology}, +title={Predicting the Personal-Best Times of Speed Skaters Using Case-Based Reasoning}, +keywords={CBR for health and exercise, speed skating, race-time prediction, case representation}, +booktitle={Case-Based Reasoning Research and Development}, +year={2020}, +type = {article; proceedings paper}, +language = {English}, +publisher={Springer International Publishing}, +address={Cham}, +pages={112--126}, +abstract={Speed skating is a form of ice skating in which the skaters race each other over a variety of standardised distances. Races take place on specialised ice-rinks and the type of track and ice conditions can have a significant impact on race-times. As race distances increase, pacing also plays an important role. In this paper we seek to extend recent work on the application of case-based reasoning to marathon-time prediction by predicting race-times for speed skaters. In particular, we propose and evaluate a number of case-based reasoning variants based on different case and feature representations to generate track-specific race predictions. We show it is possible to improve upon state-of-the-art prediction accuracy by harnessing richer case representations using shorter races and track-adjusted finish and lap-times.}, +isbn={978-3-030-58342-2} +} + +@Inproceedings{10.1007/978-3-030-58342-2_5, +author={Feely, Ciara +and Caulfield, Brian +and Lawlor, Aonghus +and Smyth, Barry}, +editor={Watson, Ian +and Weber, Rosina}, +affiliation={University College Dublin}, +title={Using Case-Based Reasoning to Predict Marathon Performance and Recommend Tailored Training Plans}, +keywords={CBR for health and exercise, marathon running, race-time prediction, plan recommendation}, +booktitle={Case-Based Reasoning Research and Development}, +year={2020}, +language = {English}, +publisher={Springer International Publishing}, +address={Cham}, +pages={67--81}, +type = {article; proceedings paper}, +abstract={Training for the marathon, especially a first marathon, is always a challenge. Many runners struggle to find the right balance between their workouts and their recovery, often leading to sub-optimal performance on race-day or even injury during training. We describe and evaluate a novel case-based reasoning system to help marathon runners as they train in two ways. First, it uses a case-base of training/workouts and race histories to predict future marathon times for a target runner, throughout their training program, helping runners to calibrate their progress and, ultimately, plan their race-day pacing. Second, the system recommends tailored training plans to runners, adapted for their current goal-time target, and based on the training plans of similar runners who have achieved this time. We evaluate the system using a dataset of more than 21,000 unique runners and 1.5 million training/workout sessions.}, +isbn={978-3-030-58342-2} +} + +@article{LALITHA2020583, +title = {Personalised Self-Directed Learning Recommendation System}, +journal = {Procedia Computer Science}, +volume = {171}, +pages = {583-592}, +year = {2020}, +type = {article}, +language = {English}, +note = {Third International Conference on Computing and Network Communications (CoCoNet'19)}, +issn = {1877-0509}, +doi = {https://doi.org/10.1016/j.procs.2020.04.063}, +url = {https://www.sciencedirect.com/science/article/pii/S1877050920310309}, +author = {T B Lalitha and P S Sreeja}, +affiliation={Hindustan Institute of Technology and Science}, +keywords = {e-Learning, PSDLR, Recommendation System, SDL, Self-Directed Learning}, +abstract = {Modern educational systems have changed drastically bringing in knowledge anywhere as needed by the learner with the evolution of Internet. Availability of knowledge in public domain, capability of exchanging large amount of information and filtering relevant information quickly has enabled disruption to conventional educational system. Thus, future trends are looking towards E-Learning (Electronic Learning) and M-Learning (Mobile Learning) technologies over the Internet for their vast knowledge acquisition. In this paper, the work gives an elaborate context of learning strategies prevailing and emerging with the classification of e-learning Techniques. It majorly focuses on the features and variety of aspects with the e-learning and the choice of learning method involved and facilitate the adoption of new ways for personalized selection on learning resources for SDL (Self-Directed Learning) from the unstructured, large web-based environment. Thereby, proposes a Personalised Self-Directed Learning Recommendation System (PSDLR) based on the personal specifications of the SDL learner. The result offers insight into the perspectives and challenges of Self-Directed Learning based on cognitive and constructive characteristics which majorly incorporates web-based learning and gives path in finding appropriate solutions using machine learning techniques and ontology for the open problems in the respective fields with personalised recommendations and guidelines for future research.} +} + +@article{Zhou2021, +author={Zhou, Lina +and Wang, Chunxia}, +affiliation={Baotou Medical College}, +title={Research on Recommendation of Personalized Exercises in English Learning Based on Data Mining}, +journal={Scientific Programming}, +year={2021}, +month={Dec}, +type = {article}, +language = {English}, +day={21}, +publisher={Hindawi}, +keywords={Recommender systems, Learning}, +volume={2021}, +pages={5042286}, +abstract={Aiming at the problems of traditional method of exercise recommendation precision, recall rate, long recommendation time, and poor recommendation comprehensiveness, this study proposes a personalized exercise recommendation method for English learning based on data mining. Firstly, a personalized recommendation model is designed, based on the model to preprocess the data in the Web access log, and cleaning the noise data to avoid its impact on the accuracy of the recommendation results is focused; secondly, the DINA model to diagnose the degree of mastery of students{\&}{\#}x2019; knowledge points is used and the students{\&}{\#}x2019; browsing patterns through fuzzy similar relationships are clustered; and finally, according to the clustering results, the similarity between students and the similarity between exercises are measured, and the collaborative filtering recommendation of personalized exercises for English learning is realized. The experimental results show that the exercise recommendation precision and recall rate of this method are higher, the recommendation time is shorter, and the recommendation results are comprehensive.}, +issn={1058-9244}, +doi={10.1155/2021/5042286}, +url={https://doi.org/10.1155/2021/5042286} +} + +@article{INGKAVARA2022100086, +title = {The use of a personalized learning approach to implementing self-regulated online learning}, +journal = {Computers and Education: Artificial Intelligence}, +volume = {3}, +pages = {100086}, +type = {article}, +language = {English}, +year = {2022}, +issn = {2666-920X}, +doi = {https://doi.org/10.1016/j.caeai.2022.100086}, +url = {https://www.sciencedirect.com/science/article/pii/S2666920X22000418}, +author = {Thanyaluck Ingkavara and Patcharin Panjaburee and Niwat Srisawasdi and Suthiporn Sajjapanroj}, +keywords = {Intelligent tutoring system, Personalization, Adaptive learning, E-learning, TAM, Artificial intelligence}, +abstract = {Nowadays, students are encouraged to learn via online learning systems to promote students' autonomy. Scholars have found that students' self-regulated actions impact their academic success in an online learning environment. However, because traditional online learning systems cannot personalize feedback to the student's personality, most students have less chance to obtain helpful suggestions for enhancing their knowledge linked to their learning problems. This paper incorporated self-regulated online learning in the Physics classroom and used a personalized learning approach to help students receive proper learning paths and material corresponding to their learning preferences. This study conducted a quasi-experimental design using a quantitative approach to evaluate the effectiveness of the proposed learning environment in secondary schools. The experimental group of students participated in self-regulated online learning with a personalized learning approach, while the control group participated in conventional self-regulated online learning. The experimental results showed that the experimental group's post-test and the learning-gain score of the experimental group were significantly higher than those of the control group. Moreover, the results also suggested that the student's perceptions about the usefulness of learning suggestions, ease of use, goal setting, learning environmental structuring, task strategies, time management, self-evaluation, impact on learning, and attitude toward the learning environment are important predictors of behavioral intention to learn with the self-regulated online learning that integrated with the personalized learning approach.} +} + +@article{HUANG2023104684, +title = {Effects of artificial Intelligence–Enabled personalized recommendations on learners’ learning engagement, motivation, and outcomes in a flipped classroom}, +journal = {Computers and Education}, +volume = {194}, +pages = {104684}, +year = {2023}, +language = {English}, +type = {article}, +issn = {0360-1315}, +doi = {https://doi.org/10.1016/j.compedu.2022.104684}, +url = {https://www.sciencedirect.com/science/article/pii/S036013152200255X}, +author = {Anna Y.Q. Huang and Owen H.T. Lu and Stephen J.H. Yang}, +keywords = {Data science applications in education, Distance education and online learning, Improving classroom teaching}, +abstract = {The flipped classroom approach is aimed at improving learning outcomes by promoting learning motivation and engagement. Recommendation systems can also be used to improve learning outcomes. With the rapid development of artificial intelligence (AI) technology, various systems have been developed to facilitate student learning. Accordingly, we applied AI-enabled personalized video recommendations to stimulate students' learning motivation and engagement during a systems programming course in a flipped classroom setting. We assigned students to control and experimental groups comprising 59 and 43 college students, respectively. The students in both groups received flipped classroom instruction, but only those in the experimental group received AI-enabled personalized video recommendations. We quantitatively measured students’ engagement based on their learning profiles in a learning management system. The results revealed that the AI-enabled personalized video recommendations could significantly improve the learning performance and engagement of students with a moderate motivation level.} +} + +@article{ZHAO2023118535, +title = {A recommendation system for effective learning strategies: An integrated approach using context-dependent DEA}, +journal = {Expert Systems with Applications}, +volume = {211}, +pages = {118535}, +year = {2023}, +language = {English}, +type = {article}, +issn = {0957-4174}, +doi = {https://doi.org/10.1016/j.eswa.2022.118535}, +url = {https://www.sciencedirect.com/science/article/pii/S0957417422016104}, +author = {Lu-Tao Zhao and Dai-Song Wang and Feng-Yun Liang and Jian Chen}, +keywords = {Recommendation system, Learning strategies, Context-dependent DEA, Efficiency analysis}, +abstract = {Universities have been focusing on increasing individualized training and providing appropriate education for students. The individual differences and learning needs of college students should be given enough attention. From the perspective of learning efficiency, we establish a clustering hierarchical progressive improvement model (CHPI), which is based on cluster analysis and context-dependent data envelopment analysis (DEA) methods. The CHPI clusters students' ontological features, employs the context-dependent DEA method to stratify students of different classes, and calculates measures, such as obstacles, to determine the reference path for individuals with inefficient learning processes. The learning strategies are determined according to the gap between the inefficient individual to be improved and the individuals on the reference path. By the study of college English courses as an example, it is found that the CHPI can accurately recommend targeted learning strategies to satisfy the individual needs of college students so that the learning of individuals with inefficient learning processes in a certain stage can be effectively improved. In addition, CHPI can provide specific, efficient suggestions to improve learning efficiency comparing to existing recommendation systems, and has great potential in promoting the integration of education-related researches and expert systems.} +} + +@article{SU2022109547, +title = {Graph-based cognitive diagnosis for intelligent tutoring systems}, +journal = {Knowledge-Based Systems}, +volume = {253}, +pages = {109547}, +year = {2022}, +language = {English}, +type = {article}, +issn = {0950-7051}, +doi = {https://doi.org/10.1016/j.knosys.2022.109547}, +url = {https://www.sciencedirect.com/science/article/pii/S095070512200778X}, +author = {Yu Su and Zeyu Cheng and Jinze Wu and Yanmin Dong and Zhenya Huang and Le Wu and Enhong Chen and Shijin Wang and Fei Xie}, +keywords = {Cognitive diagnosis, Graph neural networks, Interpretable machine learning}, +abstract = {For intelligent tutoring systems, Cognitive Diagnosis (CD) is a fundamental task that aims to estimate the mastery degree of a student on each skill according to the exercise record. The CD task is considered rather challenging since we need to model inner-relations and inter-relations among students, skills, and questions to obtain more abundant information. Most existing methods attempt to solve this problem through two-way interactions between students and questions (or between students and skills), ignoring potential high-order relations among entities. Furthermore, how to construct an end-to-end framework that can model the complex interactions among different types of entities at the same time remains unexplored. Therefore, in this paper, we propose a graph-based Cognitive Diagnosis model (GCDM) that directly discovers the interactions among students, skills, and questions through a heterogeneous cognitive graph. Specifically, we design two graph-based layers: a performance-relative propagator and an attentive knowledge aggregator. The former is applied to propagate a student’s cognitive state through different types of graph edges, while the latter selectively gathers messages from neighboring graph nodes. Extensive experimental results on two real-world datasets clearly show the effectiveness and extendibility of our proposed model.} +} + +@article{EZALDEEN2022100700, +title = {A hybrid E-learning recommendation integrating adaptive profiling and sentiment analysis}, +journal = {Journal of Web Semantics}, +volume = {72}, +pages = {100700}, +year = {2022}, +type = {article}, +language = {English}, +issn = {1570-8268}, +doi = {https://doi.org/10.1016/j.websem.2021.100700}, +url = {https://www.sciencedirect.com/science/article/pii/S1570826821000664}, +author = {Hadi Ezaldeen and Rachita Misra and Sukant Kishoro Bisoy and Rawaa Alatrash and Rojalina Priyadarshini}, +keywords = {Hybrid E-learning recommendation, Adaptive profiling, Semantic learner profile, Fine-grained sentiment analysis, Convolutional Neural Network, Word embeddings}, +abstract = {This research proposes a novel framework named Enhanced e-Learning Hybrid Recommender System (ELHRS) that provides an appropriate e-content with the highest predicted ratings corresponding to the learner’s particular needs. To accomplish this, a new model is developed to deduce the Semantic Learner Profile automatically. It adaptively associates the learning patterns and rules depending on the learner’s behavior and the semantic relations computed in the semantic matrix that mutually links e-learning materials and terms. Here, a semantic-based approach for term expansion is introduced using DBpedia and WordNet ontologies. Further, various sentiment analysis models are proposed and incorporated as a part of the recommender system to predict ratings of e-learning resources from posted text reviews utilizing fine-grained sentiment classification on five discrete classes. Qualitative Natural Language Processing (NLP) methods with tailored-made Convolutional Neural Network (CNN) are developed and evaluated on our customized dataset collected for a specific domain and a public dataset. Two improved language models are introduced depending on Skip-Gram (S-G) and Continuous Bag of Words (CBOW) techniques. In addition, a robust language model based on hybridization of these couple of methods is developed to derive better vocabulary representation, yielding better accuracy 89.1% for the CNN-Three-Channel-Concatenation model. The suggested recommendation methodology depends on the learner’s preferences, other similar learners’ experience and background, deriving their opinions from the reviews towards the best learning resources. This assists the learners in finding the desired e-content at the proper time.} +} + +@article{MUANGPRATHUB2020e05227, +title = {Learning recommendation with formal concept analysis for intelligent tutoring system}, +journal = {Heliyon}, +volume = {6}, +number = {10}, +pages = {e05227}, +language = {English}, +type = {article}, +year = {2020}, +issn = {2405-8440}, +doi = {https://doi.org/10.1016/j.heliyon.2020.e05227}, +url = {https://www.sciencedirect.com/science/article/pii/S2405844020320703}, +author = {Jirapond Muangprathub and Veera Boonjing and Kosin Chamnongthai}, +keywords = {Computer Science, Learning recommendation, Formal concept analysis, Intelligent tutoring system, Adaptive learning}, +abstract = {The aim of this research was to develop a learning recommendation component in an intelligent tutoring system (ITS) that dynamically predicts and adapts to a learner's style. In order to develop a proper ITS, we present an improved knowledge base supporting adaptive learning, which can be achieved by a suitable knowledge construction. This process is illustrated by implementing a web-based online tutor system. In addition, our knowledge structure provides adaptive presentation and personalized learning with the proposed adaptive algorithm, to retrieve content according to individual learner characteristics. To demonstrate the proposed adaptive algorithm, pre-test and post-test were used to evaluate suggestion accuracy of the course in a class for adapting to a learner's style. In addition, pre- and post-testing were also used with students in a real teaching/learning environment to evaluate the performance of the proposed model. The results show that the proposed system can be used to help students or learners achieve improved learning.} +} + +@article{min8100434, +author = {Leikola, Maria and Sauer, Christian and Rintala, Lotta and Aromaa, Jari and Lundström, Mari}, +title = {Assessing the Similarity of Cyanide-Free Gold Leaching Processes: A Case-Based Reasoning Application}, +journal = {Minerals}, +volume = {8}, +type = {article}, +language = {English}, +year = {2018}, +number = {10}, +url = {https://www.mdpi.com/2075-163X/8/10/434}, +issn = {2075-163X}, +keywords={hydrometallurgy, cyanide-free gold, knowledge modelling, case-based reasoning, information retrieval}, +abstract = {Hydrometallurgical researchers, and other professionals alike, invest significant amounts of time reading scientific articles, technical notes, and other scientific documents, while looking for the most relevant information for their particular research interest. In an attempt to save the researcher’s time, this study presents an information retrieval tool using case-based reasoning. The tool was built for comparing scientific articles concerning cyanide-free leaching of gold ores/concentrates/tailings. Altogether, 50 cases of experiments were gathered in a case base. 15 different attributes related to the treatment of the raw material and the leaching conditions were selected to compare the cases. The attributes were as follows: Pretreatment, Overall method, Complexant source, Oxidant source, Complexant concentration, Oxidant concentration, Temperature, pH, Redox-potential, Pressure, Materials of construction, Extraction, Extraction rate, Reagent consumption, and Solid-liquid ratio. The resulting retrieval tool (LeachSim) was able to rank the scientific articles according to their similarity with the user’s research interest. Such a tool could eventually aid the user in finding the most relevant information, but not replace thorough understanding and human expertise.}, +doi = {10.3390/min8100434} +} + +@article{10.1145/3459665, +author = {Cunningham, P\'{a}draig and Delany, Sarah Jane}, +title = {K-Nearest Neighbour Classifiers - A Tutorial}, +year = {2021}, +issue_date = {July 2022}, +publisher = {Association for Computing Machinery}, +address = {New York, NY, USA}, +type={article}, +language={English}, +volume = {54}, +number = {6}, +issn = {0360-0300}, +url = {https://doi.org/10.1145/3459665}, +doi = {10.1145/3459665}, +abstract = {Perhaps the most straightforward classifier in the arsenal or Machine Learning techniques is the Nearest Neighbour Classifier—classification is achieved by identifying the nearest neighbours to a query example and using those neighbours to determine the class of the query. This approach to classification is of particular importance, because issues of poor runtime performance is not such a problem these days with the computational power that is available. This article presents an overview of techniques for Nearest Neighbour classification focusing on: mechanisms for assessing similarity (distance), computational issues in identifying nearest neighbours, and mechanisms for reducing the dimension of the data.This article is the second edition of a paper previously published as a technical report [16]. Sections on similarity measures for time-series, retrieval speedup, and intrinsic dimensionality have been added. An Appendix is included, providing access to Python code for the key methods.}, +journal = {ACM Comput. Surv.}, +month = {jul}, +articleno = {128}, +numpages = {25}, +keywords = {k-Nearest neighbour classifiers} +} + +@article{9072123, +author={Sinaga, Kristina P. and Yang, Miin-Shen}, +journal={IEEE Access}, +type={article}, +language={English}, +title={Unsupervised K-Means Clustering Algorithm}, +year={2020}, +volume={8}, +number={}, +pages={80716-80727}, +doi={10.1109/ACCESS.2020.2988796} +} + +@article{WANG2021331, +title = {A new prediction strategy for dynamic multi-objective optimization using Gaussian Mixture Model}, +journal = {Information Sciences}, +volume = {580}, +type = {article}, +language = {English}, +pages = {331-351}, +year = {2021}, +issn = {0020-0255}, +doi = {https://doi.org/10.1016/j.ins.2021.08.065}, +url = {https://www.sciencedirect.com/science/article/pii/S0020025521008732}, +author = {Feng Wang and Fanshu Liao and Yixuan Li and Hui Wang}, +keywords = {Dynamic multi-objective optimization, Gaussian Mixture Model, Change type detection, Resampling}, +abstract = {Dynamic multi-objective optimization problems (DMOPs), in which the environments change over time, have attracted many researchers’ attention in recent years. Since the Pareto set (PS) or the Pareto front (PF) can change over time, how to track the movement of the PS or PF is a challenging problem in DMOPs. Over the past few years, lots of methods have been proposed, and the prediction based strategy has been considered the most effective way to track the new PS. However, the performance of most existing prediction strategies depends greatly on the quantity and quality of the historical information and will deteriorate due to non-linear changes, leading to poor results. In this paper, we propose a new prediction method, named MOEA/D-GMM, which incorporates the Gaussian Mixture Model (GMM) into the MOEA/D framework for the prediction of the new PS when changes occur. Since GMM is a powerful non-linear model to accurately fit various data distributions, it can effectively generate solutions with better quality according to the distributions. In the proposed algorithm, a change type detection strategy is first designed to estimate an approximate PS according to different change types. Then, GMM is employed to make a more accurate prediction by training it with the approximate PS. To overcome the shortcoming of a lack of training solutions for GMM, the Empirical Cumulative Distribution Function (ECDF) method is used to resample more training solutions before GMM training. Experimental results on various benchmark test problems and a classical real-world problem show that, compared with some state-of-the-art dynamic optimization algorithms, MOEA/D-GMM outperforms others in most cases.} +} + +@article{9627973, +author={Xu, Shengbing and Cai, Wei and Xia, Hongxi and Liu, Bo and Xu, Jie}, +journal={IEEE Access}, +title={Dynamic Metric Accelerated Method for Fuzzy Clustering}, +year={2021}, +type={article}, +language={English}, +volume={9}, +number={}, +pages={166838-166854}, +doi={10.1109/ACCESS.2021.3131368} +} + +@article{9434422, +author={Gupta, Samarth and Chaudhari, Shreyas and Joshi, Gauri and Yağan, Osman}, +journal={IEEE Transactions on Information Theory}, +title={Multi-Armed Bandits With Correlated Arms}, +year={2021}, +language={English}, +type={article}, +volume={67}, +number={10}, +pages={6711-6732}, +doi={10.1109/TIT.2021.3081508} +} + +@Inproceedings{8495930, +author={Supic, H.}, +booktitle={2018 IEEE 27th International Conference on Enabling Technologies: Infrastructure for Collaborative Enterprises (WETICE)}, +title={Case-Based Reasoning Model for Personalized Learning Path Recommendation in Example-Based Learning Activities}, +year={2018}, +type={article}, +language={English}, +volume={}, +number={}, +pages={175-178}, +doi={10.1109/WETICE.2018.00040} +} + +@Inproceedings{9870279, +author={Lin, Baihan}, +booktitle={2022 IEEE Congress on Evolutionary Computation (CEC)}, +title={Evolutionary Multi-Armed Bandits with Genetic Thompson Sampling}, +year={2022}, +type={article}, +language={English}, +volume={}, +number={}, +pages={1-8}, +doi={10.1109/CEC55065.2022.9870279} +} + +@article{Obeid, +author={Obeid, C. and Lahoud, C. and Khoury, H. E. and Champin, P.}, +title={A Novel Hybrid Recommender System Approach for Student Academic Advising Named COHRS, Supported by Case-based Reasoning and Ontology}, +journal={Computer Science and Information Systems}, +type={article}, +language={English}, +volume={19}, +number={2}, +pages={979–1005}, +year={2022}, +doi={https://doi.org/10.2298/CSIS220215011O} +} + +@book{Nkambou, +author = {Nkambou, R. and Bourdeau, J. and Mizoguchi, R.}, +title = {Advances in Intelligent Tutoring Systems}, +year = {2010}, +type = {article}, +language = {English}, +publisher = {Springer Berlin, Heidelberg}, +edition = {1} +} + +@book{hajduk2019cognitive, +title={Cognitive Multi-agent Systems: Structures, Strategies and Applications to Mobile Robotics and Robosoccer}, +author={Hajduk, M. and Sukop, M. and Haun, M.}, +type={book}, +language={English}, +isbn={9783319936857}, +series={Studies in Systems, Decision and Control}, +year={2019}, +publisher={Springer International Publishing} +} + +@article{RICHTER20093, +title = {The search for knowledge, contexts, and Case-Based Reasoning}, +journal = {Engineering Applications of Artificial Intelligence}, +language = {English}, +type = {article}, +volume = {22}, +number = {1}, +pages = {3-9}, +year = {2009}, +issn = {0952-1976}, +doi = {https://doi.org/10.1016/j.engappai.2008.04.021}, +url = {https://www.sciencedirect.com/science/article/pii/S095219760800078X}, +author = {Michael M. Richter}, +keywords = {Case-Based Reasoning, Knowledge, Processes, Utility, Context}, +abstract = {A major goal of this paper is to compare Case-Based Reasoning with other methods searching for knowledge. We consider knowledge as a resource that can be traded. It has no value in itself; the value is measured by the usefulness of applying it in some process. Such a process has info-needs that have to be satisfied. The concept to measure this is the economical term utility. In general, utility depends on the user and its context, i.e., it is subjective. Here, we introduce levels of contexts from general to individual. We illustrate that Case-Based Reasoning on the lower, i.e., more personal levels CBR is quite useful, in particular in comparison with traditional informational retrieval methods.} +} + +@Thesis{Marie, +author={Marie, F.}, +title={COLISEUM-3D. Une plate-forme innovante pour la segmentation d’images médicales par Raisonnement à Partir de Cas (RàPC) et méthodes d’apprentissage de type Deep Learning}, +type={diplomathesis}, +language={French}, +institution={Université de Franche-Comte}, +year={2019} +} + +@book{Hoang, +title = {La formule du savoir. Une philosophie unifiée du savoir fondée sur le théorème de Bayes}, +author = {Hoang, L.N.}, +type = {book}, +language = {French}, +isbn = {9782759822607}, +year = {2018}, +publisher = {EDP Sciences} +} + +@book{Richter2013, +title={Case-Based Reasoning (A Textbook)}, +author={Richter, M. and Weber, R.}, +type={book}, +language={English}, +isbn={9783642401664}, +year={2013}, +publisher={Springer-Verlag GmbH} +} + +@book{kedia2020hands, +title={Hands-On Python Natural Language Processing: Explore tools and techniques to analyze and process text with a view to building real-world NLP applications}, +author={Kedia, A. and Rasu, M.}, +language={English}, +type={book}, +isbn={9781838982584}, +url={https://books.google.fr/books?id=1AbuDwAAQBAJ}, +year={2020}, +publisher={Packt Publishing} +} + +@book{ghosh2019natural, +title={Natural Language Processing Fundamentals: Build intelligent applications that can interpret the human language to deliver impactful results}, +author={Ghosh, S. and Gunning, D.}, +language={English}, +type={book}, +isbn={9781789955989}, +url={https://books.google.fr/books?id=i8-PDwAAQBAJ}, +year={2019}, +publisher={Packt Publishing} +} + +@article{Akerblom, +title={Online learning of network bottlenecks via minimax paths}, +author={kerblom, Niklas and Hoseini, Fazeleh Sadat and Haghir Chehreghani, Morteza}, +language={English}, +type={article}, +volume = {122}, +year = {2023}, +issn = {1573-0565}, +doi = {https://doi.org/10.1007/s10994-022-06270-0}, +url = {https://doi.org/10.1007/s10994-022-06270-0}, +abstract={In this paper, we study bottleneck identification in networks via extracting minimax paths. Many real-world networks have stochastic weights for which full knowledge is not available in advance. Therefore, we model this task as a combinatorial semi-bandit problem to which we apply a combinatorial version of Thompson Sampling and establish an upper bound on the corresponding Bayesian regret. Due to the computational intractability of the problem, we then devise an alternative problem formulation which approximates the original objective. Finally, we experimentally evaluate the performance of Thompson Sampling with the approximate formulation on real-world directed and undirected networks.} +} + +@article{Simen, +title={Dynamic slate recommendation with gated recurrent units and Thompson sampling}, +author={Eide, Simen and Leslie, David S. and Frigessi, Arnoldo}, +language={English}, +type={article}, +volume = {36}, +year = {2022}, +issn = {1573-756X}, +doi = {https://doi.org/10.1007/s10618-022-00849-w}, +url = {https://doi.org/10.1007/s10618-022-00849-w}, +abstract={We consider the problem of recommending relevant content to users of an internet platform in the form of lists of items, called slates. We introduce a variational Bayesian Recurrent Neural Net recommender system that acts on time series of interactions between the internet platform and the user, and which scales to real world industrial situations. The recommender system is tested both online on real users, and on an offline dataset collected from a Norwegian web-based marketplace, FINN.no, that is made public for research. This is one of the first publicly available datasets which includes all the slates that are presented to users as well as which items (if any) in the slates were clicked on. Such a data set allows us to move beyond the common assumption that implicitly assumes that users are considering all possible items at each interaction. Instead we build our likelihood using the items that are actually in the slate, and evaluate the strengths and weaknesses of both approaches theoretically and in experiments. We also introduce a hierarchical prior for the item parameters based on group memberships. Both item parameters and user preferences are learned probabilistically. Furthermore, we combine our model with bandit strategies to ensure learning, and introduce ‘in-slate Thompson sampling’ which makes use of the slates to maximise explorative opportunities. We show experimentally that explorative recommender strategies perform on par or above their greedy counterparts. Even without making use of exploration to learn more effectively, click rates increase simply because of improved diversity in the recommended slates.} +} + +@Inproceedings{Arthurs, +author={Arthurs, Noah and Stenhaug, Ben and Karayev, Sergey and Piech, Chris}, +booktitle={International Conference on Educational Data Mining (EDM)}, +title={Grades Are Not Normal: Improving Exam Score Models Using the Logit-Normal Distribution}, +year={2019}, +type={article}, +language={English}, +volume={}, +number={}, +pages={6}, +url={https://eric.ed.gov/?id=ED599204} +} + +@article{Bahramian, +title={A Cold Start Context-Aware Recommender System for Tour Planning Using Artificial Neural Network and Case Based Reasoning}, +author={Bahramian, Zahra and Ali Abbaspour, Rahim and Claramunt, Christophe}, +language={English}, +type={article}, +year = {2017}, +issn = {1574-017X}, +doi = {https://doi.org/10.1155/2017/9364903}, +url = {https://doi.org/10.1155/2017/9364903}, +abstract={Nowadays, large amounts of tourism information and services are available over the Web. This makes it difficult for the user to search for some specific information such as selecting a tour in a given city as an ordered set of points of interest. Moreover, the user rarely knows all his needs upfront and his preferences may change during a recommendation process. The user may also have a limited number of initial ratings and most often the recommender system is likely to face the well-known cold start problem. The objective of the research presented in this paper is to introduce a hybrid interactive context-aware tourism recommender system that takes into account user’s feedbacks and additional contextual information. It offers personalized tours to the user based on his preferences thanks to the combination of a case based reasoning framework and an artificial neural network. The proposed method has been tried in the city of Tehran in Iran. The results show that the proposed method outperforms current artificial neural network methods and combinations of case based reasoning with -nearest neighbor methods in terms of user effort, accuracy, and user satisfaction.} +} + +@Thesis{Daubias2011, +author={Sthéphanie Jean-Daubias}, +title={Ingénierie des profils d'apprenants}, +type={diplomathesis}, +language={French}, +institution={Université Claude Bernard Lyon 1}, +year={2011} +} + +@article{Tapalova, +author = {Olga Tapalova and Nadezhda Zhiyenbayeva}, +title ={Artificial Intelligence in Education: AIEd for Personalised Learning Pathways}, +journal = {Electronic Journal of e-Learning}, +volume = {}, +number = {}, +pages = {15}, +year = {2022}, +URL = {https://eric.ed.gov/?q=Artificial+Intelligence+in+Education%3a+AIEd+for+Personalised+Learning+Pathways&id=EJ1373006}, +language={English}, +type={article}, +abstract = {Artificial intelligence is the driving force of change focusing on the needs and demands of the student. The research explores Artificial Intelligence in Education (AIEd) for building personalised learning systems for students. The research investigates and proposes a framework for AIEd: social networking sites and chatbots, expert systems for education, intelligent mentors and agents, machine learning, personalised educational systems and virtual educational environments. These technologies help educators to develop and introduce personalised approaches to master new knowledge and develop professional competencies. The research presents a case study of AIEd implementation in education. The scholars conducted the experiment in educational establishments using artificial intelligence in the curriculum. The scholars surveyed 184 second-year students of the Institute of Pedagogy and Psychology at the Abay Kazakh National Pedagogical University and the Kuban State Technological University to collect the data. The scholars considered the collective group discussions regarding the application of artificial intelligence in education to improve the effectiveness of learning. The research identified key advantages to creating personalised learning pathways such as access to training in 24/7 mode, training in virtual contexts, adaptation of educational content to personal needs of students, real-time and regular feedback, improvements in the educational process and mental stimulations. The proposed education paradigm reflects the increasing role of artificial intelligence in socio-economic life, the social and ethical concerns artificial intelligence may pose to humanity and its role in the digitalisation of education. The current article may be used as a theoretical framework for many educational institutions planning to exploit the capabilities of artificial intelligence in their adaptation to personalized learning.} +} + +@article{Auer, +title = {From monolithic systems to Microservices: An assessment framework}, +journal = {Information and Software Technology}, +volume = {137}, +pages = {106600}, +year = {2021}, +issn = {0950-5849}, +doi = {https://doi.org/10.1016/j.infsof.2021.106600}, +url = {https://www.sciencedirect.com/science/article/pii/S0950584921000793}, +author = {Florian Auer and Valentina Lenarduzzi and Michael Felderer and Davide Taibi}, +keywords = {Microservices, Cloud migration, Software measurement}, +abstract = {Context: +Re-architecting monolithic systems with Microservices-based architecture is a common trend. Various companies are migrating to Microservices for different reasons. However, making such an important decision like re-architecting an entire system must be based on real facts and not only on gut feelings. +Objective: +The goal of this work is to propose an evidence-based decision support framework for companies that need to migrate to Microservices, based on the analysis of a set of characteristics and metrics they should collect before re-architecting their monolithic system. +Method: +We conducted a survey done in the form of interviews with professionals to derive the assessment framework based on Grounded Theory. +Results: +We identified a set consisting of information and metrics that companies can use to decide whether to migrate to Microservices or not. The proposed assessment framework, based on the aforementioned metrics, could be useful for companies if they need to migrate to Microservices and do not want to run the risk of failing to consider some important information.} +} + +@Article{jmse10040464, +AUTHOR = {Zuluaga, Carlos A. and Aristizábal, Luis M. and Rúa, Santiago and Franco, Diego A. and Osorio, Dorie A. and Vásquez, Rafael E.}, +TITLE = {Development of a Modular Software Architecture for Underwater Vehicles Using Systems Engineering}, +JOURNAL = {Journal of Marine Science and Engineering}, +VOLUME = {10}, +YEAR = {2022}, +NUMBER = {4}, +ARTICLE-NUMBER = {464}, +URL = {https://www.mdpi.com/2077-1312/10/4/464}, +ISSN = {2077-1312}, +ABSTRACT = {This paper addresses the development of a modular software architecture for the design/construction/operation of a remotely operated vehicle (ROV), based on systems engineering. First, systems engineering and the Vee model are presented with the objective of defining the interactions of the stakeholders with the software architecture development team and establishing the baselines that must be met in each development phase. In the development stage, the definition of the architecture and its connection with the hardware is presented, taking into account the use of the actor model, which represents the high-level software architecture used to solve concurrency problems. Subsequently, the structure of the classes is defined both at high and low levels in the instruments using the object-oriented programming paradigm. Finally, unit tests are developed for each component in the software architecture, quality assessment tests are implemented for system functions fulfillment, and a field sea trial for testing different modules of the vehicle is described. This approach is well suited for the development of complex systems such as marine vehicles and those systems which require scalability and modularity to add functionalities.}, +DOI = {10.3390/jmse10040464} +} + +@article{doi:10.1177/1754337116651013, +author = {Julien Henriet and Lang Christophe and Philippe Laurent}, +title ={Artificial Intelligence-Virtual Trainer: An educative system based on artificial intelligence and designed to produce varied and consistent training lessons}, +journal = {Proceedings of the Institution of Mechanical Engineers, Part P: Journal of Sports Engineering and Technology}, +volume = {231}, +number = {2}, +pages = {110-124}, +year = {2017}, +doi = {10.1177/1754337116651013}, +URL = {https://doi.org/10.1177/1754337116651013}, +eprint = {https://doi.org/10.1177/1754337116651013}, +abstract = { AI-Virtual Trainer is an educative system using Artificial Intelligence to propose varied lessons to trainers. The agents of this multi-agent system apply case-based reasoning to build solutions by analogy. However, as required by the field, Artificial Intelligence-Virtual Trainer never proposes the same lesson twice, whereas the same objective may be set many times consecutively. The adaptation process of Artificial Intelligence-Virtual Trainer delivers an ordered set of exercises adapted to the objectives and sub-objectives chosen by trainers. This process has been enriched by including the notion of distance between exercises: the proposed tasks are not only appropriate but are hierarchically ordered. With this new version of the system, students are guided towards their objectives via an underlying theme. Finally, the agents responsible for the different parts of lessons collaborate with each other according to a dedicated protocol and decision-making policy since no exercise must appear more than once in the same lesson. The results prove that Artificial Intelligence-Virtual Trainer, however perfectible, meets the requirements of this field. } +} + +@InProceedings{10.1007/978-3-030-01081-2_9, +author="Henriet, Julien +and Greffier, Fran{\c{c}}oise", +editor="Cox, Michael T. +and Funk, Peter +and Begum, Shahina", +title="AI-VT: An Example of CBR that Generates a Variety of Solutions to the Same Problem", +booktitle="Case-Based Reasoning Research and Development", +year="2018", +publisher="Springer International Publishing", +address="Cham", +pages="124--139", +abstract="AI-Virtual Trainer (AI-VT) is an intelligent tutoring system based on case-based reasoning. AI-VT has been designed to generate personalised, varied, and consistent training sessions for learners. The AI-VT training sessions propose different exercises in regard to a capacity associated with sub-capacities. For example, in the field of training for algorithms, a capacity could be ``Use a control structure alternative'' and an associated sub-capacity could be ``Write a boolean condition''. AI-VT can elaborate a personalised list of exercises for each learner. One of the main requirements and challenges studied in this work is its ability to propose varied training sessions to the same learner for many weeks, which constitutes the challenge studied in our work. Indeed, if the same set of exercises is proposed time after time to learners, they will stop paying attention and lose motivation. Thus, even if the generation of training sessions is based on analogy and must integrate the repetition of some exercises, it also must introduce some diversity and AI-VT must deal with this diversity. In this paper, we have highlighted the fact that the retaining (or capitalisation) phase of CBR is of the utmost importance for diversity, and we have also highlighted that the equilibrium between repetition and variety depends on the abilities learned. This balance has an important impact on the retaining phase of AI-VT.", +isbn="978-3-030-01081-2" +} + +@article{BAKUROV2021100913, +title = {Genetic programming for stacked generalization}, +journal = {Swarm and Evolutionary Computation}, +volume = {65}, +pages = {100913}, +year = {2021}, +issn = {2210-6502}, +doi = {https://doi.org/10.1016/j.swevo.2021.100913}, +url = {https://www.sciencedirect.com/science/article/pii/S2210650221000742}, +author = {Illya Bakurov and Mauro Castelli and Olivier Gau and Francesco Fontanella and Leonardo Vanneschi}, +keywords = {Genetic Programming, Stacking, Ensemble Learning, Stacked Generalization}, +abstract = {In machine learning, ensemble techniques are widely used to improve the performance of both classification and regression systems. They combine the models generated by different learning algorithms, typically trained on different data subsets or with different parameters, to obtain more accurate models. Ensemble strategies range from simple voting rules to more complex and effective stacked approaches. They are based on adopting a meta-learner, i.e. a further learning algorithm, and are trained on the predictions provided by the single algorithms making up the ensemble. The paper aims at exploiting some of the most recent genetic programming advances in the context of stacked generalization. In particular, we investigate how the evolutionary demes despeciation initialization technique, ϵ-lexicase selection, geometric-semantic operators, and semantic stopping criterion, can be effectively used to improve GP-based systems’ performance for stacked generalization (a.k.a. stacking). The experiments, performed on a broad set of synthetic and real-world regression problems, confirm the effectiveness of the proposed approach.} +} + +@article{Liang, +author={Liang Mang and Chang Tianpeng and An Bingxing and Duan Xinghai and Du Lili and Wang Xiaoqiao and Miao Jian and Xu Lingyang and Gao Xue and Zhang Lupei and Li Junya and Gao Huijiang}, +Title={A Stacking Ensemble Learning Framework for Genomic Prediction}, +Journal={Frontiers in Genetics}, +year={2021}, +doi ={10.3389/fgene.2021.600040}, +PMID={33747037}, +PMCID={PMC7969712} +} + +@Article{cmc.2023.033417, +AUTHOR = {Jeonghoon Choi and Dongjun Suh and Marc-Oliver Otto}, +TITLE = {Boosted Stacking Ensemble Machine Learning Method for Wafer Map Pattern Classification}, +JOURNAL = {Computers, Materials \& Continua}, +VOLUME = {74}, +YEAR = {2023}, +NUMBER = {2}, +PAGES = {2945--2966}, +URL = {http://www.techscience.com/cmc/v74n2/50296}, +ISSN = {1546-2226}, +ABSTRACT = {Recently, machine learning-based technologies have been developed to automate the classification of wafer map defect patterns during semiconductor manufacturing. The existing approaches used in the wafer map pattern classification include directly learning the image through a convolution neural network and applying the ensemble method after extracting image features. This study aims to classify wafer map defects more effectively and derive robust algorithms even for datasets with insufficient defect patterns. First, the number of defects during the actual process may be limited. Therefore, insufficient data are generated using convolutional auto-encoder (CAE), and the expanded data are verified using the evaluation technique of structural similarity index measure (SSIM). After extracting handcrafted features, a boosted stacking ensemble model that integrates the four base-level classifiers with the extreme gradient boosting classifier as a meta-level classifier is designed and built for training the model based on the expanded data for final prediction. Since the proposed algorithm shows better performance than those of existing ensemble classifiers even for insufficient defect patterns, the results of this study will contribute to improving the product quality and yield of the actual semiconductor manufacturing process.}, +DOI = {10.32604/cmc.2023.033417} +} + +@ARTICLE{10.3389/fgene.2021.600040, +AUTHOR={Liang, Mang and Chang, Tianpeng and An, Bingxing and Duan, Xinghai and Du, Lili and Wang, Xiaoqiao and Miao, Jian and Xu, Lingyang and Gao, Xue and Zhang, Lupei and Li, Junya and Gao, Huijiang}, +TITLE={A Stacking Ensemble Learning Framework for Genomic Prediction}, +JOURNAL={Frontiers in Genetics}, +VOLUME={12}, +YEAR={2021}, +URL={https://www.frontiersin.org/articles/10.3389/fgene.2021.600040}, +DOI={10.3389/fgene.2021.600040}, +ISSN={1664-8021}, +ABSTRACT={Machine learning (ML) is perhaps the most useful tool for the interpretation of large genomic datasets. However, the performance of a single machine learning method in genomic selection (GS) is currently unsatisfactory. To improve the genomic predictions, we constructed a stacking ensemble learning framework (SELF), integrating three machine learning methods, to predict genomic estimated breeding values (GEBVs). The present study evaluated the prediction ability of SELF by analyzing three real datasets, with different genetic architecture; comparing the prediction accuracy of SELF, base learners, genomic best linear unbiased prediction (GBLUP) and BayesB. For each trait, SELF performed better than base learners, which included support vector regression (SVR), kernel ridge regression (KRR) and elastic net (ENET). The prediction accuracy of SELF was, on average, 7.70% higher than GBLUP in three datasets. Except for the milk fat percentage (MFP) traits, of the German Holstein dairy cattle dataset, SELF was more robust than BayesB in all remaining traits. Therefore, we believed that SEFL has the potential to be promoted to estimate GEBVs in other animals and plants.} +} + +@article{DIDDEN2023338, +title = {Decentralized learning multi-agent system for online machine shop scheduling problem}, +journal = {Journal of Manufacturing Systems}, +volume = {67}, +pages = {338-360}, +year = {2023}, +issn = {0278-6125}, +doi = {https://doi.org/10.1016/j.jmsy.2023.02.004}, +url = {https://www.sciencedirect.com/science/article/pii/S0278612523000286}, +author = {Jeroen B.H.C. Didden and Quang-Vinh Dang and Ivo J.B.F. Adan}, +keywords = {Multi-agent system, Decentralized systems, Learning algorithm, Industry 4.0, Smart manufacturing}, +abstract = {Customer profiles have rapidly changed over the past few years, with products being requested with more customization and with lower demand. In addition to the advances in technologies owing to Industry 4.0, manufacturers explore autonomous and smart factories. This paper proposes a decentralized multi-agent system (MAS), including intelligent agents that can respond to their environment autonomously through learning capabilities, to cope with an online machine shop scheduling problem. In the proposed system, agents participate in auctions to receive jobs to process, learn how to bid for jobs correctly, and decide when to start processing a job. The objective is to minimize the mean weighted tardiness of all jobs. In contrast to the existing literature, the proposed MAS is assessed on its learning capabilities, producing novel insights concerning what is relevant for learning, when re-learning is needed, and system response to dynamic events (such as rush jobs, increase in processing time, and machine unavailability). Computational experiments also reveal the outperformance of the proposed MAS to other multi-agent systems by at least 25% and common dispatching rules in mean weighted tardiness, as well as other performance measures.} +} + +@article{REZAEI20221, +title = {A Biased Inferential Naivety learning model for a network of agents}, +journal = {Cognitive Systems Research}, +volume = {76}, +pages = {1-12}, +year = {2022}, +issn = {1389-0417}, +doi = {https://doi.org/10.1016/j.cogsys.2022.07.001}, +url = {https://www.sciencedirect.com/science/article/pii/S1389041722000298}, +author = {Zeinab Rezaei and Saeed Setayeshi and Ebrahim Mahdipour}, +keywords = {Bayesian decision making, Heuristic method, Inferential naivety assumption, Observational learning, Social learning}, +abstract = {We propose a Biased Inferential Naivety social learning model. In this model, a group of agents tries to determine the true state of the world and make the best possible decisions. The agents have limited computational abilities. They receive noisy private signals about the true state and observe the history of their neighbors' decisions. The proposed model is rooted in the Bayesian method but avoids the complexity of fully Bayesian inference. In our model, the role of knowledge obtained from social observations is separated from the knowledge obtained from private observations. Therefore, the Bayesian inferences on social observations are approximated using inferential naivety assumption, while purely Bayesian inferences are made on private observations. The reduction of herd behavior is another innovation of the proposed model. This advantage is achieved by reducing the effect of social observations on agents' beliefs over time. Therefore, all the agents learn the truth, and the correct consensus is achieved effectively. In this model, using two cognitive biases, there is heterogeneity in agents' behaviors. Therefore, the growth of beliefs and the learning speed can be improved in different situations. Several Monte Carlo simulations confirm the features of the proposed model. The conditions under which the proposed model leads to asymptotic learning are proved.} +} + +@article{KAMALI2023110242, +title = {An immune inspired multi-agent system for dynamic multi-objective optimization}, +journal = {Knowledge-Based Systems}, +volume = {262}, +pages = {110242}, +year = {2023}, +issn = {0950-7051}, +doi = {https://doi.org/10.1016/j.knosys.2022.110242}, +url = {https://www.sciencedirect.com/science/article/pii/S0950705122013387}, +author = {Seyed Ruhollah Kamali and Touraj Banirostam and Homayun Motameni and Mohammad Teshnehlab}, +keywords = {Immune inspired multi-agent system, Dynamic multi-objective optimization, Severe and frequent changes}, +abstract = {In this research, an immune inspired multi-agent system (IMAS) is proposed to solve optimization problems in dynamic and multi-objective environments. The proposed IMAS uses artificial immune system metaphors to shape the local behaviors of agents to detect environmental changes, generate Pareto optimal solutions, and react to the dynamics of the problem environment. Apart from that, agents enhance their adaptive capacity in dealing with environmental changes to find the global optimum, with a hierarchical structure without any central control. This study used a combination of diversity-, multi-population- and memory-based approaches to perform better in multi-objective environments with severe and frequent changes. The proposed IMAS is compared with six state-of-the-art algorithms on various benchmark problems. The results indicate its superiority in many of the experiments.} +} + +@article{ZHANG2023110564, +title = {A novel human learning optimization algorithm with Bayesian inference learning}, +journal = {Knowledge-Based Systems}, +volume = {271}, +pages = {110564}, +year = {2023}, +issn = {0950-7051}, +doi = {https://doi.org/10.1016/j.knosys.2023.110564}, +url = {https://www.sciencedirect.com/science/article/pii/S0950705123003143}, +author = {Pinggai Zhang and Ling Wang and Zixiang Fei and Lisheng Wei and Minrui Fei and Muhammad Ilyas Menhas}, +keywords = {Human learning optimization, Meta-heuristic, Bayesian inference, Bayesian inference learning, Individual learning, Social learning}, +abstract = {Humans perform Bayesian inference in a wide variety of tasks, which can help people make selection decisions effectively and therefore enhances learning efficiency and accuracy. Inspired by this fact, this paper presents a novel human learning optimization algorithm with Bayesian inference learning (HLOBIL), in which a Bayesian inference learning operator (BILO) is developed to utilize the inference strategy for enhancing learning efficiency. The in-depth analysis shows that the proposed BILO can efficiently improve the exploitation ability of the algorithm as it can achieve the optimal values and retrieve the optimal information with the accumulated search information. Besides, the exploration ability of HLOBIL is also strengthened by the inborn characteristics of Bayesian inference. The experimental results demonstrate that the developed HLOBIL is superior to previous HLO variants and other state-of-art algorithms with its improved exploitation and exploration abilities.} +} + +@article{HIPOLITO2023103510, +title = {Breaking boundaries: The Bayesian Brain Hypothesis for perception and prediction}, +journal = {Consciousness and Cognition}, +volume = {111}, +pages = {103510}, +year = {2023}, +issn = {1053-8100}, +doi = {https://doi.org/10.1016/j.concog.2023.103510}, +url = {https://www.sciencedirect.com/science/article/pii/S1053810023000478}, +author = {Inês Hipólito and Michael Kirchhoff}, +keywords = {Bayesian Brain Hypothesis, Modularity of the Mind, Cognitive processes, Informational boundaries}, +abstract = {This special issue aims to provide a comprehensive overview of the current state of the Bayesian Brain Hypothesis and its standing across neuroscience, cognitive science and the philosophy of cognitive science. By gathering cutting-edge research from leading experts, this issue seeks to showcase the latest advancements in our understanding of the Bayesian brain, as well as its potential implications for future research in perception, cognition, and motor control. A special focus to achieve this aim is adopted in this special issue, as it seeks to explore the relation between two seemingly incompatible frameworks for the understanding of cognitive structure and function: the Bayesian Brain Hypothesis and the Modularity Theory of the Mind. In assessing the compatibility between these theories, the contributors to this special issue open up new pathways of thinking and advance our understanding of cognitive processes.} +} + +@article{LI2023424, +title = {Multi-agent evolution reinforcement learning method for machining parameters optimization based on bootstrap aggregating graph attention network simulated environment}, +journal = {Journal of Manufacturing Systems}, +volume = {67}, +pages = {424-438}, +year = {2023}, +issn = {0278-6125}, +doi = {https://doi.org/10.1016/j.jmsy.2023.02.015}, +url = {https://www.sciencedirect.com/science/article/pii/S0278612523000390}, +author = {Weiye Li and Songping He and Xinyong Mao and Bin Li and Chaochao Qiu and Jinwen Yu and Fangyu Peng and Xin Tan}, +keywords = {Surface roughness, Cutting efficiency, Machining parameters optimization, Graph attention network, Multi-agent reinforcement learning, Evolutionary learning}, +abstract = {Improving machining quality and production efficiency is the focus of the manufacturing industry. How to obtain efficient machining parameters under multiple constraints such as machining quality is a severe challenge for manufacturing industry. In this paper, a multi-agent evolutionary reinforcement learning method (MAERL) is proposed to optimize the machining parameters for high quality and high efficiency machining by combining the graph neural network and reinforcement learning. Firstly, a bootstrap aggregating graph attention network (Bagging-GAT) based roughness estimation method for machined surface is proposed, which combines the structural knowledge between machining parameters and vibration features. Secondly, a mathematical model of machining parameters optimization problem is established, which is formalized into Markov decision process (MDP), and a multi-agent reinforcement learning method is proposed to solve the MDP problem, and evolutionary learning is introduced to improve the stability of multi-agent training. Finally, a series of experiments were carried out on the commutator production line, and the results show that the proposed Bagging-GAT-based method can improve the prediction effect by about 25% in the case of small samples, and the MAERL-based optimization method can better deal with the coupling problem of reward function in the optimization process. Compared with the classical optimization method, the optimization effect is improved by 13% and a lot of optimization time is saved.} +} + +@inproceedings{10.1145/3290605.3300912, +author = {Kim, Yea-Seul and Walls, Logan A. and Krafft, Peter and Hullman, Jessica}, +title = {A Bayesian Cognition Approach to Improve Data Visualization}, +year = {2019}, +isbn = {9781450359702}, +publisher = {Association for Computing Machinery}, +address = {New York, NY, USA}, +url = {https://doi.org/10.1145/3290605.3300912}, +doi = {10.1145/3290605.3300912}, +abstract = {People naturally bring their prior beliefs to bear on how they interpret the new information, yet few formal models exist for accounting for the influence of users' prior beliefs in interactions with data presentations like visualizations. We demonstrate a Bayesian cognitive model for understanding how people interpret visualizations in light of prior beliefs and show how this model provides a guide for improving visualization evaluation. In a first study, we show how applying a Bayesian cognition model to a simple visualization scenario indicates that people's judgments are consistent with a hypothesis that they are doing approximate Bayesian inference. In a second study, we evaluate how sensitive our observations of Bayesian behavior are to different techniques for eliciting people subjective distributions, and to different datasets. We find that people don't behave consistently with Bayesian predictions for large sample size datasets, and this difference cannot be explained by elicitation technique. In a final study, we show how normative Bayesian inference can be used as an evaluation framework for visualizations, including of uncertainty.}, +booktitle = {Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems}, +pages = {1–14}, +numpages = {14}, +keywords = {bayesian cognition, uncertainty elicitation, visualization}, +location = {Glasgow, Scotland Uk}, +series = {CHI '19} +} + +@article{DYER2024104827, +title = {Black-box Bayesian inference for agent-based models}, +journal = {Journal of Economic Dynamics and Control}, +volume = {161}, +pages = {104827}, +year = {2024}, +issn = {0165-1889}, +doi = {https://doi.org/10.1016/j.jedc.2024.104827}, +url = {https://www.sciencedirect.com/science/article/pii/S0165188924000198}, +author = {Joel Dyer and Patrick Cannon and J. Doyne Farmer and Sebastian M. Schmon}, +keywords = {Agent-based models, Bayesian inference, Neural networks, Parameter estimation, Simulation-based inference, Time series}, +abstract = {Simulation models, in particular agent-based models, are gaining popularity in economics and the social sciences. The considerable flexibility they offer, as well as their capacity to reproduce a variety of empirically observed behaviours of complex systems, give them broad appeal, and the increasing availability of cheap computing power has made their use feasible. Yet a widespread adoption in real-world modelling and decision-making scenarios has been hindered by the difficulty of performing parameter estimation for such models. In general, simulation models lack a tractable likelihood function, which precludes a straightforward application of standard statistical inference techniques. A number of recent works have sought to address this problem through the application of likelihood-free inference techniques, in which parameter estimates are determined by performing some form of comparison between the observed data and simulation output. However, these approaches are (a) founded on restrictive assumptions, and/or (b) typically require many hundreds of thousands of simulations. These qualities make them unsuitable for large-scale simulations in economics and the social sciences, and can cast doubt on the validity of these inference methods in such scenarios. In this paper, we investigate the efficacy of two classes of simulation-efficient black-box approximate Bayesian inference methods that have recently drawn significant attention within the probabilistic machine learning community: neural posterior estimation and neural density ratio estimation. We present a number of benchmarking experiments in which we demonstrate that neural network-based black-box methods provide state of the art parameter inference for economic simulation models, and crucially are compatible with generic multivariate or even non-Euclidean time-series data. In addition, we suggest appropriate assessment criteria for use in future benchmarking of approximate Bayesian inference procedures for simulation models in economics and the social sciences.} +} + +@Article{Nikpour2021, +author={Nikpour, Hoda +and Aamodt, Agnar}, +title={Inference and reasoning in a Bayesian knowledge-intensive CBR system}, +journal={Progress in Artificial Intelligence}, +year={2021}, +month={Mar}, +day={01}, +volume={10}, +number={1}, +pages={49-63}, +abstract={This paper presents the inference and reasoning methods in a Bayesian supported knowledge-intensive case-based reasoning (CBR) system called BNCreek. The inference and reasoning process in this system is a combination of three methods. The semantic network inference methods and the CBR method are employed to handle the difficulties of inferencing and reasoning in uncertain domains. The Bayesian network inference methods are employed to make the process more accurate. An experiment from oil well drilling as a complex and uncertain application domain is conducted. The system is evaluated against expert estimations and compared with seven other corresponding systems. The normalized discounted cumulative gain (NDCG) as a rank-based metric, the weighted error (WE), and root-square error (RSE) as the statistical metrics are employed to evaluate different aspects of the system capabilities. The results show the efficiency of the developed inference and reasoning methods.}, +issn={2192-6360}, +doi={10.1007/s13748-020-00223-1}, +url={https://doi.org/10.1007/s13748-020-00223-1} +} + +@article{PRESCOTT2024112577, +title = {Efficient multifidelity likelihood-free Bayesian inference with adaptive computational resource allocation}, +journal = {Journal of Computational Physics}, +volume = {496}, +pages = {112577}, +year = {2024}, +issn = {0021-9991}, +doi = {https://doi.org/10.1016/j.jcp.2023.112577}, +url = {https://www.sciencedirect.com/science/article/pii/S0021999123006721}, +author = {Thomas P. Prescott and David J. Warne and Ruth E. Baker}, +keywords = {Likelihood-free Bayesian inference, Multifidelity approaches}, +abstract = {Likelihood-free Bayesian inference algorithms are popular methods for inferring the parameters of complex stochastic models with intractable likelihoods. These algorithms characteristically rely heavily on repeated model simulations. However, whenever the computational cost of simulation is even moderately expensive, the significant burden incurred by likelihood-free algorithms leaves them infeasible for many practical applications. The multifidelity approach has been introduced in the context of approximate Bayesian computation to reduce the simulation burden of likelihood-free inference without loss of accuracy, by using the information provided by simulating computationally cheap, approximate models in place of the model of interest. In this work we demonstrate that multifidelity techniques can be applied in the general likelihood-free Bayesian inference setting. Analytical results on the optimal allocation of computational resources to simulations at different levels of fidelity are derived, and subsequently implemented practically. We provide an adaptive multifidelity likelihood-free inference algorithm that learns the relationships between models at different fidelities and adapts resource allocation accordingly, and demonstrate that this algorithm produces posterior estimates with near-optimal efficiency.} +} + +@article{RISTIC202030, +title = {A tutorial on uncertainty modeling for machine reasoning}, +journal = {Information Fusion}, +volume = {55}, +pages = {30-44}, +year = {2020}, +issn = {1566-2535}, +doi = {https://doi.org/10.1016/j.inffus.2019.08.001}, +url = {https://www.sciencedirect.com/science/article/pii/S1566253519301976}, +author = {Branko Ristic and Christopher Gilliam and Marion Byrne and Alessio Benavoli}, +keywords = {Information fusion, Uncertainty, Imprecision, Model based classification, Bayesian, Random sets, Belief function theory, Possibility functions, Imprecise probability}, +abstract = {Increasingly we rely on machine intelligence for reasoning and decision making under uncertainty. This tutorial reviews the prevalent methods for model-based autonomous decision making based on observations and prior knowledge, primarily in the context of classification. Both observations and the knowledge-base available for reasoning are treated as being uncertain. Accordingly, the central themes of this tutorial are quantitative modeling of uncertainty, the rules required to combine such uncertain information, and the task of decision making under uncertainty. The paper covers the main approaches to uncertain knowledge representation and reasoning, in particular, Bayesian probability theory, possibility theory, reasoning based on belief functions and finally imprecise probability theory. The main feature of the tutorial is that it illustrates various approaches with several testing scenarios, and provides MATLAB solutions for them as a supplementary material for an interested reader.} +} + +@article{CICIRELLO2022108619, +title = {Machine learning based optimization for interval uncertainty propagation}, +journal = {Mechanical Systems and Signal Processing}, +volume = {170}, +pages = {108619}, +year = {2022}, +issn = {0888-3270}, +doi = {https://doi.org/10.1016/j.ymssp.2021.108619}, +url = {https://www.sciencedirect.com/science/article/pii/S0888327021009493}, +author = {Alice Cicirello and Filippo Giunta}, +keywords = {Bounded uncertainty, Bayesian optimization, Expensive-to-evaluate deterministic computer models, Gaussian process, Communicating uncertainty}, +abstract = {Two non-intrusive uncertainty propagation approaches are proposed for the performance analysis of engineering systems described by expensive-to-evaluate deterministic computer models with parameters defined as interval variables. These approaches employ a machine learning based optimization strategy, the so-called Bayesian optimization, for evaluating the upper and lower bounds of a generic response variable over the set of possible responses obtained when each interval variable varies independently over its range. The lack of knowledge caused by not evaluating the response function for all the possible combinations of the interval variables is accounted for by developing a probabilistic description of the response variable itself by using a Gaussian Process regression model. An iterative procedure is developed for selecting a small number of simulations to be evaluated for updating this statistical model by using well-established acquisition functions and to assess the response bounds. In both approaches, an initial training dataset is defined. While one approach builds iteratively two distinct training datasets for evaluating separately the upper and lower bounds of the response variable, the other one builds iteratively a single training dataset. Consequently, the two approaches will produce different bound estimates at each iteration. The upper and lower response bounds are expressed as point estimates obtained from the mean function of the posterior distribution. Moreover, a confidence interval on each estimate is provided for effectively communicating to engineers when these estimates are obtained at a combination of the interval variables for which no deterministic simulation has been run. Finally, two metrics are proposed to define conditions for assessing if the predicted bound estimates can be considered satisfactory. The applicability of these two approaches is illustrated with two numerical applications, one focusing on vibration and the other on vibro-acoustics.} +} + +@INPROCEEDINGS{9278071, + author={Petit, Maxime and Dellandrea, Emmanuel and Chen, Liming}, + booktitle={2020 Joint IEEE 10th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob)}, + title={Bayesian Optimization for Developmental Robotics with Meta-Learning by Parameters Bounds Reduction}, + year={2020}, + volume={}, + number={}, + pages={1-8}, + keywords={Optimization;Robots;Task analysis;Bayes methods;Visualization;Service robots;Cognition;developmental robotics;long-term memory;meta learning;hyperparmeters automatic optimization;case-based reasoning}, + doi={10.1109/ICDL-EpiRob48136.2020.9278071} +} + +@article{LI2023477, +title = {Hierarchical and partitioned planning strategy for closed-loop devices in low-voltage distribution network based on improved KMeans partition method}, +journal = {Energy Reports}, +volume = {9}, +pages = {477-485}, +year = {2023}, +note = {2022 The 3rd International Conference on Power and Electrical Engineering}, +issn = {2352-4847}, +doi = {https://doi.org/10.1016/j.egyr.2023.05.161}, +url = {https://www.sciencedirect.com/science/article/pii/S2352484723009137}, +author = {Jingqi Li and Junlin Li and Dan Wang and Chengxiong Mao and Zhitao Guan and Zhichao Liu and Miaomiao Du and Yuanzhuo Qi and Lexiang Wang and Wenge Liu and Pengfei Tang}, +keywords = {Closed-loop device, Distribution network partition, Device planning, Hierarchical planning, Improved KMeans partition method}, +abstract = {To improve the reliability of power supply, this paper proposes a hierarchical and partitioned planning strategy for closed-loop devices in low-voltage distribution network. Based on the geographic location and load situation of the distribution network area, an improved KMeans partition method is used to partition the area in the upper layer. In the lower layer, an intelligent algorithm is adopted to decide the numbers and placement locations of mobile low-voltage contact boxes and mobile seamless closed-loop load transfer devices in each partition with the goal of the highest closed-loop safety, the greatest improvement in annual power outage amount and the lowest cost. Finally, the feasibility and effectiveness of the proposed strategy are proved by an example.} +} + +@article{SAXENA2024100838, +title = {Hybrid KNN-SVM machine learning approach for solar power forecasting}, +journal = {Environmental Challenges}, +volume = {14}, +pages = {100838}, +year = {2024}, +issn = {2667-0100}, +doi = {https://doi.org/10.1016/j.envc.2024.100838}, +url = {https://www.sciencedirect.com/science/article/pii/S2667010024000040}, +author = {Nishant Saxena and Rahul Kumar and Yarrapragada K S S Rao and Dilbag Singh Mondloe and Nishikant Kishor Dhapekar and Abhishek Sharma and Anil Singh Yadav}, +keywords = {Solar power forecasting, Hybrid model, KNN, Optimization, Solar energy, SVM}, +abstract = {Predictions about solar power will have a significant impact on large-scale renewable energy plants. Photovoltaic (PV) power generation forecasting is particularly sensitive to measuring the uncertainty in weather conditions. Although several conventional techniques like long short-term memory (LSTM), support vector machine (SVM), etc. are available, but due to some restrictions, their application is limited. To enhance the precision of forecasting solar power from solar farms, a hybrid machine learning model that includes blends of the K-Nearest Neighbor (KNN) machine learning technique with the SVM to increase reliability for power system operators is proposed in this investigation. The conventional LSTM technique is also implemented to compare the performance of the proposed hybrid technique. The suggested hybrid model is improved by the use of structural diversity and data diversity in KNN and SVM, respectively. For the solar power predictions, the suggested method was tested on the Jodhpur real-time series dataset obtained from the data centers of weather stations using Meteonorm. The data set includes metrics such as Hourly Average Temperature (HAT), Hourly Total Sunlight Duration (HTSD), Hourly Total Global Solar Radiation (HTGSR), and Hourly Total Photovoltaic Energy Generation (HTPEG). The collated data has been segmented into training data, validation data, and testing data. Furthermore, the proposed technique performed better when evaluated on the three performance indices, viz., accuracy, sensitivity, and specificity. Compared with the conventional LSTM technique, the hybrid technique improved the prediction with 98\% accuracy.} +} + +@article{RAKESH2023100898, +title = {Moving object detection using modified GMM based background subtraction}, +journal = {Measurement: Sensors}, +volume = {30}, +pages = {100898}, +year = {2023}, +issn = {2665-9174}, +doi = {https://doi.org/10.1016/j.measen.2023.100898}, +url = {https://www.sciencedirect.com/science/article/pii/S2665917423002349}, +author = {S. Rakesh and Nagaratna P. Hegde and M. {Venu Gopalachari} and D. Jayaram and Bhukya Madhu and Mohd Abdul Hameed and Ramdas Vankdothu and L.K. {Suresh Kumar}}, +keywords = {Background subtraction, Gaussian mixture models, Intelligent video surveillance, Object detection}, +abstract = {Academics have become increasingly interested in creating cutting-edge technologies to enhance Intelligent Video Surveillance (IVS) performance in terms of accuracy, speed, complexity, and deployment. It has been noted that precise object detection is the only way for IVS to function well in higher level applications including event interpretation, tracking, classification, and activity recognition. Through the use of cutting-edge techniques, the current study seeks to improve the performance accuracy of object detection techniques based on Gaussian Mixture Models (GMM). It is achieved by developing crucial phases in the object detecting process. In this study, it is discussed how to model each pixel as a mixture of Gaussians and how to update the model using an online k-means approximation. The adaptive mixture model's Gaussian distributions are then analyzed to identify which ones are more likely to be the product of a background process. Each pixel is categorized according to whether the background model is thought to include the Gaussian distribution that best depicts it.} +} + +@article{JIAO2022540, +title = {Interpretable fuzzy clustering using unsupervised fuzzy decision trees}, +journal = {Information Sciences}, +volume = {611}, +pages = {540-563}, +year = {2022}, +issn = {0020-0255}, +doi = {https://doi.org/10.1016/j.ins.2022.08.077}, +url = {https://www.sciencedirect.com/science/article/pii/S0020025522009872}, +author = {Lianmeng Jiao and Haoyu Yang and Zhun-ga Liu and Quan Pan}, +keywords = {Fuzzy clustering, Interpretable clustering, Unsupervised decision tree, Cluster merging}, +abstract = {In clustering process, fuzzy partition performs better than hard partition when the boundaries between clusters are vague. Whereas, traditional fuzzy clustering algorithms produce less interpretable results, limiting their application in security, privacy, and ethics fields. To that end, this paper proposes an interpretable fuzzy clustering algorithm—fuzzy decision tree-based clustering which combines the flexibility of fuzzy partition with the interpretability of the decision tree. We constructed an unsupervised multi-way fuzzy decision tree to achieve the interpretability of clustering, in which each cluster is determined by one or several paths from the root to leaf nodes. The proposed algorithm comprises three main modules: feature and cutting point-selection, node fuzzy splitting, and cluster merging. The first two modules are repeated to generate an initial unsupervised decision tree, and the final module is designed to combine similar leaf nodes to form the final compact clustering model. Our algorithm optimizes an internal clustering validation metric to automatically determine the number of clusters without their initial positions. The synthetic and benchmark datasets were used to test the performance of the proposed algorithm. Furthermore, we provided two examples demonstrating its interest in solving practical problems.} +} + +@article{ARNAUGONZALEZ2023101516, +title = {A methodological approach to enable natural language interaction in an Intelligent Tutoring System}, +journal = {Computer Speech and Language}, +volume = {81}, +pages = {101516}, +year = {2023}, +issn = {0885-2308}, +doi = {https://doi.org/10.1016/j.csl.2023.101516}, +url = {https://www.sciencedirect.com/science/article/pii/S0885230823000359}, +author = {Pablo Arnau-González and Miguel Arevalillo-Herráez and Romina Albornoz-De Luise and David Arnau}, +keywords = {Intelligent tutoring systems (ITS), Interactive learning environments (ILE), Conversational agents, Rasa, Natural language understanding (NLU), Natural language processing (NLP)}, +abstract = {In this paper, we present and evaluate the recent incorporation of a conversational agent into an Intelligent Tutoring System (ITS), using the open-source machine learning framework Rasa. Once it has been appropriately trained, this tool is capable of identifying the intention of a given text input and extracting the relevant entities related to the message content. We describe both the generation of a realistic training set in Spanish language that enables the creation of the required Natural Language Understanding (NLU) models and the evaluation of the resulting system. For the generation of the training set, we have followed a methodology that can be easily exported to other ITS. The model evaluation shows that the conversational agent can correctly identify the majority of the user intents, reporting an f1-score above 95%, and cooperate with the ITS to produce a consistent dialogue flow that makes interaction more natural.} +} + +@article{MAO20224065, +title = {An Exploratory Approach to Intelligent Quiz Question Recommendation}, +journal = {Procedia Computer Science}, +volume = {207}, +pages = {4065-4074}, +year = {2022}, +note = {Knowledge-Based and Intelligent Information and Engineering Systems: Proceedings of the 26th International Conference KES2022}, +issn = {1877-0509}, +doi = {https://doi.org/10.1016/j.procs.2022.09.469}, +url = {https://www.sciencedirect.com/science/article/pii/S1877050922013631}, +author = {Kejie Mao and Qiwen Dong and Ye Wang and Daocheng Honga}, +keywords = {question recommendation, two-sided recommender systems, reinforcement learning, intelligent tutoring}, +abstract = {With the rapid advancement of ICT, the digital transformation on education is greatly accelerating in various applications. As a particularly prominent application of digital education, quiz question recommendation is playing a vital role in precision teaching, smart tutoring, and personalized learning. However, the looming challenge of quiz question recommender for students is to satisfy the question diversity demands for students ZPD (the zone of proximal development) stage dynamically online. Therefore, we propose to formalize quiz question recommendation with a novel approach of reinforcement learning based two-sided recommender system. We develop a recommendation framework RTR (Reinforcement-Learning based Two-sided Recommender Systems) for taking into account the interests of both sides of the system, learning and adapting to those interests in real time, and resulting in more satisfactory recommended content. This established recommendation framework captures question characters and student dynamic preferences by considering the emergence of both sides of the system, and it yields a better learning experience in the context of practical quiz question generation.} +} + +@article{CLEMENTE2022118171, +title = {A proposal for an adaptive Recommender System based on competences and ontologies}, +journal = {Expert Systems with Applications}, +volume = {208}, +pages = {118171}, +year = {2022}, +issn = {0957-4174}, +doi = {https://doi.org/10.1016/j.eswa.2022.118171}, +url = {https://www.sciencedirect.com/science/article/pii/S0957417422013392}, +author = {Julia Clemente and Héctor Yago and Javier {de Pedro-Carracedo} and Javier Bueno}, +keywords = {Recommender system, , Ontology network, Methodological development, Student modeling}, +abstract = {Context: +Competences represent an interesting pedagogical support in many processes like diagnosis or recommendation. From these, it is possible to infer information about the progress of the student to provide help targeted both, trainers who must make adaptive tutoring decisions for each learner, and students to detect and correct their learning weaknesses. For the correct development of any of these tasks, it is important to have a suitable student model that allows the representation of the most significant information possible about the student. Additionally, it would be very advantageous for this modeling to incorporate mechanisms from which it would be possible to infer more information about the student’s state of knowledge. +Objective: +To facilitate this goal, in this paper a new approach to develop an adaptive competence-based recommender system is proposed. +Method: +We present a methodological development guide as well as a set of ontological and non-ontological resources to develop and adapt the prototype of the proposed recommender system. +Results: +A modular flexible ontology network previously built for this purpose has been extended, which is responsible for recording the instructional design and student information. Furthermore, we describe a case study based on a first aid learning experience to assess the prototype with the proposed methodology. +Conclusions: +We highlight the relevance of flexibility and adaptability in learning modeling and recommendation processes. In order to promote improvement in the personalized learning of students, we present a Recommender System prototype taking advantages of ontologies, with a methodological guide, a broad taxonomy of recommendation criteria and the nature of competences. Future lines of research lines, including a more comprehensive evaluation of the system, will allow us to demonstrate in depth its adaptability according to the characteristics of the student, flexibility and extensibility for its integration in various environments and domains.} +} + +@article{https://doi.org/10.1155/2023/2578286, +author = {Li, Linqing and Wang, Zhifeng}, +title = {Knowledge Graph-Enhanced Intelligent Tutoring System Based on Exercise Representativeness and Informativeness}, +journal = {International Journal of Intelligent Systems}, +volume = {2023}, +number = {1}, +pages = {2578286}, +doi = {https://doi.org/10.1155/2023/2578286}, +url = {https://onlinelibrary.wiley.com/doi/abs/10.1155/2023/2578286}, +eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1155/2023/2578286}, +abstract = {In the realm of online tutoring intelligent systems, e-learners are exposed to a substantial volume of learning content. The extraction and organization of exercises and skills hold significant importance in establishing clear learning objectives and providing appropriate exercise recommendations. Presently, knowledge graph-based recommendation algorithms have garnered considerable attention among researchers. However, these algorithms solely consider knowledge graphs with single relationships and do not effectively model exercise-rich features, such as exercise representativeness and informativeness. Consequently, this paper proposes a framework, namely, the Knowledge Graph Importance-Exercise Representativeness and Informativeness Framework, to address these two issues. The framework consists of four intricate components and a novel cognitive diagnosis model called the Neural Attentive Cognitive Diagnosis model to recommend the proper exercises. These components encompass the informativeness component, exercise representation component, knowledge importance component, and exercise representativeness component. The informativeness component evaluates the informational value of each exercise and identifies the candidate exercise set (EC) that exhibits the highest exercise informativeness. Moreover, the exercise representation component utilizes a graph neural network to process student records. The output of the graph neural network serves as the input for exercise-level attention and skill-level attention, ultimately generating exercise embeddings and skill embeddings. Furthermore, the skill embeddings are employed as input for the knowledge importance component. This component transforms a one-dimensional knowledge graph into a multidimensional one through four class relations and calculates skill importance weights based on novelty and popularity. Subsequently, the exercise representativeness component incorporates exercise weight knowledge coverage to select exercises from the candidate exercise set for the tested exercise set. Lastly, the cognitive diagnosis model leverages exercise representation and skill importance weights to predict student performance on the test set and estimate their knowledge state. To evaluate the effectiveness of our selection strategy, extensive experiments were conducted on two types of publicly available educational datasets. The experimental results demonstrate that our framework can recommend appropriate exercises to students, leading to improved student performance.}, +year = {2023} +} + +@inproceedings{badier:hal-04092828, + TITLE = {{Comprendre les usages et effets d'un syst{\`e}me de recommandations p{\'e}dagogiques en contexte d'apprentissage non-formel}}, + AUTHOR = {Badier, Ana{\"e}lle and Lefort, Mathieu and Lefevre, Marie}, + URL = {https://hal.science/hal-04092828}, + BOOKTITLE = {{EIAH'23}}, + ADDRESS = {Brest, France}, + YEAR = {2023}, + MONTH = Jun, + HAL_ID = {hal-04092828}, + HAL_VERSION = {v1}, +} + +@article{BADRA2023108920, +title = {Case-based prediction – A survey}, +journal = {International Journal of Approximate Reasoning}, +volume = {158}, +pages = {108920}, +year = {2023}, +issn = {0888-613X}, +doi = {https://doi.org/10.1016/j.ijar.2023.108920}, +url = {https://www.sciencedirect.com/science/article/pii/S0888613X23000440}, +author = {Fadi Badra and Marie-Jeanne Lesot}, +keywords = {Case-based prediction, Analogical transfer, Similarity}, +abstract = {This paper clarifies the relation between case-based prediction and analogical transfer. Case-based prediction consists in predicting the outcome associated with a new case directly from its comparison with a set of cases retrieved from a case base, by relying solely on a structured memory and some similarity measures. Analogical transfer is a cognitive process that allows to derive some new information about a target situation by applying a plausible inference principle, according to which if two situations are similar with respect to some criteria, then it is plausible that they are also similar with respect to other criteria. Case-based prediction algorithms are known to apply analogical transfer to make predictions, but the existing approaches are diverse, and developing a unified theory of case-based prediction remains a challenge. In this paper, we show that a common principle underlying case-based prediction methods is that they interpret the plausible inference as a transfer of similarity knowledge from a situation space to an outcome space. Among all potential outcomes, the predicted outcome is the one that optimizes this transfer, i.e., that makes the similarities in the outcome space most compatible with the observed similarities in the situation space. Based on this observation, a systematic analysis of the different theories of case-based prediction is presented, where the approaches are distinguished according to the type of knowledge used to measure the compatibility between the two sets of similarity relations.} +} + + +@Article{jmse11050890 , +AUTHOR = {Louvros, Panagiotis and Stefanidis, Fotios and Boulougouris, Evangelos and Komianos, Alexandros and Vassalos, Dracos}, +TITLE = {Machine Learning and Case-Based Reasoning for Real-Time Onboard Prediction of the Survivability of Ships}, +JOURNAL = {Journal of Marine Science and Engineering}, +VOLUME = {11}, +YEAR = {2023}, +NUMBER = {5}, +ARTICLE-NUMBER = {890}, +URL = {https://www.mdpi.com/2077-1312/11/5/890}, +ISSN = {2077-1312}, +ABSTRACT = {The subject of damaged stability has greatly profited from the development of new tools and techniques in recent history. Specifically, the increased computational power and the probabilistic approach have transformed the subject, increasing accuracy and fidelity, hence allowing for a universal application and the inclusion of the most probable scenarios. Currently, all ships are evaluated for their stability and are expected to survive the dangers they will most likely face. However, further advancements in simulations have made it possible to further increase the fidelity and accuracy of simulated casualties. Multiple time domain and, to a lesser extent, Computational Fluid dynamics (CFD) solutions have been suggested as the next “evolutionary” step for damage stability. However, while those techniques are demonstrably more accurate, the computational power to utilize them for the task of probabilistic evaluation is not there yet. In this paper, the authors present a novel approach that aims to serve as a stopgap measure for introducing the time domain simulations in the existing framework. Specifically, the methodology presented serves the purpose of a fast decision support tool which is able to provide information regarding the ongoing casualty utilizing prior knowledge gained from simulations. This work was needed and developed for the purposes of the EU-funded project SafePASS.}, +DOI = {10.3390/jmse11050890} +} + + +@Article{su14031366, +AUTHOR = {Chun, Se-Hak and Jang, Jae-Won}, +TITLE = {A New Trend Pattern-Matching Method of Interactive Case-Based Reasoning for Stock Price Predictions}, +JOURNAL = {Sustainability}, +VOLUME = {14}, +YEAR = {2022}, +NUMBER = {3}, +ARTICLE-NUMBER = {1366}, +URL = {https://www.mdpi.com/2071-1050/14/3/1366}, +ISSN = {2071-1050}, +ABSTRACT = {In this paper, we suggest a new case-based reasoning method for stock price predictions using the knowledge of traders to select similar past patterns among nearest neighbors obtained from a traditional case-based reasoning machine. Thus, this method overcomes the limitation of conventional case-based reasoning, which does not consider how to retrieve similar neighbors from previous patterns in terms of a graphical pattern. In this paper, we show how the proposed method can be used when traders find similar time series patterns among nearest cases. For this, we suggest an interactive prediction system where traders can select similar patterns with individual knowledge among automatically recommended neighbors by case-based reasoning. In this paper, we demonstrate how traders can use their knowledge to select similar patterns using a graphical interface, serving as an exemplar for the target. These concepts are investigated against the backdrop of a practical application involving the prediction of three individual stock prices, i.e., Zoom, Airbnb, and Twitter, as well as the prediction of the Dow Jones Industrial Average (DJIA). The verification of the prediction results is compared with a random walk model based on the RMSE and Hit ratio. The results show that the proposed technique is more effective than the random walk model but it does not statistically surpass the random walk model.}, +DOI = {10.3390/su14031366} +} + +@Article{fire7040107, +AUTHOR = {Pei, Qiuyan and Jia, Zhichao and Liu, Jia and Wang, Yi and Wang, Junhui and Zhang, Yanqi}, +TITLE = {Prediction of Coal Spontaneous Combustion Hazard Grades Based on Fuzzy Clustered Case-Based Reasoning}, +JOURNAL = {Fire}, +VOLUME = {7}, +YEAR = {2024}, +NUMBER = {4}, +ARTICLE-NUMBER = {107}, +URL = {https://www.mdpi.com/2571-6255/7/4/107}, +ISSN = {2571-6255}, +ABSTRACT = {Accurate prediction of the coal spontaneous combustion hazard grades is of great significance to ensure the safe production of coal mines. However, traditional coal temperature prediction models have low accuracy and do not predict the coal spontaneous combustion hazard grades. In order to accurately predict coal spontaneous combustion hazard grades, a prediction model of coal spontaneous combustion based on principal component analysis (PCA), case-based reasoning (CBR), fuzzy clustering (FM), and the snake optimization (SO) algorithm was proposed in this manuscript. Firstly, based on the change rule of the concentration of signature gases in the process of coal warming, a new method of classifying the risk of spontaneous combustion of coal was established. Secondly, MeanRadius-SMOTE was adopted to balance the data structure. The weights of the prediction indicators were calculated through PCA to enhance the prediction precision of the CBR model. Then, by employing FM in the case base, the computational cost of CBR was reduced and its computational efficiency was improved. The SO algorithm was used to determine the hyperparameters in the PCA-FM-CBR model. In addition, multiple comparative experiments were conducted to verify the superiority of the model proposed in this manuscript. The results indicated that SO-PCA-FM-CBR possesses good prediction performance and also improves computational efficiency. Finally, the authors of this manuscript adopted the Random Balance Designs—Fourier Amplitude Sensitivity Test (RBD-FAST) to explain the output of the model and analyzed the global importance of input variables. The results demonstrated that CO is the most important variable affecting the coal spontaneous combustion hazard grades.}, +DOI = {10.3390/fire7040107} +} + +@Article{Desmarais2012, +author={Desmarais, Michel C. +and Baker, Ryan S. J. d.}, +title={A review of recent advances in learner and skill modeling in intelligent learning environments}, +journal={User Modeling and User-Adapted Interaction}, +year={2012}, +month={Apr}, +day={01}, +volume={22}, +number={1}, +pages={9-38}, +abstract={In recent years, learner models have emerged from the research laboratory and research classrooms into the wider world. Learner models are now embedded in real world applications which can claim to have thousands, or even hundreds of thousands, of users. Probabilistic models for skill assessment are playing a key role in these advanced learning environments. In this paper, we review the learner models that have played the largest roles in the success of these learning environments, and also the latest advances in the modeling and assessment of learner skills. We conclude by discussing related advancements in modeling other key constructs such as learner motivation, emotional and attentional state, meta-cognition and self-regulated learning, group learning, and the recent movement towards open and shared learner models.}, +issn={1573-1391}, +doi={10.1007/s11257-011-9106-8}, +url={https://doi.org/10.1007/s11257-011-9106-8} +} + +@article{Eide, +title={Dynamic slate recommendation with gated recurrent units and Thompson sampling}, +author={Eide, Simen and Leslie, David S. and Frigessi, Arnoldo}, +language={English}, +type={article}, +volume = {36}, +year = {2022}, +issn = {1573-756X}, +doi = {https://doi.org/10.1007/s10618-022-00849-w}, +url = {https://doi.org/10.1007/s10618-022-00849-w}, +abstract={We consider the problem of recommending relevant content to users of an internet platform in the form of lists of items, called slates. We introduce a variational Bayesian Recurrent Neural Net recommender system that acts on time series of interactions between the internet platform and the user, and which scales to real world industrial situations. The recommender system is tested both online on real users, and on an offline dataset collected from a Norwegian web-based marketplace, FINN.no, that is made public for research. This is one of the first publicly available datasets which includes all the slates that are presented to users as well as which items (if any) in the slates were clicked on. Such a data set allows us to move beyond the common assumption that implicitly assumes that users are considering all possible items at each interaction. Instead we build our likelihood using the items that are actually in the slate, and evaluate the strengths and weaknesses of both approaches theoretically and in experiments. We also introduce a hierarchical prior for the item parameters based on group memberships. Both item parameters and user preferences are learned probabilistically. Furthermore, we combine our model with bandit strategies to ensure learning, and introduce ‘in-slate Thompson sampling’ which makes use of the slates to maximise explorative opportunities. We show experimentally that explorative recommender strategies perform on par or above their greedy counterparts. Even without making use of exploration to learn more effectively, click rates increase simply because of improved diversity in the recommended slates.} +} + +@InProceedings{10.1007/978-3-031-09680-8_14, +author={Sablayrolles, Louis +and Lefevre, Marie +and Guin, Nathalie +and Broisin, Julien}, +editor={Crossley, Scott +and Popescu, Elvira}, +title={Design and Evaluation of a Competency-Based Recommendation Process}, +booktitle={Intelligent Tutoring Systems}, +year={2022}, +publisher={Springer International Publishing}, +address={Cham}, +pages={148--160}, +abstract={The purpose of recommending activities to learners is to provide them with resources adapted to their needs, to facilitate the learning process. However, when teachers face a large number of students, it is difficult for them to recommend a personalized list of resources to each learner. In this paper, we are interested in the design of a system that automatically recommends resources to learners using their cognitive profile expressed in terms of competencies, but also according to a specific strategy defined by teachers. Our contributions relate to (1) a competency-based pedagogical strategy allowing to express the teacher's expertise, and (2) a recommendation process based on this strategy. This process has been experimented and assessed with students learning Shell programming in a first-year computer science degree. The first results show that (i) the items selected by our system from the set of possible items were relevant according to the experts; (ii) our system provided recommendations in a reasonable time; (iii) the recommendations were consulted by the learners but lacked usability.}, +isbn={978-3-031-09680-8} +} + +@inproceedings{10.1145/3578337.3605122, +author = {Xu, Shuyuan and Ge, Yingqiang and Li, Yunqi and Fu, Zuohui and Chen, Xu and Zhang, Yongfeng}, +title = {Causal Collaborative Filtering}, +year = {2023}, +isbn = {9798400700736}, +publisher = {Association for Computing Machinery}, +address = {New York, NY, USA}, +url = {https://doi.org/10.1145/3578337.3605122}, +doi = {10.1145/3578337.3605122}, +abstract = {Many of the traditional recommendation algorithms are designed based on the fundamental idea of mining or learning correlative patterns from data to estimate the user-item correlative preference. However, pure correlative learning may lead to Simpson's paradox in predictions, and thus results in sacrificed recommendation performance. Simpson's paradox is a well-known statistical phenomenon, which causes confusions in statistical conclusions and ignoring the paradox may result in inaccurate decisions. Fortunately, causal and counterfactual modeling can help us to think outside of the observational data for user modeling and personalization so as to tackle such issues. In this paper, we propose Causal Collaborative Filtering (CCF) --- a general framework for modeling causality in collaborative filtering and recommendation. We provide a unified causal view of CF and mathematically show that many of the traditional CF algorithms are actually special cases of CCF under simplified causal graphs. We then propose a conditional intervention approach for do-operations so that we can estimate the user-item causal preference based on the observational data. Finally, we further propose a general counterfactual constrained learning framework for estimating the user-item preferences. Experiments are conducted on two types of real-world datasets---traditional and randomized trial data---and results show that our framework can improve the recommendation performance and reduce the Simpson's paradox problem of many CF algorithms.}, +booktitle = {Proceedings of the 2023 ACM SIGIR International Conference on Theory of Information Retrieval}, +pages = {235–245}, +numpages = {11}, +keywords = {recommender systems, counterfactual reasoning, collaborative filtering, causal analysis, Simpson's paradox}, +location = {Taipei, Taiwan}, +series = {ICTIR '23} +} + +@inproceedings{10.1145/3583780.3615048, +author = {Zhu, Zheqing and Van Roy, Benjamin}, +title = {Scalable Neural Contextual Bandit for Recommender Systems}, +year = {2023}, +isbn = {9798400701245}, +publisher = {Association for Computing Machinery}, +address = {New York, NY, USA}, +url = {https://doi.org/10.1145/3583780.3615048}, +doi = {10.1145/3583780.3615048}, +abstract = {High-quality recommender systems ought to deliver both innovative and relevant content through effective and exploratory interactions with users. Yet, supervised learning-based neural networks, which form the backbone of many existing recommender systems, only leverage recognized user interests, falling short when it comes to efficiently uncovering unknown user preferences. While there has been some progress with neural contextual bandit algorithms towards enabling online exploration through neural networks, their onerous computational demands hinder widespread adoption in real-world recommender systems. In this work, we propose a scalable sample-efficient neural contextual bandit algorithm for recommender systems. To do this, we design an epistemic neural network architecture, Epistemic Neural Recommendation (ENR), that enables Thompson sampling at a large scale. In two distinct large-scale experiments with real-world tasks, ENR significantly boosts click-through rates and user ratings by at least 9\% and 6\% respectively compared to state-of-the-art neural contextual bandit algorithms. Furthermore, it achieves equivalent performance with at least 29\% fewer user interactions compared to the best-performing baseline algorithm. Remarkably, while accomplishing these improvements, ENR demands orders of magnitude fewer computational resources than neural contextual bandit baseline algorithms.}, +booktitle = {Proceedings of the 32nd ACM International Conference on Information and Knowledge Management}, +pages = {3636–3646}, +numpages = {11}, +keywords = {contextual bandits, decision making under uncertainty, exploration vs exploitation, recommender systems, reinforcement learning}, +location = {Birmingham, United Kingdom}, +series = {CIKM '23} +} + +@ARTICLE{10494875, + author={Ghoorchian, Saeed and Kortukov, Evgenii and Maghsudi, Setareh}, + journal={IEEE Open Journal of Signal Processing}, + title={Non-Stationary Linear Bandits With Dimensionality Reduction for Large-Scale Recommender Systems}, + year={2024}, + volume={5}, + number={}, + pages={548-558}, + keywords={Vectors;Recommender systems;Decision making;Runtime;Signal processing algorithms;Covariance matrices;Robustness;Decision-making;multi-armed bandit;non-stationary environment;online learning;recommender systems}, + doi={10.1109/OJSP.2024.3386490} +} + +@article{GIANNIKIS2024111752, +title = {Reinforcement learning for addressing the cold-user problem in recommender systems}, +journal = {Knowledge-Based Systems}, +volume = {294}, +pages = {111752}, +year = {2024}, +issn = {0950-7051}, +doi = {https://doi.org/10.1016/j.knosys.2024.111752}, +url = {https://www.sciencedirect.com/science/article/pii/S0950705124003873}, +author = {Stelios Giannikis and Flavius Frasincar and David Boekestijn}, +keywords = {Recommender systems, Reinforcement learning, Active learning, Cold-user problem}, +abstract = {Recommender systems are widely used in webshops because of their ability to provide users with personalized recommendations. However, the cold-user problem (i.e., recommending items to new users) is an important issue many webshops face. With the recent General Data Protection Regulation in Europe, the use of additional user information such as demographics is not possible without the user’s explicit consent. Several techniques have been proposed to solve the cold-user problem. Many of these techniques utilize Active Learning (AL) methods, which let cold users rate items to provide better recommendations for them. In this research, we propose two novel approaches that combine reinforcement learning with AL to elicit the users’ preferences and provide them with personalized recommendations. We compare reinforcement learning approaches that are either AL-based or item-based, where the latter predicts users’ ratings of an item by using their ratings of similar items. Differently than many of the existing approaches, this comparison is made based on implicit user information. Using a large real-world dataset, we show that the item-based strategy is more accurate than the AL-based strategy as well as several existing AL strategies.} +} + +@article{IFTIKHAR2024121541, +title = {A reinforcement learning recommender system using bi-clustering and Markov Decision Process}, +journal = {Expert Systems with Applications}, +volume = {237}, +pages = {121541}, +year = {2024}, +issn = {0957-4174}, +doi = {https://doi.org/10.1016/j.eswa.2023.121541}, +url = {https://www.sciencedirect.com/science/article/pii/S0957417423020432}, +author = {Arta Iftikhar and Mustansar Ali Ghazanfar and Mubbashir Ayub and Saad {Ali Alahmari} and Nadeem Qazi and Julie Wall}, +keywords = {Reinforcement learning, Markov Decision Process, Bi-clustering, Q-learning, Policy}, +abstract = {Collaborative filtering (CF) recommender systems are static in nature and does not adapt well with changing user preferences. User preferences may change after interaction with a system or after buying a product. Conventional CF clustering algorithms only identifies the distribution of patterns and hidden correlations globally. However, the impossibility of discovering local patterns by these algorithms, headed to the popularization of bi-clustering algorithms. Bi-clustering algorithms can analyze all dataset dimensions simultaneously and consequently, discover local patterns that deliver a better understanding of the underlying hidden correlations. In this paper, we modelled the recommendation problem as a sequential decision-making problem using Markov Decision Processes (MDP). To perform state representation for MDP, we first converted user-item votings matrix to a binary matrix. Then we performed bi-clustering on this binary matrix to determine a subset of similar rows and columns. A bi-cluster merging algorithm is designed to merge similar and overlapping bi-clusters. These bi-clusters are then mapped to a squared grid (SG). RL is applied on this SG to determine best policy to give recommendation to users. Start state is determined using Improved Triangle Similarity (ITR similarity measure. Reward function is computed as grid state overlapping in terms of users and items in current and prospective next state. A thorough comparative analysis was conducted, encompassing a diverse array of methodologies, including RL-based, pure Collaborative Filtering (CF), and clustering methods. The results demonstrate that our proposed method outperforms its competitors in terms of precision, recall, and optimal policy learning.} +} + +@article{Soto2, +author={Soto-Forero, Daniel and Ackermann, Simha and Betbeder, Marie-Laure and Henriet, Julien}, +title={Automatic Real-Time Adaptation of Training Session Difficulty Using Rules and Reinforcement Learning in the AI-VT ITS}, +journal = {International Journal of Modern Education and Computer Science(IJMECS)}, +volume = {16}, +pages = {56-71}, +year = {2024}, +issn = {2075-0161}, +doi = { https://doi.org/10.5815/ijmecs.2024.03.05}, +url = {https://www.mecs-press.org/ijmecs/ijmecs-v16-n3/v16n3-5.html}, +keywords={Real Time Adaptation, Intelligent Training System, Thompson Sampling, Case-Based Reasoning, Automatic Adaptation}, +abstract={Some of the most common and typical issues in the field of intelligent tutoring systems (ITS) are (i) the correct identification of learners’ difficulties in the learning process, (ii) the adaptation of content or presentation of the system according to the difficulties encountered, and (iii) the ability to adapt without initial data (cold-start). In some cases, the system tolerates modifications after the realization and assessment of competences. Other systems require complicated real-time adaptation since only a limited number of data can be captured. In that case, it must be analyzed properly and with a certain precision in order to obtain the appropriate adaptations. Generally, for the adaptation step, the ITS gathers common learners together and adapts their training similarly. Another type of adaptation is more personalized, but requires acquired or estimated information about each learner (previous grades, probability of success, etc.). Some of these parameters may be difficult to obtain, and others are imprecise and can lead to misleading adaptations. The adaptation using machine learning requires prior training with a lot of data. This article presents a model for the real time automatic adaptation of a predetermined session inside an ITS called AI-VT. This adaptation process is part of a case-based reasoning global model. The characteristics of the model proposed in this paper (i) require a limited number of data in order to generate a personalized adaptation, (ii) do not require training, (iii) are based on the correlation to complexity levels, and (iv) are able to adapt even at the cold-start stage. The proposed model is presented with two different configurations, deterministic and stochastic. The model has been tested with a database of 1000 learners, corresponding to different knowledge levels in three different scenarios. The results show the dynamic adaptation of the proposed model in both versions, with the adaptations obtained helping the system to evolve more rapidly and identify learner weaknesses in the different levels of complexity as well as the generation of pertinent recommendations in specific cases for each learner capacity.} +} + +@InProceedings{10.1007/978-3-031-63646-2_11 , +author={Soto-Forero, Daniel and Betbeder, Marie-Laure and Henriet, Julien}, +editor={Recio-Garcia, Juan A. and Orozco-del-Castillo, Mauricio G. and Bridge, Derek}, +title={Ensemble Stacking Case-Based Reasoning for Regression}, +booktitle={Case-Based Reasoning Research and Development}, +year={2024}, +publisher={Springer Nature Switzerland}, +address={Cham}, +pages={159--174}, +abstract={This paper presents a case-based reasoning algorithm with a two-stage iterative double stacking to find approximate solutions to one and multidimensional regression problems. This approach does not require training, so it can work with dynamic data at run time. The solutions are generated using stochastic algorithms in order to allow exploration of the solution space. The evaluation is performed by transforming the regression problem into an optimization problem with an associated objective function. The algorithm has been tested in comparison with nine classical regression algorithms on ten different regression databases extracted from the UCI site. The results show that the proposed algorithm generates solutions in most cases quite close to the real solutions. According to the RMSE, the proposed algorithm globally among the four best algorithms, according to MAE, to the fourth best algorithms of the ten evaluated, suggesting that the results are reasonably good.}, +isbn={978-3-031-63646-2} +} + +@article{ZHANG2018189, +title = {A three learning states Bayesian knowledge tracing model}, +journal = {Knowledge-Based Systems}, +volume = {148}, +pages = {189-201}, +year = {2018}, +issn = {0950-7051}, +doi = {https://doi.org/10.1016/j.knosys.2018.03.001}, +url = {https://www.sciencedirect.com/science/article/pii/S0950705118301199}, +author = {Kai Zhang and Yiyu Yao}, +keywords = {Bayesian knowledge tracing, Three-way decisions}, +abstract = {This paper proposes a Bayesian knowledge tracing model with three learning states by extending the original two learning states. We divide a learning process into three sections by using an evaluation function for three-way decisions. Advantages of such a trisection over traditional bisection are demonstrated by comparative experiments. We develop a three learning states model based on the trisection of the learning process. We apply the model to a series of comparative experiments with the original model. Qualitative and quantitative analyses of the experimental results indicate the superior performance of the proposed model over the original model in terms of prediction accuracies and related statistical measures.} +} + +@article{Li_2024, +doi = {10.3847/1538-4357/ad3215}, +url = {https://dx.doi.org/10.3847/1538-4357/ad3215}, +year = {2024}, +month = {apr}, +publisher = {The American Astronomical Society}, +volume = {965}, +number = {2}, +pages = {125}, +author = {Zhigang Li and Zhejie Ding and Yu Yu and Pengjie Zhang}, +title = {The Kullback–Leibler Divergence and the Convergence Rate of Fast Covariance Matrix Estimators in Galaxy Clustering Analysis}, +journal = {The Astrophysical Journal}, +abstract = {We present a method to quantify the convergence rate of the fast estimators of the covariance matrices in the large-scale structure analysis. Our method is based on the Kullback–Leibler (KL) divergence, which describes the relative entropy of two probability distributions. As a case study, we analyze the delete-d jackknife estimator for the covariance matrix of the galaxy correlation function. We introduce the information factor or the normalized KL divergence with the help of a set of baseline covariance matrices to diagnose the information contained in the jackknife covariance matrix. Using a set of quick particle mesh mock catalogs designed for the Baryon Oscillation Spectroscopic Survey DR11 CMASS galaxy survey, we find that the jackknife resampling method succeeds in recovering the covariance matrix with 10 times fewer simulation mocks than that of the baseline method at small scales (s ≤ 40 h −1 Mpc). However, the ability to reduce the number of mock catalogs is degraded at larger scales due to the increasing bias on the jackknife covariance matrix. Note that the analysis in this paper can be applied to any fast estimator of the covariance matrix for galaxy clustering measurements.} +} + +@Article{Kim2024, +author={Kim, Wonjik}, +title={A Random Focusing Method with Jensen--Shannon Divergence for Improving Deep Neural Network Performance Ensuring Architecture Consistency}, +journal={Neural Processing Letters}, +year={2024}, +month={Jun}, +day={17}, +volume={56}, +number={4}, +pages={199}, +abstract={Multiple hidden layers in deep neural networks perform non-linear transformations, enabling the extraction of meaningful features and the identification of relationships between input and output data. However, the gap between the training and real-world data can result in network overfitting, prompting the exploration of various preventive methods. The regularization technique called 'dropout' is widely used for deep learning models to improve the training of robust and generalized features. During the training phase with dropout, neurons in a particular layer are randomly selected to be ignored for each input. This random exclusion of neurons encourages the network to depend on different subsets of neurons at different times, fostering robustness and reducing sensitivity to specific neurons. This study introduces a novel approach called random focusing, departing from complete neuron exclusion in dropout. The proposed random focusing selectively highlights random neurons during training, aiming for a smoother transition between training and inference phases while keeping network architecture consistent. This study also incorporates Jensen--Shannon Divergence to enhance the stability and efficacy of the random focusing method. Experimental validation across tasks like image classification and semantic segmentation demonstrates the adaptability of the proposed methods across different network architectures, including convolutional neural networks and transformers.}, +issn={1573-773X}, +doi={10.1007/s11063-024-11668-z}, +url={https://doi.org/10.1007/s11063-024-11668-z} +} + +@InProceedings{pmlr-v238-ou24a, + title = {Thompson Sampling Itself is Differentially Private}, + author = {Ou, Tingting and Cummings, Rachel and Avella Medina, Marco}, + booktitle = {Proceedings of The 27th International Conference on Artificial Intelligence and Statistics}, + pages = {1576--1584}, + year = {2024}, + editor = {Dasgupta, Sanjoy and Mandt, Stephan and Li, Yingzhen}, + volume = {238}, + series = {Proceedings of Machine Learning Research}, + month = {02--04 May}, + publisher = {PMLR}, + pdf = {https://proceedings.mlr.press/v238/ou24a/ou24a.pdf}, + url = {https://proceedings.mlr.press/v238/ou24a.html}, + abstract = {In this work we first show that the classical Thompson sampling algorithm for multi-arm bandits is differentially private as-is, without any modification. We provide per-round privacy guarantees as a function of problem parameters and show composition over $T$ rounds; since the algorithm is unchanged, existing $O(\sqrt{NT\log N})$ regret bounds still hold and there is no loss in performance due to privacy. We then show that simple modifications – such as pre-pulling all arms a fixed number of times, increasing the sampling variance – can provide tighter privacy guarantees. We again provide privacy guarantees that now depend on the new parameters introduced in the modification, which allows the analyst to tune the privacy guarantee as desired. We also provide a novel regret analysis for this new algorithm, and show how the new parameters also impact expected regret. Finally, we empirically validate and illustrate our theoretical findings in two parameter regimes and demonstrate that tuning the new parameters substantially improve the privacy-regret tradeoff.} +} + +@Article{math12111758, +AUTHOR = {Uguina, Antonio R. and Gomez, Juan F. and Panadero, Javier and Martínez-Gavara, Anna and Juan, Angel A.}, +TITLE = {A Learnheuristic Algorithm Based on Thompson Sampling for the Heterogeneous and Dynamic Team Orienteering Problem}, +JOURNAL = {Mathematics}, +VOLUME = {12}, +YEAR = {2024}, +NUMBER = {11}, +ARTICLE-NUMBER = {1758}, +URL = {https://www.mdpi.com/2227-7390/12/11/1758}, +ISSN = {2227-7390}, +ABSTRACT = {The team orienteering problem (TOP) is a well-studied optimization challenge in the field of Operations Research, where multiple vehicles aim to maximize the total collected rewards within a given time limit by visiting a subset of nodes in a network. With the goal of including dynamic and uncertain conditions inherent in real-world transportation scenarios, we introduce a novel dynamic variant of the TOP that considers real-time changes in environmental conditions affecting reward acquisition at each node. Specifically, we model the dynamic nature of environmental factors—such as traffic congestion, weather conditions, and battery level of each vehicle—to reflect their impact on the probability of obtaining the reward when visiting each type of node in a heterogeneous network. To address this problem, a learnheuristic optimization framework is proposed. It combines a metaheuristic algorithm with Thompson sampling to make informed decisions in dynamic environments. Furthermore, we conduct empirical experiments to assess the impact of varying reward probabilities on resource allocation and route planning within the context of this dynamic TOP, where nodes might offer a different reward behavior depending upon the environmental conditions. Our numerical results indicate that the proposed learnheuristic algorithm outperforms static approaches, achieving up to 25% better performance in highly dynamic scenarios. Our findings highlight the effectiveness of our approach in adapting to dynamic conditions and optimizing decision-making processes in transportation systems.}, +DOI = {10.3390/math12111758} +} + +@inproceedings{NEURIPS2023_9d8cf124, + author = {Abel, David and Barreto, Andre and Van Roy, Benjamin and Precup, Doina and van Hasselt, Hado P and Singh, Satinder}, + booktitle = {Advances in Neural Information Processing Systems}, + editor = {A. Oh and T. Naumann and A. Globerson and K. Saenko and M. Hardt and S. Levine}, + pages = {50377--50407}, + publisher = {Curran Associates, Inc.}, + title = {A Definition of Continual Reinforcement Learning}, + url = {https://proceedings.neurips.cc/paper_files/paper/2023/file/9d8cf1247786d6dfeefeeb53b8b5f6d7-Paper-Conference.pdf}, + volume = {36}, + year = {2023} +} + +@article{NGUYEN2024111566, +title = {Dynamic metaheuristic selection via Thompson Sampling for online optimization}, +journal = {Applied Soft Computing}, +volume = {158}, +pages = {111566}, +year = {2024}, +issn = {1568-4946}, +doi = {https://doi.org/10.1016/j.asoc.2024.111566}, +url = {https://www.sciencedirect.com/science/article/pii/S1568494624003405}, +author = {Alain Nguyen}, +keywords = {Selection hyper-heuristic, Multi-armed-bandit, Thompson Sampling, Online optimization}, +abstract = {It is acknowledged that no single heuristic can outperform all the others in every optimization problem. This has given rise to hyper-heuristic methods for providing solutions to a wider range of problems. In this work, a set of five non-competing low-level heuristics is proposed in a hyper-heuristic framework. The multi-armed bandit problem analogy is efficiently leveraged and Thompson Sampling is used to actively select the best heuristic for online optimization. The proposed method is compared against ten population-based metaheuristic algorithms on the well-known CEC’05 optimizing benchmark consisting of 23 functions of various landscapes. The results show that the proposed algorithm is the only one able to find the global minimum of all functions with remarkable consistency.} +} + +@Article{Malladi2024, +author={Malladi, Rama K.}, +title={Application of Supervised Machine Learning Techniques to Forecast the COVID-19 U.S. Recession and Stock Market Crash}, +journal={Computational Economics}, +year={2024}, +month={Mar}, +day={01}, +volume={63}, +number={3}, +pages={1021-1045}, +abstract={Machine learning (ML), a transformational technology, has been successfully applied to forecasting events down the road. This paper demonstrates that supervised ML techniques can be used in recession and stock market crash (more than 20{\%} drawdown) forecasting. After learning from strictly past monthly data, ML algorithms detected the Covid-19 recession by December 2019, six months before the official NBER announcement. Moreover, ML algorithms foresaw the March 2020 S{\&}P500 crash two months before it happened. The current labor market and housing are harbingers of a future U.S. recession (in 3 months). Financial factors have a bigger role to play in stock market crashes than economic factors. The labor market appears as a top-two feature in predicting both recessions and crashes. ML algorithms detect that the U.S. exited recession before December 2020, even though the official NBER announcement has not yet been made. They also do not anticipate a U.S. stock market crash before March 2021. ML methods have three times higher false discovery rates of recessions compared to crashes.}, +issn={1572-9974}, +doi={10.1007/s10614-022-10333-8}, +url={https://doi.org/10.1007/s10614-022-10333-8} +} + +@INPROCEEDINGS{10493943, + author={Raaa Subha and Naaa Gayathri and Saaa Sasireka and Raaa Sathiyabanu and Baaa Santhiyaa and Baaa Varshini}, + booktitle={2024 5th International Conference on Mobile Computing and Sustainable Informatics (ICMCSI)}, + title={Intelligent Tutoring Systems using Long Short-Term Memory Networks and Bayesian Knowledge Tracing}, + year={2024}, + volume={0}, + number={0}, + pages={24-29}, + keywords={Knowledge engineering;Filtering;Estimation;Transforms;Real-time systems;Bayes methods;Problem-solving;Intelligent Tutoring System (ITS);Long Short-Term Memory (LSTM);Bayesian Knowledge Tracing (BKT);Reinforcement Learning}, + doi={10.1109/ICMCSI61536.2024.00010} +} + +@article{https://doi.org/10.1155/2024/4067721, +author = {Ahmed, Esmael}, +title = {Student Performance Prediction Using Machine Learning Algorithms}, +journal = {Applied Computational Intelligence and Soft Computing}, +volume = {2024}, +number = {1}, +pages = {4067721}, +doi = {https://doi.org/10.1155/2024/4067721}, +url = {https://onlinelibrary.wiley.com/doi/abs/10.1155/2024/4067721}, +eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1155/2024/4067721}, +abstract = {Education is crucial for a productive life and providing necessary resources. With the advent of technology like artificial intelligence, higher education institutions are incorporating technology into traditional teaching methods. Predicting academic success has gained interest in education as a strong academic record improves a university’s ranking and increases student employment opportunities. Modern learning institutions face challenges in analyzing performance, providing high-quality education, formulating strategies for evaluating students’ performance, and identifying future needs. E-learning is a rapidly growing and advanced form of education, where students enroll in online courses. Platforms like Intelligent Tutoring Systems (ITS), learning management systems (LMS), and massive open online courses (MOOC) use educational data mining (EDM) to develop automatic grading systems, recommenders, and adaptative systems. However, e-learning is still considered a challenging learning environment due to the lack of direct interaction between students and course instructors. Machine learning (ML) is used in developing adaptive intelligent systems that can perform complex tasks beyond human abilities. Some areas of applications of ML algorithms include cluster analysis, pattern recognition, image processing, natural language processing, and medical diagnostics. In this research work, K-means, a clustering data mining technique using Davies’ Bouldin method, obtains clusters to find important features affecting students’ performance. The study found that the SVM algorithm had the best prediction results after parameter adjustment, with a 96\% accuracy rate. In this paper, the researchers have examined the functions of the Support Vector Machine, Decision Tree, naive Bayes, and KNN classifiers. The outcomes of parameter adjustment greatly increased the accuracy of the four prediction models. Naïve Bayes model’s prediction accuracy is the lowest when compared to other prediction methods, as it assumes a strong independent relationship between features.}, +year = {2024} +} + +@Inproceedings{Arthurs, +author={Arthurs, Noah and Stenhaug, Ben and Karayev, Sergey and Piech, Chris}, +booktitle={International Conference on Educational Data Mining (EDM)}, +title={Grades Are Not Normal: Improving Exam Score Models Using the Logit-Normal Distribution}, +year={2019}, +type={article}, +language={English}, +volume={}, +number={}, +pages={6}, +url={https://eric.ed.gov/?id=ED599204} +} + +@article{HAZEM, +author = {Hazem A. Alrakhawi and Nurullizam Jamiat and Samy S. Abu-Naser}, +title = {Intelligent Tutoring Systems in education: A systematic review of usage, tools, effects and evaluation}, +journal = {Journal of Theoretical and Applied Information Technology}, +volume = {2023}, +number = {4}, +pages = {4067721}, +doi = {}, +url = {}, +abstract = {}, +year = {2023} +} + +@Article{Liu2023, +author={Liu, Mengchi +and Yu, Dongmei}, +title={Towards intelligent E-learning systems}, +journal={Education and Information Technologies}, +year={2023}, +month={Jul}, +day={01}, +volume={28}, +number={7}, +pages={7845-7876}, +abstract={The prevalence of e-learning systems has made educational resources more accessible, interactive and effective to learners without the geographic and temporal boundaries. However, as the number of users increases and the volume of data grows, current e-learning systems face some technical and pedagogical challenges. This paper provides a comprehensive review on the efforts of applying new information and communication technologies to improve e-learning services. We first systematically investigate current e-learning systems in terms of their classification, architecture, functions, challenges, and current trends. We then present a general architecture for big data based e-learning systems to meet the ever-growing demand for e-learning. We also describe how to use data generated in big data based e-learning systems to support more flexible and customized course delivery and personalized learning.}, +issn={1573-7608}, +doi={10.1007/s10639-022-11479-6}, +url={https://doi.org/10.1007/s10639-022-11479-6} +} + +@InProceedings{10.1007/978-3-031-63646-2_13, +author="Soto-Forero, Daniel +and Ackermann, Simha +and Betbeder, Marie-Laure +and Henriet, Julien", +editor="Recio-Garcia, Juan A. +and Orozco-del-Castillo, Mauricio G. +and Bridge, Derek", +title="The Intelligent Tutoring System AI-VT with Case-Based Reasoning and Real Time Recommender Models", +booktitle="Case-Based Reasoning Research and Development", +year="2024", +publisher="Springer Nature Switzerland", +address="Cham", +pages="191--205", +abstract="This paper presents a recommendation model coupled on an existing CBR system model through a new modular architecture designed to integrate multiple services in a learning system called AI-VT (Artificial Intelligence Training System). The recommendation model provides a semi-automatic review of the CBR, two variants of the recommendation model have been implemented: deterministic and stochastic. The model has been tested with 1000 simulated learners, and compared with an original CBR system and BKT (Bayesian Knowledge Tracing) recommender system. The results show that the proposed model identifies learners' weaknesses correctly and revises the content of the ITS (Intelligent Tutoring System) better than the original ITS with CBR. Compared to BKT, the results at each level of complexity are variable, but overall the proposed stochastic model obtains better results.", +isbn="978-3-031-63646-2" +} + +@article{doi:10.1137/23M1592420, +author = {Minsker, Stanislav and Strawn, Nate}, +title = {The Geometric Median and Applications to Robust Mean Estimation}, +journal = {SIAM Journal on Mathematics of Data Science}, +volume = {6}, +number = {2}, +pages = {504-533}, +year = {2024}, +doi = {10.1137/23M1592420}, +URL = { https://doi.org/10.1137/23M1592420}, +eprint = {https://doi.org/10.1137/23M1592420}, +abstract = { Abstract.This paper is devoted to the statistical and numerical properties of the geometric median and its applications to the problem of robust mean estimation via the median of means principle. Our main theoretical results include (a) an upper bound for the distance between the mean and the median for general absolutely continuous distributions in \(\mathbb R^d\), and examples of specific classes of distributions for which these bounds do not depend on the ambient dimension \(d\); (b) exponential deviation inequalities for the distance between the sample and the population versions of the geometric median, which again depend only on the trace-type quantities and not on the ambient dimension. As a corollary, we deduce improved bounds for the (geometric) median of means estimator that hold for large classes of heavy-tailed distributions. Finally, we address the error of numerical approximation, which is an important practical aspect of any statistical estimation procedure. We demonstrate that the objective function minimized by the geometric median satisfies a “local quadratic growth” condition that allows one to translate suboptimality bounds for the objective function to the corresponding bounds for the numerical approximation to the median itself and propose a simple stopping rule applicable to any optimization method which yields explicit error guarantees. We conclude with the numerical experiments, including the application to estimation of mean values of log-returns for S\&P 500 data. } +} + +@article{lei2024analysis, + title={Analysis of Simpson’s Paradox and Its Applications}, + author={Lei, Zhihao}, + journal={Highlights in Science, Engineering and Technology}, + volume={88}, + pages={357--362}, + year={2024} +} + +@InProceedings{pmlr-v108-seznec20a, + title = {A single algorithm for both restless and rested rotting bandits}, + author = {Seznec, Julien and Menard, Pierre and Lazaric, Alessandro and Valko, Michal}, + booktitle = {Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics}, + pages = {3784--3794}, + year = {2020}, + editor = {Chiappa, Silvia and Calandra, Roberto}, + volume = {108}, + series = {Proceedings of Machine Learning Research}, + month = {26--28 Aug}, + publisher = {PMLR}, + pdf = {http://proceedings.mlr.press/v108/seznec20a/seznec20a.pdf}, + url = {https://proceedings.mlr.press/v108/seznec20a.html}, + abstract = {In many application domains (e.g., recommender systems, intelligent tutoring systems), the rewards associated to the available actions tend to decrease over time. This decay is either caused by the actions executed in the past (e.g., a user may get bored when songs of the same genre are recommended over and over) or by an external factor (e.g., content becomes outdated). These two situations can be modeled as specific instances of the rested and restless bandit settings, where arms are rotting (i.e., their value decrease over time). These problems were thought to be significantly different, since Levine et al. (2017) showed that state-of-the-art algorithms for restless bandit perform poorly in the rested rotting setting. In this paper, we introduce a novel algorithm, Rotting Adaptive Window UCB (RAW-UCB), that achieves near-optimal regret in both rotting rested and restless bandit, without any prior knowledge of the setting (rested or restless) and the type of non-stationarity (e.g., piece-wise constant, bounded variation). This is in striking contrast with previous negative results showing that no algorithm can achieve similar results as soon as rewards are allowed to increase. We confirm our theoretical findings on a number of synthetic and dataset-based experiments.} +} diff --git a/main.blg b/main.blg new file mode 100644 index 0000000..5836276 --- /dev/null +++ b/main.blg @@ -0,0 +1,49 @@ +This is BibTeX, Version 0.99d (TeX Live 2023) +Capacity: max_strings=200000, hash_size=200000, hash_prime=170003 +The top-level auxiliary file: main.aux +The style file: apalike.bst +Database file #1: main.bib +Warning--entry type for "Marie" isn't style-file defined +--line 591 of file main.bib +You've used 28 entries, + 1935 wiz_defined-function locations, + 660 strings with 9503 characters, +and the built_in function-call counts, 11836 in all, are: += -- 1120 +> -- 556 +< -- 15 ++ -- 204 +- -- 188 +* -- 1030 +:= -- 2077 +add.period$ -- 87 +call.type$ -- 28 +change.case$ -- 229 +chr.to.int$ -- 28 +cite$ -- 28 +duplicate$ -- 424 +empty$ -- 798 +format.name$ -- 230 +if$ -- 2311 +int.to.chr$ -- 1 +int.to.str$ -- 0 +missing$ -- 31 +newline$ -- 143 +num.names$ -- 88 +pop$ -- 196 +preamble$ -- 1 +purify$ -- 232 +quote$ -- 0 +skip$ -- 320 +stack$ -- 0 +substring$ -- 802 +swap$ -- 54 +text.length$ -- 4 +text.prefix$ -- 0 +top$ -- 0 +type$ -- 160 +warning$ -- 0 +while$ -- 84 +width$ -- 0 +write$ -- 367 +(There was 1 warning) diff --git a/main.log b/main.log new file mode 100644 index 0000000..7961327 --- /dev/null +++ b/main.log @@ -0,0 +1,1556 @@ +This is pdfTeX, Version 3.141592653-2.6-1.40.25 (TeX Live 2023) (preloaded format=pdflatex 2023.5.31) 21 FEB 2025 14:14 +entering extended mode + restricted \write18 enabled. + %&-line parsing enabled. +**main.tex +(./main.tex +LaTeX2e <2022-11-01> patch level 1 +L3 programming layer <2023-05-22> (./spimufcphdthesis.cls +Document Class: spimufcphdthesis 2022/02/10 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-docum +ent.cls +Document Class: upmethodology-document 2022/10/04 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-p-com +mon.sty +Package: upmethodology-p-common 2022/02/10 + (/usr/local/texlive/2023/texmf-dist/tex/latex/base/ifthen.sty +Package: ifthen 2022/04/13 v1.1d Standard LaTeX ifthen package (DPC) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/tools/xspace.sty +Package: xspace 2014/10/28 v1.13 Space after command names (DPC,MH) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/xcolor/xcolor.sty +Package: xcolor 2022/06/12 v2.14 LaTeX color extensions (UK) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics-cfg/color.cfg +File: color.cfg 2016/01/02 v1.6 sample color configuration +) +Package xcolor Info: Driver file: pdftex.def on input line 227. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics-def/pdftex.def +File: pdftex.def 2022/09/22 v1.2b Graphics/color driver for pdftex +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics/mathcolor.ltx) +Package xcolor Info: Model `cmy' substituted by `cmy0' on input line 1353. +Package xcolor Info: Model `hsb' substituted by `rgb' on input line 1357. +Package xcolor Info: Model `RGB' extended on input line 1369. +Package xcolor Info: Model `HTML' substituted by `rgb' on input line 1371. +Package xcolor Info: Model `Hsb' substituted by `hsb' on input line 1372. +Package xcolor Info: Model `tHsb' substituted by `hsb' on input line 1373. +Package xcolor Info: Model `HSB' substituted by `hsb' on input line 1374. +Package xcolor Info: Model `Gray' substituted by `gray' on input line 1375. +Package xcolor Info: Model `wave' substituted by `hsb' on input line 1376. +) +(/usr/local/texlive/2023/texmf-dist/tex/generic/iftex/ifpdf.sty +Package: ifpdf 2019/10/25 v3.4 ifpdf legacy package. Use iftex instead. + +(/usr/local/texlive/2023/texmf-dist/tex/generic/iftex/iftex.sty +Package: iftex 2022/02/03 v1.0f TeX engine tests +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/UPMVERSION.def)) +*********** UPMETHODOLOGY BOOK CLASS (WITH PART AND CHAPTER) +(/usr/local/texlive/2023/texmf-dist/tex/latex/base/book.cls +Document Class: book 2022/07/02 v1.4n Standard LaTeX document class +(/usr/local/texlive/2023/texmf-dist/tex/latex/base/bk11.clo +File: bk11.clo 2022/07/02 v1.4n Standard LaTeX file (size option) +) +\c@part=\count185 +\c@chapter=\count186 +\c@section=\count187 +\c@subsection=\count188 +\c@subsubsection=\count189 +\c@paragraph=\count190 +\c@subparagraph=\count191 +\c@figure=\count192 +\c@table=\count193 +\abovecaptionskip=\skip48 +\belowcaptionskip=\skip49 +\bibindent=\dimen140 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/a4wide/a4wide.sty +Package: a4wide 1994/08/30 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/ntgclass/a4.sty +Package: a4 2023/01/10 v1.2g A4 based page layout +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-docum +ent.sty +Package: upmethodology-document 2022/02/10 + **** upmethodology-document is using French language **** +(/usr/local/texlive/2023/texmf-dist/tex/generic/babel/babel.sty +Package: babel 2023/05/11 v3.89 The Babel package +\babel@savecnt=\count194 +\U@D=\dimen141 +\l@unhyphenated=\language87 + +(/usr/local/texlive/2023/texmf-dist/tex/generic/babel/txtbabel.def) +\bbl@readstream=\read2 +\bbl@dirlevel=\count195 +Package babel Info: Main language set with 'main='. Except if you have +(babel) problems, prefer the default mechanism for setting +(babel) the main language, ie, as the last declared. +(babel) Reported on input line 4170. + +(/usr/local/texlive/2023/texmf-dist/tex/generic/babel-french/french.ldf +Language: french 2023/03/08 v3.5q French support from the babel system +Package babel Info: Hyphen rules for 'acadian' set to \l@french +(babel) (\language29). Reported on input line 91. +Package babel Info: Hyphen rules for 'canadien' set to \l@french +(babel) (\language29). Reported on input line 92. +\FB@nonchar=\count196 +Package babel Info: Making : an active character on input line 395. +Package babel Info: Making ; an active character on input line 396. +Package babel Info: Making ! an active character on input line 397. +Package babel Info: Making ? an active character on input line 398. +\FBguill@level=\count197 +\FBold@everypar=\toks16 +\FB@Mht=\dimen142 +\mc@charclass=\count198 +\mc@charfam=\count199 +\mc@charslot=\count266 +\std@mcc=\count267 +\dec@mcc=\count268 +\FB@parskip=\dimen143 +\listindentFB=\dimen144 +\descindentFB=\dimen145 +\labelindentFB=\dimen146 +\labelwidthFB=\dimen147 +\leftmarginFB=\dimen148 +\parindentFFN=\dimen149 +\FBfnindent=\dimen150 +)) +(/usr/local/texlive/2023/texmf-dist/tex/generic/babel/locale/fr/babel-french.te +x +Package babel Info: Importing font and identification data for french +(babel) from babel-fr.ini. Reported on input line 11. +) (/usr/local/texlive/2023/texmf-dist/tex/latex/carlisle/scalefnt.sty) +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics/keyval.sty +Package: keyval 2022/05/29 v1.15 key=value parser (DPC) +\KV@toks@=\toks17 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/vmargin/vmargin.sty +Package: vmargin 2004/07/15 V2.5 set document margins (VK) + +Package: vmargin 2004/07/15 V2.5 set document margins (VK) +\PaperWidth=\dimen151 +\PaperHeight=\dimen152 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-exten +sion.sty +Package: upmethodology-extension 2022/02/10 +\upmext@tmp@putx=\skip50 + *** define the extension value frontillustrationsize **** +*** define the extension value watermarksize **** +*** undefine the extension value publisher **** +*** undefine the extension value copyrighter **** +*** undefine the extension value printedin ****) +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-fmt.s +ty +Package: upmethodology-fmt 2022/10/04 + **** upmethodology-fmt is using French language **** +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics/graphicx.sty +Package: graphicx 2021/09/16 v1.2d Enhanced LaTeX Graphics (DPC,SPQR) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics/graphics.sty +Package: graphics 2022/03/10 v1.4e Standard LaTeX Graphics (DPC,SPQR) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics/trig.sty +Package: trig 2021/08/11 v1.11 sin cos tan (DPC) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/graphics-cfg/graphics.cfg +File: graphics.cfg 2016/06/04 v1.11 sample graphics configuration +) +Package graphics Info: Driver file: pdftex.def on input line 107. +) +\Gin@req@height=\dimen153 +\Gin@req@width=\dimen154 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/caption/subcaption.sty +Package: subcaption 2023/02/19 v1.6 Sub-captions (AR) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/caption/caption.sty +Package: caption 2023/03/12 v3.6j Customizing captions (AR) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/caption/caption3.sty +Package: caption3 2023/03/12 v2.4 caption3 kernel (AR) +\caption@tempdima=\dimen155 +\captionmargin=\dimen156 +\caption@leftmargin=\dimen157 +\caption@rightmargin=\dimen158 +\caption@width=\dimen159 +\caption@indent=\dimen160 +\caption@parindent=\dimen161 +\caption@hangindent=\dimen162 +Package caption Info: Standard document class detected. +Package caption Info: french babel package is loaded. +) +\c@caption@flags=\count269 +\c@continuedfloat=\count270 +) +Package caption Info: New subtype `subfigure' on input line 239. +\c@subfigure=\count271 +Package caption Info: New subtype `subtable' on input line 239. +\c@subtable=\count272 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/tools/tabularx.sty +Package: tabularx 2020/01/15 v2.11c `tabularx' package (DPC) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/tools/array.sty +Package: array 2022/09/04 v2.5g Tabular extension package (FMi) +\col@sep=\dimen163 +\ar@mcellbox=\box51 +\extrarowheight=\dimen164 +\NC@list=\toks18 +\extratabsurround=\skip51 +\backup@length=\skip52 +\ar@cellbox=\box52 +) +\TX@col@width=\dimen165 +\TX@old@table=\dimen166 +\TX@old@col=\dimen167 +\TX@target=\dimen168 +\TX@delta=\dimen169 +\TX@cols=\count273 +\TX@ftn=\toks19 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/tools/multicol.sty +Package: multicol 2021/11/30 v1.9d multicolumn formatting (FMi) +\c@tracingmulticols=\count274 +\mult@box=\box53 +\multicol@leftmargin=\dimen170 +\c@unbalance=\count275 +\c@collectmore=\count276 +\doublecol@number=\count277 +\multicoltolerance=\count278 +\multicolpretolerance=\count279 +\full@width=\dimen171 +\page@free=\dimen172 +\premulticols=\dimen173 +\postmulticols=\dimen174 +\multicolsep=\skip53 +\multicolbaselineskip=\skip54 +\partial@page=\box54 +\last@line=\box55 +\maxbalancingoverflow=\dimen175 +\mult@rightbox=\box56 +\mult@grightbox=\box57 +\mult@firstbox=\box58 +\mult@gfirstbox=\box59 +\@tempa=\box60 +\@tempa=\box61 +\@tempa=\box62 +\@tempa=\box63 +\@tempa=\box64 +\@tempa=\box65 +\@tempa=\box66 +\@tempa=\box67 +\@tempa=\box68 +\@tempa=\box69 +\@tempa=\box70 +\@tempa=\box71 +\@tempa=\box72 +\@tempa=\box73 +\@tempa=\box74 +\@tempa=\box75 +\@tempa=\box76 +\@tempa=\box77 +\@tempa=\box78 +\@tempa=\box79 +\@tempa=\box80 +\@tempa=\box81 +\@tempa=\box82 +\@tempa=\box83 +\@tempa=\box84 +\@tempa=\box85 +\@tempa=\box86 +\@tempa=\box87 +\@tempa=\box88 +\@tempa=\box89 +\@tempa=\box90 +\@tempa=\box91 +\@tempa=\box92 +\@tempa=\box93 +\@tempa=\box94 +\@tempa=\box95 +\c@minrows=\count280 +\c@columnbadness=\count281 +\c@finalcolumnbadness=\count282 +\last@try=\dimen176 +\multicolovershoot=\dimen177 +\multicolundershoot=\dimen178 +\mult@nat@firstbox=\box96 +\colbreak@box=\box97 +\mc@col@check@num=\count283 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/colortbl/colortbl.sty +Package: colortbl 2022/06/20 v1.0f Color table columns (DPC) +\everycr=\toks20 +\minrowclearance=\skip55 +\rownum=\count284 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/picinpar/picinpar.sty +Pictures in Paragraphs. Version 1.3, November 22, 2022 +\br=\count285 +\bl=\count286 +\na=\count287 +\nb=\count288 +\tcdsav=\count289 +\tcl=\count290 +\tcd=\count291 +\tcn=\count292 +\cumtcl=\count293 +\cumpartcl=\count294 +\lftside=\dimen179 +\rtside=\dimen180 +\hpic=\dimen181 +\vpic=\dimen182 +\strutilg=\dimen183 +\picwd=\dimen184 +\topheight=\dimen185 +\ilg=\dimen186 +\lpic=\dimen187 +\lwindowsep=\dimen188 +\rwindowsep=\dimen189 +\cumpar=\dimen190 +\twa=\toks21 +\la=\toks22 +\ra=\toks23 +\ha=\toks24 +\pictoc=\toks25 +\rawtext=\box98 +\holder=\box99 +\windowbox=\box100 +\wartext=\box101 +\finaltext=\box102 +\aslice=\box103 +\bslice=\box104 +\wbox=\box105 +\wstrutbox=\box106 +\picbox=\box107 +\waslice=\box108 +\wbslice=\box109 +\fslice=\box110 +) (/usr/local/texlive/2023/texmf-dist/tex/latex/amsmath/amsmath.sty +Package: amsmath 2022/04/08 v2.17n AMS math features +\@mathmargin=\skip56 + +For additional information on amsmath, use the `?' option. +(/usr/local/texlive/2023/texmf-dist/tex/latex/amsmath/amstext.sty +Package: amstext 2021/08/26 v2.01 AMS text + +(/usr/local/texlive/2023/texmf-dist/tex/latex/amsmath/amsgen.sty +File: amsgen.sty 1999/11/30 v2.0 generic functions +\@emptytoks=\toks26 +\ex@=\dimen191 +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/amsmath/amsbsy.sty +Package: amsbsy 1999/11/29 v1.2d Bold Symbols +\pmbraise@=\dimen192 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/amsmath/amsopn.sty +Package: amsopn 2022/04/08 v2.04 operator names +) +\inf@bad=\count295 +LaTeX Info: Redefining \frac on input line 234. +\uproot@=\count296 +\leftroot@=\count297 +LaTeX Info: Redefining \overline on input line 399. +LaTeX Info: Redefining \colon on input line 410. +\classnum@=\count298 +\DOTSCASE@=\count299 +LaTeX Info: Redefining \ldots on input line 496. +LaTeX Info: Redefining \dots on input line 499. +LaTeX Info: Redefining \cdots on input line 620. +\Mathstrutbox@=\box111 +\strutbox@=\box112 +LaTeX Info: Redefining \big on input line 722. +LaTeX Info: Redefining \Big on input line 723. +LaTeX Info: Redefining \bigg on input line 724. +LaTeX Info: Redefining \Bigg on input line 725. +\big@size=\dimen193 +LaTeX Font Info: Redeclaring font encoding OML on input line 743. +LaTeX Font Info: Redeclaring font encoding OMS on input line 744. +\macc@depth=\count300 +LaTeX Info: Redefining \bmod on input line 905. +LaTeX Info: Redefining \pmod on input line 910. +LaTeX Info: Redefining \smash on input line 940. +LaTeX Info: Redefining \relbar on input line 970. +LaTeX Info: Redefining \Relbar on input line 971. +\c@MaxMatrixCols=\count301 +\dotsspace@=\muskip16 +\c@parentequation=\count302 +\dspbrk@lvl=\count303 +\tag@help=\toks27 +\row@=\count304 +\column@=\count305 +\maxfields@=\count306 +\andhelp@=\toks28 +\eqnshift@=\dimen194 +\alignsep@=\dimen195 +\tagshift@=\dimen196 +\tagwidth@=\dimen197 +\totwidth@=\dimen198 +\lineht@=\dimen199 +\@envbody=\toks29 +\multlinegap=\skip57 +\multlinetaggap=\skip58 +\mathdisplay@stack=\toks30 +LaTeX Info: Redefining \[ on input line 2953. +LaTeX Info: Redefining \] on input line 2954. +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/amscls/amsthm.sty +Package: amsthm 2020/05/29 v2.20.6 +\thm@style=\toks31 +\thm@bodyfont=\toks32 +\thm@headfont=\toks33 +\thm@notefont=\toks34 +\thm@headpunct=\toks35 +\thm@preskip=\skip59 +\thm@postskip=\skip60 +\thm@headsep=\skip61 +\dth@everypar=\toks36 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/thmtools.sty +Package: thmtools 2023/05/04 v0.76 +\thmt@toks=\toks37 +\c@thmt@dummyctr=\count307 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/thm-patch.sty +Package: thm-patch 2023/05/04 v0.76 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/parseargs.sty +Package: parseargs 2023/05/04 v0.76 +\@parsespec=\toks38 +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/thm-kv.sty +Package: thm-kv 2023/05/04 v0.76 +Package thm-kv Info: Theorem names will be uppercased on input line 42. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/kvsetkeys/kvsetkeys.sty +Package: kvsetkeys 2022-10-05 v1.19 Key value parser (HO) +) +Package thm-kv Info: kvsetkeys patch (v1.16 or later) on input line 158. +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/thm-autoref.sty +Package: thm-autoref 2023/05/04 v0.76 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/aliasctr.sty +Package: aliasctr 2023/05/04 v0.76 +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/thm-listof.sty +Package: thm-listof 2023/05/04 v0.76 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/thm-restate.sty +Package: thm-restate 2023/05/04 v0.76 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/thmtools/thm-amsthm.sty +Package: thm-amsthm 2023/05/04 v0.76 +\thmt@style@headstyle=\toks39 +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/psnfss/pifont.sty +Package: pifont 2020/03/25 PSNFSS-v9.3 Pi font support (SPQR) +LaTeX Font Info: Trying to load font information for U+pzd on input line 63. + + +(/usr/local/texlive/2023/texmf-dist/tex/latex/psnfss/upzd.fd +File: upzd.fd 2001/06/04 font definitions for U/pzd. +) +LaTeX Font Info: Trying to load font information for U+psy on input line 64. + + +(/usr/local/texlive/2023/texmf-dist/tex/latex/psnfss/upsy.fd +File: upsy.fd 2001/06/04 font definitions for U/psy. +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/setspace/setspace.sty +Package: setspace 2022/12/04 v6.7b set line spacing +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/tools/varioref.sty +Package: varioref 2022/01/09 v1.6f package for extended references (FMi) +\c@vrcnt=\count308 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/txfonts.sty +Package: txfonts 2008/01/22 v3.2.1 +LaTeX Font Info: Redeclaring symbol font `operators' on input line 21. +LaTeX Font Info: Overwriting symbol font `operators' in version `normal' +(Font) OT1/cmr/m/n --> OT1/txr/m/n on input line 21. +LaTeX Font Info: Overwriting symbol font `operators' in version `bold' +(Font) OT1/cmr/bx/n --> OT1/txr/m/n on input line 21. +LaTeX Font Info: Overwriting symbol font `operators' in version `bold' +(Font) OT1/txr/m/n --> OT1/txr/bx/n on input line 22. +\symitalic=\mathgroup4 +LaTeX Font Info: Overwriting symbol font `italic' in version `bold' +(Font) OT1/txr/m/it --> OT1/txr/bx/it on input line 26. +LaTeX Font Info: Redeclaring math alphabet \mathbf on input line 29. +LaTeX Font Info: Overwriting math alphabet `\mathbf' in version `normal' +(Font) OT1/cmr/bx/n --> OT1/txr/bx/n on input line 29. +LaTeX Font Info: Overwriting math alphabet `\mathbf' in version `bold' +(Font) OT1/cmr/bx/n --> OT1/txr/bx/n on input line 29. +LaTeX Font Info: Redeclaring math alphabet \mathit on input line 30. +LaTeX Font Info: Overwriting math alphabet `\mathit' in version `normal' +(Font) OT1/cmr/m/it --> OT1/txr/m/it on input line 30. +LaTeX Font Info: Overwriting math alphabet `\mathit' in version `bold' +(Font) OT1/cmr/bx/it --> OT1/txr/m/it on input line 30. +LaTeX Font Info: Overwriting math alphabet `\mathit' in version `bold' +(Font) OT1/txr/m/it --> OT1/txr/bx/it on input line 31. +LaTeX Font Info: Redeclaring math alphabet \mathsf on input line 40. +LaTeX Font Info: Overwriting math alphabet `\mathsf' in version `normal' +(Font) OT1/cmss/m/n --> OT1/txss/m/n on input line 40. +LaTeX Font Info: Overwriting math alphabet `\mathsf' in version `bold' +(Font) OT1/cmss/bx/n --> OT1/txss/m/n on input line 40. +LaTeX Font Info: Overwriting math alphabet `\mathsf' in version `bold' +(Font) OT1/txss/m/n --> OT1/txss/b/n on input line 41. +LaTeX Font Info: Redeclaring math alphabet \mathtt on input line 50. +LaTeX Font Info: Overwriting math alphabet `\mathtt' in version `normal' +(Font) OT1/cmtt/m/n --> OT1/txtt/m/n on input line 50. +LaTeX Font Info: Overwriting math alphabet `\mathtt' in version `bold' +(Font) OT1/cmtt/m/n --> OT1/txtt/m/n on input line 50. +LaTeX Font Info: Overwriting math alphabet `\mathtt' in version `bold' +(Font) OT1/txtt/m/n --> OT1/txtt/b/n on input line 51. +LaTeX Font Info: Redeclaring symbol font `letters' on input line 58. +LaTeX Font Info: Overwriting symbol font `letters' in version `normal' +(Font) OML/cmm/m/it --> OML/txmi/m/it on input line 58. +LaTeX Font Info: Overwriting symbol font `letters' in version `bold' +(Font) OML/cmm/b/it --> OML/txmi/m/it on input line 58. +LaTeX Font Info: Overwriting symbol font `letters' in version `bold' +(Font) OML/txmi/m/it --> OML/txmi/bx/it on input line 59. +\symlettersA=\mathgroup5 +LaTeX Font Info: Overwriting symbol font `lettersA' in version `bold' +(Font) U/txmia/m/it --> U/txmia/bx/it on input line 67. +LaTeX Font Info: Redeclaring symbol font `symbols' on input line 77. +LaTeX Font Info: Overwriting symbol font `symbols' in version `normal' +(Font) OMS/cmsy/m/n --> OMS/txsy/m/n on input line 77. +LaTeX Font Info: Overwriting symbol font `symbols' in version `bold' +(Font) OMS/cmsy/b/n --> OMS/txsy/m/n on input line 77. +LaTeX Font Info: Overwriting symbol font `symbols' in version `bold' +(Font) OMS/txsy/m/n --> OMS/txsy/bx/n on input line 78. +\symAMSa=\mathgroup6 +LaTeX Font Info: Overwriting symbol font `AMSa' in version `bold' +(Font) U/txsya/m/n --> U/txsya/bx/n on input line 94. +\symAMSb=\mathgroup7 +LaTeX Font Info: Overwriting symbol font `AMSb' in version `bold' +(Font) U/txsyb/m/n --> U/txsyb/bx/n on input line 103. +\symsymbolsC=\mathgroup8 +LaTeX Font Info: Overwriting symbol font `symbolsC' in version `bold' +(Font) U/txsyc/m/n --> U/txsyc/bx/n on input line 113. +LaTeX Font Info: Redeclaring symbol font `largesymbols' on input line 120. +LaTeX Font Info: Overwriting symbol font `largesymbols' in version `normal' +(Font) OMX/cmex/m/n --> OMX/txex/m/n on input line 120. +LaTeX Font Info: Overwriting symbol font `largesymbols' in version `bold' +(Font) OMX/cmex/m/n --> OMX/txex/m/n on input line 120. +LaTeX Font Info: Overwriting symbol font `largesymbols' in version `bold' +(Font) OMX/txex/m/n --> OMX/txex/bx/n on input line 121. +\symlargesymbolsA=\mathgroup9 +LaTeX Font Info: Overwriting symbol font `largesymbolsA' in version `bold' +(Font) U/txexa/m/n --> U/txexa/bx/n on input line 129. +LaTeX Font Info: Redeclaring math symbol \mathsterling on input line 164. +LaTeX Font Info: Redeclaring math symbol \hbar on input line 591. +LaTeX Info: Redefining \not on input line 1043. +LaTeX Info: Redefining \textsquare on input line 1063. +LaTeX Info: Redefining \openbox on input line 1064. +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/relsize/relsize.sty +Package: relsize 2013/03/29 ver 4.1 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/xkeyval/xkeyval.sty +Package: xkeyval 2022/06/16 v2.9 package option processing (HA) + +(/usr/local/texlive/2023/texmf-dist/tex/generic/xkeyval/xkeyval.tex +(/usr/local/texlive/2023/texmf-dist/tex/generic/xkeyval/xkvutils.tex +\XKV@toks=\toks40 +\XKV@tempa@toks=\toks41 +) +\XKV@depth=\count309 +File: xkeyval.tex 2014/12/03 v2.7a key=value parser (HA) +)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/hyphenat/hyphenat.sty +Package: hyphenat 2009/09/02 v2.3c hyphenation utilities +\langwohyphens=\language88 +LaTeX Info: Redefining \_ on input line 43. +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/bbm-macros/bbm.sty +Package: bbm 1999/03/15 V 1.2 provides fonts for set symbols - TH +LaTeX Font Info: Overwriting math alphabet `\mathbbm' in version `bold' +(Font) U/bbm/m/n --> U/bbm/bx/n on input line 33. +LaTeX Font Info: Overwriting math alphabet `\mathbbmss' in version `bold' +(Font) U/bbmss/m/n --> U/bbmss/bx/n on input line 35. +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/environ/environ.sty +Package: environ 2014/05/04 v0.3 A new way to define environments + +(/usr/local/texlive/2023/texmf-dist/tex/latex/trimspaces/trimspaces.sty +Package: trimspaces 2009/09/17 v1.1 Trim spaces around a token list +)) +\c@upm@subfigure@count=\count310 +\c@upm@fmt@mtabular@columnnumber=\count311 +\c@upm@format@section@sectionlevel=\count312 +\c@upm@fmt@savedcounter=\count313 +\c@@@upm@fmt@inlineenumeration=\count314 +\c@@upm@fmt@enumdescription@cnt@=\count315 +\upm@framed@minipage=\box113 +\upm@highlight@box@save=\box114 +\c@upmdefinition=\count316 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-versi +on.sty +Package: upmethodology-version 2022/02/10 + **** upmethodology-version is using French language **** +\upm@tmp@a=\count317 +) +\listendskip=\skip62 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-front +page.sty +Package: upmethodology-frontpage 2022/02/10 + **** upmethodology-frontpage is using French language **** +\upm@front@tmpa=\dimen256 +\upm@front@tmpb=\dimen257 + +*** define the extension value frontillustrationsize ****) +(/usr/local/texlive/2023/texmf-dist/tex/latex/upmethodology/upmethodology-backp +age.sty +Package: upmethodology-backpage 2022/02/10 + **** upmethodology-backpage is using French language ****) +(/usr/local/texlive/2023/texmf-dist/tex/latex/url/url.sty +\Urlmuskip=\muskip17 +Package: url 2013/09/16 ver 3.4 Verb mode for urls, etc. +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/hyperref/hyperref.sty +Package: hyperref 2023-05-16 v7.00y Hypertext links for LaTeX + +(/usr/local/texlive/2023/texmf-dist/tex/generic/ltxcmds/ltxcmds.sty +Package: ltxcmds 2020-05-10 v1.25 LaTeX kernel commands for general use (HO) +) +(/usr/local/texlive/2023/texmf-dist/tex/generic/pdftexcmds/pdftexcmds.sty +Package: pdftexcmds 2020-06-27 v0.33 Utility functions of pdfTeX for LuaTeX (HO +) + +(/usr/local/texlive/2023/texmf-dist/tex/generic/infwarerr/infwarerr.sty +Package: infwarerr 2019/12/03 v1.5 Providing info/warning/error messages (HO) +) +Package pdftexcmds Info: \pdf@primitive is available. +Package pdftexcmds Info: \pdf@ifprimitive is available. +Package pdftexcmds Info: \pdfdraftmode found. +) +(/usr/local/texlive/2023/texmf-dist/tex/generic/kvdefinekeys/kvdefinekeys.sty +Package: kvdefinekeys 2019-12-19 v1.6 Define keys (HO) +) +(/usr/local/texlive/2023/texmf-dist/tex/generic/pdfescape/pdfescape.sty +Package: pdfescape 2019/12/09 v1.15 Implements pdfTeX's escape features (HO) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/hycolor/hycolor.sty +Package: hycolor 2020-01-27 v1.10 Color options for hyperref/bookmark (HO) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/letltxmacro/letltxmacro.sty +Package: letltxmacro 2019/12/03 v1.6 Let assignment for LaTeX macros (HO) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/auxhook/auxhook.sty +Package: auxhook 2019-12-17 v1.6 Hooks for auxiliary files (HO) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/hyperref/nameref.sty +Package: nameref 2023-05-16 v2.51 Cross-referencing by name of section + +(/usr/local/texlive/2023/texmf-dist/tex/latex/refcount/refcount.sty +Package: refcount 2019/12/15 v3.6 Data extraction from label references (HO) +) +(/usr/local/texlive/2023/texmf-dist/tex/generic/gettitlestring/gettitlestring.s +ty +Package: gettitlestring 2019/12/15 v1.6 Cleanup title references (HO) + (/usr/local/texlive/2023/texmf-dist/tex/latex/kvoptions/kvoptions.sty +Package: kvoptions 2022-06-15 v3.15 Key value format for package options (HO) +)) +\c@section@level=\count318 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/etoolbox/etoolbox.sty +Package: etoolbox 2020/10/05 v2.5k e-TeX tools for LaTeX (JAW) +\etb@tempcnta=\count319 +) +\@linkdim=\dimen258 +\Hy@linkcounter=\count320 +\Hy@pagecounter=\count321 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/hyperref/pd1enc.def +File: pd1enc.def 2023-05-16 v7.00y Hyperref: PDFDocEncoding definition (HO) +Now handling font encoding PD1 ... +... no UTF-8 mapping file for font encoding PD1 +) +(/usr/local/texlive/2023/texmf-dist/tex/generic/intcalc/intcalc.sty +Package: intcalc 2019/12/15 v1.3 Expandable calculations with integers (HO) +) +\Hy@SavedSpaceFactor=\count322 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/hyperref/puenc.def +File: puenc.def 2023-05-16 v7.00y Hyperref: PDF Unicode definition (HO) +Now handling font encoding PU ... +... no UTF-8 mapping file for font encoding PU +) +Package hyperref Info: Option `breaklinks' set `true' on input line 4050. +Package hyperref Info: Option `pageanchor' set `true' on input line 4050. +Package hyperref Info: Option `bookmarks' set `false' on input line 4050. +Package hyperref Info: Option `hyperfigures' set `true' on input line 4050. +Package hyperref Info: Option `hyperindex' set `true' on input line 4050. +Package hyperref Info: Option `linktocpage' set `true' on input line 4050. +Package hyperref Info: Option `bookmarks' set `true' on input line 4050. +Package hyperref Info: Option `bookmarksopen' set `true' on input line 4050. +Package hyperref Info: Option `bookmarksnumbered' set `true' on input line 4050 +. +Package hyperref Info: Option `colorlinks' set `false' on input line 4050. +Package hyperref Info: Hyper figures ON on input line 4165. +Package hyperref Info: Link nesting OFF on input line 4172. +Package hyperref Info: Hyper index ON on input line 4175. +Package hyperref Info: Plain pages OFF on input line 4182. +Package hyperref Info: Backreferencing OFF on input line 4187. +Package hyperref Info: Implicit mode ON; LaTeX internals redefined. +Package hyperref Info: Bookmarks ON on input line 4434. +LaTeX Info: Redefining \href on input line 4683. +\c@Hy@tempcnt=\count323 +LaTeX Info: Redefining \url on input line 4772. +\XeTeXLinkMargin=\dimen259 + +(/usr/local/texlive/2023/texmf-dist/tex/generic/bitset/bitset.sty +Package: bitset 2019/12/09 v1.3 Handle bit-vector datatype (HO) + +(/usr/local/texlive/2023/texmf-dist/tex/generic/bigintcalc/bigintcalc.sty +Package: bigintcalc 2019/12/15 v1.5 Expandable calculations on big integers (HO +) +)) +\Fld@menulength=\count324 +\Field@Width=\dimen260 +\Fld@charsize=\dimen261 +Package hyperref Info: Hyper figures ON on input line 6049. +Package hyperref Info: Link nesting OFF on input line 6056. +Package hyperref Info: Hyper index ON on input line 6059. +Package hyperref Info: backreferencing OFF on input line 6066. +Package hyperref Info: Link coloring OFF on input line 6071. +Package hyperref Info: Link coloring with OCG OFF on input line 6076. +Package hyperref Info: PDF/A mode OFF on input line 6081. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/base/atbegshi-ltx.sty +Package: atbegshi-ltx 2021/01/10 v1.0c Emulation of the original atbegshi +package with kernel methods +) +\Hy@abspage=\count325 +\c@Item=\count326 +\c@Hfootnote=\count327 +) +Package hyperref Info: Driver: hpdftex. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/hyperref/hpdftex.def +File: hpdftex.def 2023-05-16 v7.00y Hyperref driver for pdfTeX + +(/usr/local/texlive/2023/texmf-dist/tex/latex/base/atveryend-ltx.sty +Package: atveryend-ltx 2020/08/19 v1.0a Emulation of the original atveryend pac +kage +with kernel methods +) +\Fld@listcount=\count328 +\c@bookmark@seq@number=\count329 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/rerunfilecheck/rerunfilecheck.sty +Package: rerunfilecheck 2022-07-10 v1.10 Rerun checks for auxiliary files (HO) + +(/usr/local/texlive/2023/texmf-dist/tex/generic/uniquecounter/uniquecounter.sty +Package: uniquecounter 2019/12/15 v1.4 Provide unlimited unique counter (HO) +) +Package uniquecounter Info: New unique counter `rerunfilecheck' on input line 2 +85. +) +\Hy@SectionHShift=\skip63 +) +\upm@smalllogo@height=\dimen262 +) (./spimbasephdthesis.sty +Package: spimbasephdthesis 2022/10/04 + +(/usr/local/texlive/2023/texmf-dist/tex/latex/lettrine/lettrine.sty +File: lettrine.sty 2023-04-18 v2.40 (Daniel Flipo) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/l3packages/xfp/xfp.sty +(/usr/local/texlive/2023/texmf-dist/tex/latex/l3kernel/expl3.sty +Package: expl3 2023-05-22 L3 programming layer (loader) + +(/usr/local/texlive/2023/texmf-dist/tex/latex/l3backend/l3backend-pdftex.def +File: l3backend-pdftex.def 2023-04-19 L3 backend support: PDF output (pdfTeX) +\l__color_backend_stack_int=\count330 +\l__pdf_internal_box=\box115 +)) +Package: xfp 2023-02-02 L3 Floating point unit +) +\c@DefaultLines=\count331 +\c@DefaultDepth=\count332 +\DefaultFindent=\dimen263 +\DefaultNindent=\dimen264 +\DefaultSlope=\dimen265 +\DiscardVskip=\dimen266 +\L@lbox=\box116 +\L@tbox=\box117 +\c@L@lines=\count333 +\c@L@depth=\count334 +\L@Pindent=\dimen267 +\L@Findent=\dimen268 +\L@Nindent=\dimen269 +\L@lraise=\dimen270 +\L@first=\dimen271 +\L@next=\dimen272 +\L@slope=\dimen273 +\L@height=\dimen274 +\L@novskip=\dimen275 +\L@target@ht=\dimen276 +\L@target@dp=\dimen277 +\L@target@tht=\dimen278 +\LettrineWidth=\dimen279 +\LettrineHeight=\dimen280 +\LettrineDepth=\dimen281 +Loading lettrine.cfg +(/usr/local/texlive/2023/texmf-dist/tex/latex/lettrine/lettrine.cfg) +\Llist@everypar=\toks42 +) +*** define the extension value backcovermessage ****) +**** including upm extension spimufcphdthesis (upmext-spimufcphdthesis.cfg) *** +* (./upmext-spimufcphdthesis.cfg *** define the extension value copyright **** +*** style extension spimufcphdthesis, Copyright {(c)} 2012--22 Prof. Dr. St\unh +box \voidb@x \bgroup \let \unhbox \voidb@x \setbox \@tempboxa \hbox {e\global \ +mathchardef \accent@spacefactor \spacefactor }\let \begingroup \let \typeout \p +rotect \begingroup \def \MessageBreak { +(Font) }\let \protect \immediate\write \m@ne {LaTeX Font Info: + on input line 5.}\endgroup \endgroup \relax \let \ignorespaces \relax \accent +19 e\egroup \spacefactor \accent@spacefactor phane GALLAND. **** +*** define the extension value trademarks **** +(/usr/local/texlive/2023/texmf-dist/tex/latex/psnfss/helvet.sty +Package: helvet 2020/03/25 PSNFSS-v9.3 (WaS) +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/base/textcomp.sty +Package: textcomp 2020/02/02 v2.0n Standard LaTeX package +) +*** define the extension value watermarksize **** +*** define the extension value doctoralschoolnumber **** +*** define the extension value doctoralschoolname **** +*** define the extension value universityname **** +*** define the extension value speciality **** +*** define the extension value defensedate **** +*** define the extension value ubfcaddress **** +*** define the extension value jurytabwidth **** +*** define the extension value defensemessagestyle **** +*** define the extension value jurystyle **** +*** define the extension value defensemessage **** +*** define the extension value doctoralschoollogo **** +*** define the extension value minorlanguagethesistitle ****)) +(/usr/local/texlive/2023/texmf-dist/tex/latex/base/inputenc.sty +Package: inputenc 2021/02/14 v1.3d Input encoding file +\inpenc@prehook=\toks43 +\inpenc@posthook=\toks44 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/algorithms/algorithm.sty +Package: algorithm 2009/08/24 v0.1 Document Style `algorithm' - floating enviro +nment + +(/usr/local/texlive/2023/texmf-dist/tex/latex/float/float.sty +Package: float 2001/11/08 v1.3d Float enhancements (AL) +\c@float@type=\count335 +\float@exts=\toks45 +\float@box=\box118 +\@float@everytoks=\toks46 +\@floatcapt=\box119 +) +\@float@every@algorithm=\toks47 +\c@algorithm=\count336 +) +(/usr/local/texlive/2023/texmf-dist/tex/latex/algorithmicx/algpseudocode.sty +Package: algpseudocode + +(/usr/local/texlive/2023/texmf-dist/tex/latex/algorithmicx/algorithmicx.sty +Package: algorithmicx 2005/04/27 v1.2 Algorithmicx + +Document Style algorithmicx 1.2 - a greatly improved `algorithmic' style +\c@ALG@line=\count337 +\c@ALG@rem=\count338 +\c@ALG@nested=\count339 +\ALG@tlm=\skip64 +\ALG@thistlm=\skip65 +\c@ALG@Lnr=\count340 +\c@ALG@blocknr=\count341 +\c@ALG@storecount=\count342 +\c@ALG@tmpcounter=\count343 +\ALG@tmplength=\skip66 +) +Document Style - pseudocode environments for use with the `algorithmicx' style +) *** define the extension value defensedate **** +*** define the extension value defenselocation **** +LaTeX Font Info: Trying to load font information for OT1+phv on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/psnfss/ot1phv.fd +File: ot1phv.fd 2020/03/25 scalable font definitions for OT1/phv. +) (./main.aux + +LaTeX Warning: Label `fig:figCycle' multiply defined. + +) +\openout1 = `main.aux'. + +LaTeX Font Info: Checking defaults for OML/txmi/m/it on input line 224. +LaTeX Font Info: Trying to load font information for OML+txmi on input line +224. + (/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/omltxmi.fd +File: omltxmi.fd 2000/12/15 v3.1 +) +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for OMS/txsy/m/n on input line 224. +LaTeX Font Info: Trying to load font information for OMS+txsy on input line +224. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/omstxsy.fd +File: omstxsy.fd 2000/12/15 v3.1 +) +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for OT1/cmr/m/n on input line 224. +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for T1/cmr/m/n on input line 224. +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for TS1/cmr/m/n on input line 224. +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for OMX/txex/m/n on input line 224. +LaTeX Font Info: Trying to load font information for OMX+txex on input line +224. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/omxtxex.fd +File: omxtxex.fd 2000/12/15 v3.1 +) +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for U/txexa/m/n on input line 224. +LaTeX Font Info: Trying to load font information for U+txexa on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/utxexa.fd +File: utxexa.fd 2000/12/15 v3.1 +) +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for PD1/pdf/m/n on input line 224. +LaTeX Font Info: ... okay on input line 224. +LaTeX Font Info: Checking defaults for PU/pdf/m/n on input line 224. +LaTeX Font Info: ... okay on input line 224. + +(/usr/local/texlive/2023/texmf-dist/tex/context/base/mkii/supp-pdf.mkii +[Loading MPS to PDF converter (version 2006.09.02).] +\scratchcounter=\count344 +\scratchdimen=\dimen282 +\scratchbox=\box120 +\nofMPsegments=\count345 +\nofMParguments=\count346 +\everyMPshowfont=\toks48 +\MPscratchCnt=\count347 +\MPscratchDim=\dimen283 +\MPnumerator=\count348 +\makeMPintoPDFobject=\count349 +\everyMPtoPDFconversion=\toks49 +) (/usr/local/texlive/2023/texmf-dist/tex/latex/epstopdf-pkg/epstopdf-base.sty +Package: epstopdf-base 2020-01-24 v2.11 Base part for package epstopdf +Package epstopdf-base Info: Redefining graphics rule for `.eps' on input line 4 +85. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/latexconfig/epstopdf-sys.cfg +File: epstopdf-sys.cfg 2010/07/13 v1.3 Configuration of (r)epstopdf for TeX Liv +e +)) +LaTeX Info: Redefining \degres on input line 224. +LaTeX Info: Redefining \up on input line 224. + + +Package french.ldf Warning: OT1 encoding should not be used for French. +(french.ldf) Add \usepackage[T1]{fontenc} to the preamble +(french.ldf) of your document; reported on input line 224. + +Package caption Info: Begin \AtBeginDocument code. +Package caption Info: float package is loaded. +Package caption Info: hyperref package is loaded. +Package caption Info: picinpar package is loaded. +Package caption Info: End \AtBeginDocument code. + +*** Overriding the 'enumerate' environment. Pass option 'standardlists' for avo +iding this override. +*** Overriding the 'description' environment. Pass option 'standardlists' for a +voiding this override. ************ USE CUSTOM FRONT COVER + +File: ubfc.pdf Graphic file (type pdf) + +Package pdftex.def Info: ubfc.pdf used on input line 224. +(pdftex.def) Requested size: 91.06479pt x 56.9055pt. + +File: ufc.pdf Graphic file (type pdf) + +Package pdftex.def Info: ufc.pdf used on input line 224. +(pdftex.def) Requested size: 170.7282pt x 42.67912pt. +LaTeX Font Info: Trying to load font information for TS1+phv on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/psnfss/ts1phv.fd +File: ts1phv.fd 2020/03/25 scalable font definitions for TS1/phv. +) +LaTeX Font Info: Trying to load font information for OT1+txr on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/ot1txr.fd +File: ot1txr.fd 2000/12/15 v3.1 +) +LaTeX Font Info: Trying to load font information for U+txmia on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/utxmia.fd +File: utxmia.fd 2000/12/15 v3.1 +) +LaTeX Font Info: Trying to load font information for U+txsya on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/utxsya.fd +File: utxsya.fd 2000/12/15 v3.1 +) +LaTeX Font Info: Trying to load font information for U+txsyb on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/utxsyb.fd +File: utxsyb.fd 2000/12/15 v3.1 +) +LaTeX Font Info: Trying to load font information for U+txsyc on input line 2 +24. + +(/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/utxsyc.fd +File: utxsyc.fd 2000/12/15 v3.1 +) [1 + + +{/usr/local/texlive/2023/texmf-var/fonts/map/pdftex/updmap/pdftex.map}{/usr/loc +al/texlive/2023/texmf-dist/fonts/enc/dvips/base/8r.enc} <./ubfc.pdf> <./ufc.pdf + + +pdfTeX warning: /Library/TeX/texbin/pdflatex (file ./ufc.pdf): PDF inclusion: m +ultiple pdfs with page group included in a single page +>] +Package hyperref Info: Link coloring OFF on input line 224. + (./main.out) (./main.out) +\@outlinefile=\write3 +\openout3 = `main.out'. + + [2 + +] [3] [4] (./main.toc +LaTeX Font Info: Font shape `OT1/phv/m/it' in size <10.95> not available +(Font) Font shape `OT1/phv/m/sl' tried instead on input line 20. + +Underfull \vbox (badness 10000) has occurred while \output is active [] + + [5 + +]) +\tf@toc=\write4 +\openout4 = `main.toc'. + + +[6] [1 + + +] [2] +Chapitre 1. +Package lettrine.sty Info: Targeted height = 23.28528pt +(lettrine.sty) (for loversize=0, accent excluded), +(lettrine.sty) Lettrine height = 23.75279pt (\uppercase {C}); +(lettrine.sty) reported on input line 280. + +Underfull \hbox (badness 10000) in paragraph at lines 280--280 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 280--280 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 280--280 + + [] + + +Underfull \vbox (badness 10000) has occurred while \output is active [] + + [3 + +] +Overfull \hbox (6.79999pt too wide) in paragraph at lines 280--280 +[][][][] + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 280--280 + + [] + +[4] +Underfull \hbox (badness 10000) in paragraph at lines 301--302 + + [] + +[5] [6] +Chapitre 2. + +Underfull \hbox (badness 10000) in paragraph at lines 311--312 + + [] + +[7 + +] +Underfull \hbox (badness 10000) in paragraph at lines 316--317 + + [] + +<./Figures/EIAH.png, id=222, 546.54187pt x 468.24937pt> +File: ./Figures/EIAH.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/EIAH.png used on input line 319. +(pdftex.def) Requested size: 273.27026pt x 234.1241pt. + +Underfull \hbox (badness 1270) in paragraph at lines 325--326 +\OT1/phv/m/n/10.95 por-te-ment li[]e []a l'exp[]erience ou []a la pra-tique. L' +ap-pren-tis-sage im-plique r[]eflexion, + [] + +<./Figures/TLearning.png, id=228, 1004.25188pt x 275.52937pt> +File: ./Figures/TLearning.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/TLearning.png used on input line 327. +(pdftex.def) Requested size: 401.69363pt x 110.20978pt. + +Underfull \hbox (badness 10000) in paragraph at lines 336--337 + + [] + +[8 <./Figures/EIAH.png>] +Underfull \hbox (badness 3219) in paragraph at lines 338--339 +[]\OT1/phv/m/n/10.95 L'autre strat[]egie glo-bale est l'ap-pren-tis-sage on-lin +e, qui est bas[]e sur les res- + [] + + +Underfull \hbox (badness 5592) in paragraph at lines 338--339 +\OT1/phv/m/n/10.95 sources et l'in-for-ma-tion dis-po-nible sur le web. Cette s +trat[]egie in-cite ac-ti-ve- + [] + + +Underfull \hbox (badness 3815) in paragraph at lines 338--339 +\OT1/phv/m/n/10.95 ment les ap-pre-nants []a acqu[]erir de nou-velles connais-s +ances ind[]ependamment + [] + + +Underfull \hbox (badness 1924) in paragraph at lines 338--339 +[][][][][][] \OT1/phv/m/n/10.95 gr[]ace aux vastes res-sources + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 338--339 + + [] + + +Underfull \hbox (badness 3128) in paragraph at lines 365--366 +\OT1/phv/m/n/10.95 saie d'imi-ter le com-por-te-ment hu-main pour r[]esoudre de + nou-veaux probl[]emes, + [] + + +Underfull \hbox (badness 2435) in paragraph at lines 365--366 +\OT1/phv/m/n/10.95 en se sou-ve-nant des exp[]eriences pass[]ees. Le prin-cipe +de base est que les + [] + + +Underfull \hbox (badness 2538) in paragraph at lines 365--366 +\OT1/phv/m/n/10.95 probl[]emes si-mi-laires ont des so-lu-tions si-mi-laires. S +i un nou-veau probl[]eme ar- + [] + + +Underfull \hbox (badness 2818) in paragraph at lines 365--366 +\OT1/phv/m/n/10.95 rive et que le mod[]ele connait d[]ej[]a la so-lu-tion d'un +probl[]eme si-mi-laire, alors + [] + + +Underfull \hbox (badness 1675) in paragraph at lines 365--366 +\OT1/phv/m/n/10.95 la so-lu-tion au nou-veau probl[]eme peut []etre inf[]er[]ee + des so-lu-tions d[]ej[]a trouv[]ees + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 365--366 +[][][][][][] + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 365--366 + + [] + +[9 <./Figures/TLearning.png>] +Underfull \hbox (badness 10000) in paragraph at lines 367--368 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 369--370 + + [] + +<./Figures/cycle.png, id=245, 6.97093pt x 9.97743pt> +File: ./Figures/cycle.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/cycle.png used on input line 371. +(pdftex.def) Requested size: 209.12704pt x 299.3216pt. +[10 <./Figures/cycle.png>] +Underfull \hbox (badness 10000) in paragraph at lines 383--384 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 394--395 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 396--398 + + [] + +[11] +<./Figures/Reuse.png, id=275, 13.65732pt x 3.78421pt> +File: ./Figures/Reuse.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/Reuse.png used on input line 399. +(pdftex.def) Requested size: 273.14514pt x 75.6836pt. + +Underfull \hbox (badness 10000) in paragraph at lines 410--411 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 412--413 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 414--415 + + [] + +[12 <./Figures/Reuse.png>] +Underfull \hbox (badness 10000) in paragraph at lines 429--430 +[]\OT1/phv/m/n/10.95 Conteneur du vo-ca-bu-laire : il contient l'in-for-ma-tion + de la + [] + + +Underfull \hbox (badness 7613) in paragraph at lines 429--430 +\OT1/phv/m/n/10.95 repr[]esentation.[][][][][][] Les []el[]ements + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 430--431 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 433--434 + + [] + +[13] +<./Figures/CycleCBR.png, id=292, 147.1899pt x 83.8332pt> +File: ./Figures/CycleCBR.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/CycleCBR.png used on input line 435. +(pdftex.def) Requested size: 367.97382pt x 209.58244pt. + +Underfull \hbox (badness 10000) in paragraph at lines 446--447 + + [] + +[14 <./Figures/CycleCBR.png>] +Underfull \hbox (badness 10000) in paragraph at lines 454--455 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 474--475 + + [] + +[15] +Underfull \hbox (badness 10000) in paragraph at lines 480--481 + + [] + + +Underfull \hbox (badness 1668) in paragraph at lines 507--508 +[][][][][][][]\OT1/phv/m/n/10.95 Souvent, pour d[]ecider de la + [] + +[16] + +LaTeX Warning: Command \textperiodcentered invalid in math mode on input line 5 +42. + + +LaTeX Warning: Command \textperiodcentered invalid in math mode on input line 5 +42. + + +LaTeX Warning: Command \textperiodcentered invalid in math mode on input line 5 +42. + + +LaTeX Warning: Command \textperiodcentered invalid in math mode on input line 5 +42. + + +LaTeX Warning: Command \textperiodcentered invalid in math mode on input line 5 +42. + + +LaTeX Warning: Command \textperiodcentered invalid in math mode on input line 5 +42. + +Missing character: There is no in font txr! +Missing character: There is no in font txr! +Missing character: There is no in font txr! + +Underfull \hbox (badness 2626) in paragraph at lines 549--550 +[][][][][][][]\OT1/phv/m/n/10.95 L'algorithme des k moyennes est it[]er[]e []a +tra-vers les condi-tions + [] + +[17] + +LaTeX Font Warning: Font shape `OT1/phv/m/scit' undefined +(Font) using `OT1/phv/m/it' instead on input line 561. + +LaTeX Font Info: Font shape `OT1/phv/m/it' in size <12> not available +(Font) Font shape `OT1/phv/m/sl' tried instead on input line 561. +[18] + +LaTeX Font Warning: Font shape `OT1/phv/m/scit' undefined +(Font) using `OT1/phv/m/it' instead on input line 615. + +[19] + +LaTeX Font Warning: Font shape `OT1/phv/m/scit' undefined +(Font) using `OT1/phv/m/it' instead on input line 619. + + +Underfull \hbox (badness 1028) in paragraph at lines 621--622 +[]\OT1/phv/m/n/10.95 Comme in-diqu[]e dans [[]], l'al-go-rithme d'[]echantillon +nage de Thomp-son est + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 621--622 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 623--624 + + [] + +<./Figures/beta-distribution.png, id=377, 621.11293pt x 480.07928pt> +File: ./Figures/beta-distribution.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/beta-distribution.png used on input line 62 +5. +(pdftex.def) Requested size: 310.55571pt x 240.03905pt. + +Underfull \hbox (badness 10000) in paragraph at lines 643--644 + + [] + +[20] [21 <./Figures/beta-distribution.png>] [22 + +] [23] [24] +Chapitre 3. + +Underfull \hbox (badness 10000) in paragraph at lines 654--655 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 656--657 + + [] + +[25 + +] +Underfull \hbox (badness 2376) in paragraph at lines 658--659 +[]\OT1/phv/m/n/10.95 Les tech-niques d'IA peuvent aussi ai-der []a prendre des +d[]ecisions strat[]egiques + [] + + +Underfull \hbox (badness 6526) in paragraph at lines 658--659 +\OT1/phv/m/n/10.95 qui visent des ob-jec-tifs []a long []ech[]eance comme le mo +ntre le tra-vail de + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 662--663 + + [] + +[26] +<./Figures/architecture.png, id=420, 776.9025pt x 454.69875pt> +File: ./Figures/architecture.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/architecture.png used on input line 666. +(pdftex.def) Requested size: 388.45029pt x 227.34882pt. + +Underfull \hbox (badness 8170) in paragraph at lines 667--667 +[]\OT1/phv/m/sc/10.95 Figure 3.1 \OT1/phv/m/n/10.95 { |Ar-chi-tec-ture du syst[ +]eme de re-com-man-da-tion pro-pos[]e dans + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 671--672 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 673--674 + + [] + + +Underfull \vbox (badness 5893) has occurred while \output is active [] + + [27 <./Figures/architecture.png>] +Underfull \hbox (badness 10000) in paragraph at lines 675--676 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 677--678 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 679--680 + + [] + + +Underfull \vbox (badness 10000) has occurred while \output is active [] + + [28] +Underfull \hbox (badness 10000) in paragraph at lines 681--682 + + [] + +<./Figures/ELearningLevels.png, id=449, 602.25pt x 612.78937pt> +File: ./Figures/ELearningLevels.png Graphic file (type png) + +Package pdftex.def Info: ./Figures/ELearningLevels.png used on input line 683. + +(pdftex.def) Requested size: 301.12425pt x 306.39394pt. + +Underfull \hbox (badness 10000) in paragraph at lines 692--693 + + [] + + +Underfull \hbox (badness 10000) in paragraph at lines 694--695 + + [] + + +Overfull \hbox (2.56369pt too wide) in paragraph at lines 703--703 +[]|\OT1/phv/m/n/9 [[]]| + [] + + +Overfull \hbox (0.5975pt too wide) in paragraph at lines 698--714 +[][] + [] + +(./main.bbl [29] [30 <./Figures/ELearningLevels.png>] +Underfull \hbox (badness 1394) in paragraph at lines 34--40 +[]\OT1/phv/m/n/10.95 Henriet, J., Chris-tophe, L., and Laurent, P. (2017). Ar- +ti-fi-cial + [] + +[31 + +] +Underfull \hbox (badness 1939) in paragraph at lines 95--99 +[]\OT1/phv/m/n/10.95 Muangprathub, J., Boon-jing, V., and Cham-nong-thai, K. + [] + + +Underfull \vbox (badness 2150) has occurred while \output is active [] + + [32]) +[33] [34 + +] +LaTeX Font Info: Trying to load font information for OT1+txtt on input line +724. + (/usr/local/texlive/2023/texmf-dist/tex/latex/txfonts/ot1txtt.fd +File: ot1txtt.fd 2000/12/15 v3.1 +) +[35] + +File: spim.pdf Graphic file (type pdf) + +Package pdftex.def Info: spim.pdf used on input line 724. +(pdftex.def) Requested size: 170.0pt x 30.07556pt. +File: ubfc.pdf Graphic file (type pdf) + +Package pdftex.def Info: ubfc.pdf used on input line 724. +(pdftex.def) Requested size: 50.08586pt x 31.2982pt. + [36 + + <./spim.pdf>] (./main.aux) + +LaTeX Warning: There were multiply-defined labels. + +Package rerunfilecheck Info: File `main.out' has not changed. +(rerunfilecheck) Checksum: F7B40CA71D6DF7DEBE42C418DB3FFD5B;6176. + ) +Here is how much of TeX's memory you used: + 18747 strings out of 476038 + 317239 string characters out of 5790170 + 1867785 words of memory out of 5000000 + 38645 multiletter control sequences out of 15000+600000 + 580741 words of font info for 148 fonts, out of 8000000 for 9000 + 1141 hyphenation exceptions out of 8191 + 126i,16n,133p,2134b,752s stack positions out of 10000i,1000n,20000p,200000b,200000s + +pdfTeX warning (dest): name{cite.doi\\penalty\040\\@M\040\040:10.1177/175433711 +6651013} has been referenced but does not exist, replaced by a fixed one + + +Output written on main.pdf (42 pages, 2085290 bytes). +PDF statistics: + 606 PDF objects out of 1000 (max. 8388607) + 499 compressed objects within 5 object streams + 130 named destinations out of 1000 (max. 500000) + 264 words of extra memory for PDF output out of 10000 (max. 10000000) + diff --git a/main.out b/main.out new file mode 100644 index 0000000..ff0437d --- /dev/null +++ b/main.out @@ -0,0 +1,26 @@ +\BOOKMARK [-1][]{part.1}{\376\377\000I\000\040\000C\000o\000n\000t\000e\000x\000t\000e\000\040\000e\000t\000\040\000P\000r\000o\000b\000l\000\351\000m\000a\000t\000i\000q\000u\000e\000s}{}% 1 +\BOOKMARK [0][]{chapter.1}{\376\377\0001\000\040\000I\000n\000t\000r\000o\000d\000u\000c\000t\000i\000o\000n}{part.1}% 2 +\BOOKMARK [1][]{section.1.1}{\376\377\0001\000.\0001\000\040\000C\000o\000n\000t\000r\000i\000b\000u\000t\000i\000o\000n\000s\000\040\000P\000r\000i\000n\000c\000i\000p\000a\000l\000e\000s}{chapter.1}% 3 +\BOOKMARK [1][]{section.1.2}{\376\377\0001\000.\0002\000\040\000P\000l\000a\000n\000\040\000d\000e\000\040\000l\000a\000\040\000t\000h\000\350\000s\000e}{chapter.1}% 4 +\BOOKMARK [0][]{chapter.2}{\376\377\0002\000\040\000C\000o\000n\000t\000e\000x\000t\000e}{part.1}% 5 +\BOOKMARK [1][]{section.2.1}{\376\377\0002\000.\0001\000\040\000N\000o\000t\000i\000o\000n\000s\000\040\000e\000t\000\040\000A\000l\000g\000o\000r\000i\000t\000h\000m\000e\000s}{chapter.2}% 6 +\BOOKMARK [2][]{subsection.2.1.1}{\376\377\0002\000.\0001\000.\0001\000\040\000E\000n\000v\000i\000r\000o\000n\000n\000e\000m\000e\000n\000t\000s\000\040\000I\000n\000f\000o\000r\000m\000a\000t\000i\000q\000u\000e\000s\000\040\000p\000o\000u\000r\000\040\000l\000'\000A\000p\000p\000r\000e\000n\000t\000i\000s\000s\000a\000g\000e\000\040\000H\000u\000m\000a\000i\000n\000\040\000\050\000E\000I\000A\000H\000\051}{section.2.1}% 7 +\BOOKMARK [2][]{subsection.2.1.2}{\376\377\0002\000.\0001\000.\0002\000\040\000S\000t\000r\000a\000t\000\351\000g\000i\000e\000s\000\040\000d\000'\000A\000p\000p\000r\000e\000n\000t\000i\000s\000s\000a\000g\000e}{section.2.1}% 8 +\BOOKMARK [2][]{subsection.2.1.3}{\376\377\0002\000.\0001\000.\0003\000\040\000R\000a\000i\000s\000o\000n\000n\000e\000m\000e\000n\000t\000\040\000\340\000\040\000p\000a\000r\000t\000i\000r\000\040\000d\000e\000\040\000c\000a\000s\000\040\000\050\000R\000\340\000P\000C\000\051}{section.2.1}% 9 +\BOOKMARK [3][]{subsubsection.2.1.3.1}{\376\377\0002\000.\0001\000.\0003\000.\0001\000\040\000R\000e\000c\000h\000e\000r\000c\000h\000e\000r}{subsection.2.1.3}% 10 +\BOOKMARK [3][]{subsubsection.2.1.3.2}{\376\377\0002\000.\0001\000.\0003\000.\0002\000\040\000A\000d\000a\000p\000t\000e\000r\000\040\000\050\000R\000\351\000u\000t\000i\000l\000i\000s\000e\000r\000\051}{subsection.2.1.3}% 11 +\BOOKMARK [3][]{subsubsection.2.1.3.3}{\376\377\0002\000.\0001\000.\0003\000.\0003\000\040\000R\000\351\000v\000i\000s\000e\000r\000\040\000e\000t\000\040\000R\000\351\000p\000a\000r\000e\000r}{subsection.2.1.3}% 12 +\BOOKMARK [3][]{subsubsection.2.1.3.4}{\376\377\0002\000.\0001\000.\0003\000.\0004\000\040\000S\000t\000o\000c\000k\000e\000r\000\040\000\050\000R\000e\000t\000e\000n\000i\000r\000\051}{subsection.2.1.3}% 13 +\BOOKMARK [3][]{subsubsection.2.1.3.5}{\376\377\0002\000.\0001\000.\0003\000.\0005\000\040\000C\000o\000n\000t\000e\000n\000e\000u\000r\000s\000\040\000d\000e\000\040\000C\000o\000n\000n\000a\000i\000s\000s\000a\000n\000c\000e}{subsection.2.1.3}% 14 +\BOOKMARK [2][]{subsection.2.1.4}{\376\377\0002\000.\0001\000.\0004\000\040\000S\000y\000s\000t\000\350\000m\000e\000s\000\040\000M\000u\000l\000t\000i\000-\000A\000g\000e\000n\000t\000s}{section.2.1}% 15 +\BOOKMARK [2][]{subsection.2.1.5}{\376\377\0002\000.\0001\000.\0005\000\040\000P\000e\000n\000s\000\351\000e\000\040\000B\000a\000y\000e\000s\000i\000e\000n\000n\000e}{section.2.1}% 16 +\BOOKMARK [2][]{subsection.2.1.6}{\376\377\0002\000.\0001\000.\0006\000\040\000M\000\351\000t\000h\000o\000d\000e\000\040\000d\000e\000s\000\040\000k\000\040\000p\000l\000u\000s\000\040\000p\000r\000o\000c\000h\000e\000s\000\040\000v\000o\000i\000s\000i\000n\000s\000\040\000\050\000K\000-\000N\000e\000a\000r\000e\000s\000t\000\040\000N\000e\000i\000g\000h\000b\000o\000r\000h\000o\000o\000d\000\040\000-\000\040\000K\000N\000N\000\051}{section.2.1}% 17 +\BOOKMARK [2][]{subsection.2.1.7}{\376\377\0002\000.\0001\000.\0007\000\040\000K\000-\000M\000o\000y\000e\000n\000n\000e\000s}{section.2.1}% 18 +\BOOKMARK [2][]{subsection.2.1.8}{\376\377\0002\000.\0001\000.\0008\000\040\000M\000o\000d\000\350\000l\000e\000\040\000d\000e\000\040\000M\000\351\000l\000a\000n\000g\000e\000\040\000G\000a\000u\000s\000s\000i\000e\000n\000\040\000G\000M\000M\000\040\000\050\000G\000a\000u\000s\000s\000i\000a\000n\000\040\000M\000i\000x\000t\000u\000r\000e\000\040\000M\000o\000d\000e\000l\000\051}{section.2.1}% 19 +\BOOKMARK [2][]{subsection.2.1.9}{\376\377\0002\000.\0001\000.\0009\000\040\000F\000u\000z\000z\000y\000-\000C}{section.2.1}% 20 +\BOOKMARK [2][]{subsection.2.1.10}{\376\377\0002\000.\0001\000.\0001\0000\000\040\000B\000a\000n\000d\000i\000t\000\040\000M\000a\000n\000c\000h\000o\000t\000\040\000M\000A\000B\000\040\000\050\000M\000u\000l\000t\000i\000-\000A\000r\000m\000e\000d\000\040\000B\000a\000n\000d\000i\000t\000s\000\051}{section.2.1}% 21 +\BOOKMARK [2][]{subsection.2.1.11}{\376\377\0002\000.\0001\000.\0001\0001\000\040\000\311\000c\000h\000a\000n\000t\000i\000l\000l\000o\000n\000n\000a\000g\000e\000\040\000d\000e\000\040\000T\000h\000o\000m\000p\000s\000o\000n\000\040\000T\000S\000\040\000\050\000T\000h\000o\000m\000p\000s\000o\000n\000\040\000S\000a\000m\000p\000l\000i\000n\000g\000\051}{section.2.1}% 22 +\BOOKMARK [-1][]{part.2}{\376\377\000I\000I\000\040\000\311\000t\000a\000t\000\040\000d\000e\000\040\000l\000'\000a\000r\000t}{}% 23 +\BOOKMARK [0][]{chapter.3}{\376\377\0003\000\040\000E\000n\000v\000i\000r\000o\000n\000n\000e\000m\000e\000n\000t\000s\000\040\000I\000n\000f\000o\000r\000m\000a\000t\000i\000q\000u\000e\000s\000\040\000d\000'\000A\000p\000p\000r\000e\000n\000t\000i\000s\000s\000a\000g\000e\000\040\000H\000u\000m\000a\000i\000n}{part.2}% 24 +\BOOKMARK [1][]{section.3.1}{\376\377\0003\000.\0001\000\040\000L\000'\000I\000n\000t\000e\000l\000l\000i\000g\000e\000n\000c\000e\000\040\000A\000r\000t\000i\000f\000i\000c\000i\000e\000l\000l\000e}{chapter.3}% 25 +\BOOKMARK [1][]{section.3.2}{\376\377\0003\000.\0002\000\040\000S\000y\000s\000t\000\350\000m\000e\000s\000\040\000d\000e\000\040\000R\000e\000c\000o\000m\000m\000a\000n\000d\000a\000t\000i\000o\000n\000\040\000d\000a\000n\000s\000\040\000l\000e\000s\000\040\000E\000I\000A\000H}{chapter.3}% 26 diff --git a/main.pdf b/main.pdf index 77d8fff05356b13cc5b98d43d595b811ae5552b6..793bd080ec028851d899b52fd18921c9a604c6f4 100644 GIT binary patch delta 70241 zcmV)PK()V^qJOHRe~?;#*{bY#?jvs!!V-=SELKiZwjT(rZK_}w19IYjpPmbLVUbbF z4{@n#SuxXJ_w?~~4>XxEd z_~PYYL6VEA^9KIn62lB(rMD6YV=S0oti2zBqY>_y!5Nr~I8U;Fn#t#R=jAJ<5`y6y zLz4VsLYYL2i9~XYurx_NP2P3$tMo(uIw7zmc|yS&?HAr-Cf!m-kft=rYHR&bKh3Dp z-HZ5n*UM9myLb0zslnYm2!cXxMHxW^2Zd4OIwc_R!tJCQ5wO*2P(eGDe5ajw?P0B- zTODWT*Mrfhf78Ez%{}15?ci40v5VSyx--UY-j4ez7NjS?s04O4>Q)Ls_QAUr3_dSi$|3O7BnjWI=J zPw1=>rlxb1K@iU8xi-fA(6L|_)B5D-+XYB%4T8uBL2YV(XEr*9m?YyAX`>Hej8PVb zX%S#1)x7W)O{>YKT7i4JqxqL?W!EFqH^KvRw|WnO$VnHA-vSo`h6RN!v)xbpwa zUku7m{%E%Dg-I}?lpNZ`7)!F+T2ao52vThP=gn-gsy4;sb2P5T=ZH&i3m1s$ zV9~|y6hU-<3U+)<8CYg9MRzgYSPTat7-H5Vb%f=3<>3wZ36n<7UTnQ?pbEI-6OELC zvu)Ew((WxhNq6Ah`|f$&7A^)fMdVJj4O=K0H{3d;3y1Ad@7;6co_9g>Ru4yvOQ%Oj z$QvGEMs-s-@D1z$dZc=52|!WW&E4~^4UhqIs<``qRSTX$T(cV7*4DnW9^Uy8=!|NN z1Tb|5%q3z>&Zgt#2I5wgAuYn^rdTXP^7)_<4({dBrkWIosU}^b#g(F%#}s)iL(2Qq zQqV5oBu7vL#Jhnrh8l=SMY-|m=XPsoD4N^dmZ$;}!EOyh?m+{stzb&ZVgU#q%x0o{ znU$M=Vm_aJvb7eBz!V~83mzF3oMd+cNo{8Dp)*5a&gVrfrHtiY=U@o7l2#UD#6{qf z{ImBTMT*&j6R`TA0TQT#thQFH!?BmE*ua!&O4r=hA`K^Xasm>^oGe!~xnW&BZXhAR z=b@^lLDH~O4>1lsuUJDEQ%)2pS2)cy?^fl1FM-Q~N}__wVhO4D*D8EXVG@eH9^wL- zoY}Ug4jxZ1`NLawl*2LoUW@JurXzb> zFcj0194Ve~qz6;PQ3t{nQXSxZ`{2J&Dp?(gNWyJXU{H1L!;Oux)_-&Qxl-rYZ5rb-g765&btpdnZlAE)phc>+qke*n|(0m$*A)t z{KnKUO~Mg0N#Lpsk#H*iIMk;&lI+LmgFBNZ-_vJcRsI-#?&`q5r;rdRd-ESO)?&hu z#v&owP8?4F2m{w4o5dbr|0lwK0X>=38#&+5~R#2%xf}v!S@JktO z$O56{V%x(j%YV`|s%=!8z4IU)I`0_b1B?+QFW{s{A_MUb*Jf&O@mQT4&`YqznJJsR zaJCGsw$hv`7S4I7@Lk2jS6_S1mB)&ga^&~G(&wusd#L$zC+LJhTQAu8C16TvH9bW2 z2JjC!laoOs6O)NdP6D@3li5r!1hp&>(38P16q6uL7MCK09x0PwOFzf8~Vv1 z$i{*EVe{k1V{`_8K>y5hsKLB6FD3{CJD5)6e4hH0ass))vS=U1r!SC$bACe<)^95{?m)t&Pwp^(`TW!;6Q7KIGWSX9lkf56s(u=7ukL*INfyjJ_+*q< z5Q_Qy#V5VcZs>oLfB9q)49<>g==;yT={(Jv9AN9+U&h!=AcN3FKkj?!ZN0}CxJL*i zY-68Et@kGSve45*fKcr{ceq3*dMQ#4Qle26<2q1 z&z=jtY49uEtY`x^zHdlw_MJDW$D7sr6nfgJPd)BW0v~@*p)@lisJbuo*7CGR} zJ<+uks4mNE=ySfl+IeZ7x37Y?qj;pe4WsX==~=uJe@lr}ZH<>!7B)q!N34Gcl)lDp z4Py)Ac3kUrSaprK6}WO0HE9l$K{NYbYmbys(4)Re$>IU`+-&n(s5Ojf8E<^*{2m1V z6l{N1rj=Q3CMv5ALzFy+-9)({D`_dYS=o5Ie(`&-`l6!Vc60TO*;Obj_e6_0 zur^SVmfQ0-ap2mC;MRLQ4`}`!lGhL$r@MbQL3EE@5N)2m#=+rXZGc@hJU5=>onTb; zB$%$s)?$*+;%CI5J=|Gqk=Q=lMi#HW8H+jSHS8!xA!&LbKeZTKUt*45!qi*YGeq@o znke7b(oMkm5zV+irMrl0V<_q8E3;)4|Ydumc zo7ZyOy)?{Uy-jb^BKy7Ob-Hi#dRw$`R_Yhoszufl%zlAlw|D=II4A>%j-5e zwRRY?Isw^rps~^8x{UG%Wt1&tJfIh*c5>=ZeE*BH)06Rd0)w~tiLyuJ50sZVD{@}s zqR3^Dt0LD$Zi?I%d0|f%_H7mT2v5gKG+ZLLZP5G#jf8cH$zykkM&PlgrEG&n91;y?FHVG3iP9xC!l;A` zw-HS9$;CktNyLC!qeo#WBV=nt{P2jFPW~9*k6p5#_1|>(@j%(*}%Pf%y;V1KduNK_e5B zTU>)5QX+tR@PZUu;(({t66Z7rFG;Z>&&i*f{wf}BmR_Bc{ahviIA$lt^Rb7rs5WP#ot zaUC$SE`q%iW%k}wkE>%?C5)1QkqV?-d=VS|!h1n7^A`LFr^l0NUnCBeM^09}y1UKI z7s9c09g{&L6O*oAdIB*ylc7u~lTu(|fAWp?BF|v*?qQRqO5|Q{oHQcWZgaQad`#r_ z)7^irq#Xz=BV6tz65dH;gx7Mjvr70ZBg(f(IV_bRgYrR!Tu2R^!h-i(SfAX%UzvQ^ z?u<%)3?pyC-(SM`>uHSwg7}x#yq}BBPtsTkndX7pPFX8*=RiV`!KI!UzkYV}e~gGk zwSCYI6wSQfNDC_Rp;aiGPs*p_SPv)srP8Eshf}v7%b|Xt=`VF(A*YSROTZ&cCGGz$ z)i%uRq)^gsb{BI6ZLYRDP5wAG_2;oV5rd}cX>Py~l_iHbupenfKmJ|~yni6pJ{4N4 zE19z@byTe>XoFTx;9*`I#qdOYe}Ik50vp9_lr?8Ij@a;Abtnc@47JFUBWNQtU|QD4 zq6vd8Dek+|o=7VzjL~1GZRfPBEf5n#`nDlzzyg=5e3ae`?dMcyH|Ca7qP{PQM7+3x@5^f8tmPZCv?G|2e{vu~{qR7? zqJ-lEmfz5W=c*H)3)3h&;7HdO4rUzcuH`v6tYg=+2jmM6c<$gMgGd~08J|Vlc5PYF zkA|Perf6A@k)8(_kRU=|vP6Sm7S;<1=mc(5D-{-!8Dk7No` z36tCP)<$=EBEmstqzH!B7I-pkkXaMt5wnWd62DA}EO)DoFX&t~-7%=U;CVPQ+NC0O zYBkNFI!n1P+CF+`DvY;tJRR|Jy-xNWUNqjiS4mH7of9TgtMRL>e@jz^c1?h2yh(oA zW?+rM?tlr|=1AAtf7U1BMs@+yOS^!f>H%q%1r8AG;Aj)&{Ie(I0kMS6{ntTkA!=~N zpmjYEDL}HY8Zru%ugC01M1WI8aKAHfPbOD$BOSx3<{2#GumIVMOwDSG#JEr(&Jh_p zuG}t7GR%!Mb-qm)L_3{rJn)(2BK7OVD2ul450Mjt2~b0af3N#^1UiMp$C|m3kuicx z&qXFP(awgriSyw;2*8!F>tpaE#BoHyupqNSE#tRLzq&UhRLC4KH9SGbk;W|fHjJI{ zoJzSmtR-5ZnRiOSeAgcjG<^fEUD`3dycfOKiT(1ylw~u1@l*Lp<)V08-2p zlC+(ILr$6ef4h)m+OtnV^fK13&I=9|<=-B!Xu@SU_|ddoCgdMlipl79?duWon0-cj^IO{;z& za)^OAfWeL49yt?~!a&Kf_AXmCU7^i-1K@arut|6Ce>0(>rkObqg%$_Uv|gGotb=gc zuHJ-ki0sQj?Y15V!GJA<;SmivWpT9@nO;xgSny4gCkKN2n`aXm<`Pf^R!t@=9fX0* zT@ePD0!G?~N+A40aFqt5d2{>oYqFUj5c9?yt{>`Q9HK`7>}fssSX`X>9Du)ks!s5a zet!4%f4_HIQ1r*$&t%8~$6fLo5d{~79Ex7QqCQShc2&_{G+`X9+P*pvOz^V(&4pg4 zGKAL4RLWhZQn{R`Qh7@GLzqE|_$&)TtejIH7~-tkf$65}Ei~p*Kmeq~#GV}*4M5OD z@L6IOL$rh{L#urCEpv>%R3fERq`>Hpqo2nce+<*)fKatb@8(xY3R+>^b?U(#r2bMn zqLah299O5xEkYC&wwufGRLPb-0!j5uvIpCX)WZ$oTd)FnmVcH_oaE=R8T+~#PrQ=! zQ3e5SeHncjwlmP?UB%=m_BqIW{d#IgsNANAms8hcZj;Aj$Cwc7#w=e=o=OYwlYlE& zf63iUw#`{JR09O!^@~NQ84QsW9W6@|RiU@i^r!>0)u(A4S#)!YEv{x%j6)7cn zCKFP*)g;b=h^#H>GV0LjJA%= z3S?fg5TD1Ir8yUvGyr}IpneD$9{dc|e>RET5r}Y)D-&MY)%e3i_RRj$48y?0ofqsU z>(PF_zT7^Ga?@o&1?SZw8HbFs-*JHu!k>;9Z?UrK!2lZmg;5Bf$7&1#*?<=)F4MbE zBID+XN=?Pu*uzb z)6)y4haUo!+=A4rd#J@vlu=GSe*>Q{eDvcFRm*rB8-mX^2HhuYJUb2kf=`oNKno6+ zPhH*D1I4kxZ zPK8l1=`#lMH3VyXnbKWc5V}?{I0XO35soYpivL`%_wc4Z)NO&t$vrC|e-*MCd?Av# z6jrXUpn7nuBD7;czV3dgxJk@nUi366%EGEWYjtdjGMw}zUT2n!RM1hRqR&ZS8TTTOs3e!-oQ2J_IWIZ^JfwhdJRZz@_L7vlD^98 zaX*2&h86Z70QGNiJTXV8f3HSzhRjfBFX%yrVy1rD>hgfW9&#$V@P%Su9665WB9rop z?@>FWECuPRZkpoU7Sq>YAQEYAxZ(n%S+5x+SWSI@VnGlvsz`2*9+-h!=dUb;UPfZ% z4Cn6^$ZX(ARv`A|F`iw~$?irvzSxmqV%52k?!BD*W*hf3^pVD!e`mX$#%Ia9t+Z)! zQq9Y~XIh9cGPM+f+EmPN(q47A=Jb4r8(x**)rRu)i8tH(POSik+ zBpu5!KSUUy^cLUFYw0cWII-LTWO-)^#tjj*?k1%;*LGaf&-&i;4m}3#_?>SVNG0}S zuK?!IhE}me3!fz~zL74uvKtr;2(`ZM`klTH3%4Gc3r&=cDVO$P3@U%cT1%7Twh_MjSIljwsuheU z37%5x^(0Pe%C)g4x!5_-5_gskMQZtQ?0=tbbVDLJnhoi4RW1UE_oMskuN#tg&v)|f zFJFB1^u?Ed&dr^YBA2&|!4=9ogCY}#jU6|B>1$iM`ay284reeZtQd+(EX?deoD!)|Z9Opak~UJ0er zJ1LYiylkCXS~d&|Rr^U@c_{0$qyC@9`>~;Hxl?~G`E{ot*X^`S4%MOLnm){?l5;)w zRMd`?p4zJ6qT*PbhcKtMnXuBVUJv&lf2GV%WgkaX$+4tv2Cl|PI3JaY-0%g|_r>|# zl}&-ip-i6q>DGN@Qe;ZS`*`=Ut@;YLZ*@|aIE_`X1|wnf1~-oz?-%vFb11rQQ-PEj z?_htMTB=vKd8vzIHgrZ>IPF*Wg%$PHP|^s;`(me(zHbjz$r~9PuKQ)DT~h2>WK$

K$(jq2>N~u=M9-tOGlnPuH&Z z85OF-+lzZYg2OO)CH4&r|e zl{9ToD##TGIAITM!$s$=J-0or*-IhadSzk6$hEMr)0yyxyG+B>VY-q~2TG!t_*}Sv zrCX!~`p1!3N-V<9@0Ai5<^y68oPFq$8KKkFgYkm;aPg3-d1dFK!)c0R^?c%@;}{o> z6&2`bQ=XzwDmaQXTYVJ8tz&0N#;JezRwe@iFB)7tEldwZPZ?C>P*B!m729xK|69*1 z0^?|ud0hsxxgeb42q)a%SP#{{c!uruXzJ$1^@TxJE(a^FAd?Fvc^~GcMgIF()~M1~ zAp#|n#s>rV`pI4DO1vTL9u5Jt`O{^<;kPr_0qEDyZC4E+X;;wfT#moQ%hG>R3UBmm zO0DFpQw|-E8=#o<%;v({#Sl|D%jQG$Gyq{SuBR&xz)gxKTAw@<9!&8<0R!_u6vNa8 z(LlC=qrk27REr>qUm%c0T%K}DSfe)43KMAs83+KNU?cj^&;=#Itb@^HzVIFal|U3n zJD*z?}qDU_rO;-U&PRD9Ja*tpmE z292Tt0@?-4qANcZ13?$m&7e!!6of#p$8K6*IBys~GQ6Scdr)1f4QN8MUu4ZWfXIcc z(L%ZdK^nBNd~7M)w)d+6m5)jS_Sg+X1?M}oAac!NXH^Q#Gr8NN#0Yow;jS5z#+sKi;U0W&sfX*Ci_Z|~ zv$6<!aBQc9z8cJDA4kS1@&1KXBsy*I#gCpWPnawwEgQOw7eG^c8 z=9PkZFjB2v>4AD_k(u?x;}}dU;SzWr9<*i7XM@oPr7rknLtA42_tfhEu|Ob{2^uEe z9GVWc=%M@_-{eK`q3l@B?hVi)V@9DsV}^k!)o`XX~ZDugd0s(Lx}_=ta(Gq5XWI#sanhr1foG1q8+K z;5T5VID8r_&ZCrK3Uq%pDH`_9!yfdE_1xl;k@a*uMllZU39K@@U9Xd!9ONQ1k?Z`U z;m)j?z_;9b^(ToA`)Jv@-y1Oh)-Q0N3AzM^urEPDPYA04Y@wuLI_ooG^=eFD_o<^X z$c%V>6B0dnYFOP=6BiIG$h^WRfF%`s%S!m%tzHPva4I6BRw92l>$g$Vk*at-ixaR1 z4wO0-CL?>D8MiXbR+#V=`+QwAu>})N$hKz^HE8yR znl)|kF|aafwDG{w^cLrurN~<$)m(_CY7V-X`gHv*RV70W(Ls}M5w-~%TrU`mnhgur z;bZi1?<48~dBcAo-`LL`uX>=;MDee7vQPr||L3boVW(`se&#~v=z*xYCCIlq<1Ixoe zfle#Y;JSyacVi7pYosm2(APg5s}JV(yRz`UL>lgLl#deZk6x z((6R{^=T{+jF=-r!6P>hTMR&peLfap9{QpY9Z!=^PQ#cK_G&ksMi^mw8d&fni2h|# zLONrEYdV8E4eb=U7me*O>Wzqh{4nHUOeMzx$TIXd81CfAX za=vmA6MiQ8F0^0I6zLI18Gw~+bth848e{+;)dos~X|AgS2T-g>kW9s%&P7Hk5vj$1hIn#3Niqp2njEuwBQZnaqX)cA-)!()Cn7+)fNx z86o}hsiXSvfn76^rR&|`F1VZ_7<4mEf)(Y7Y#_NiDL57Dqd3fA!6dkcs*8WocrLFZRU=v$@lpY)x_H_w(w9Zc7@sxL z-Uy~a^~Vv{AgIM77hc5zOZ$IVT=mF4@lb7{ckPlnIga!f%LVk&jom=X2!x%5G9!oEn|! z5T<}N(V@$03>}=4G*hcZ{Zvx_w*-IjI3z^~fnrb?2jgL5U!Y(*mc<3~a>^`RSaO{63Z!wt` zGT&gqdB7KMo1v6ZAI4{tglku710ie2W4J~aEfWsfwG&f{V32PR`#z3FDXj~fJe}L$ z`%pd^E=WtUcxh>)^)UE%8~cCbeL=Q4JQ6&R?Ban;osKNYD2b*Gyiv6QZ;54v5f-%a z>d<8TvIGV5KS5wsRLN;`X5)X_fmh~!=SAT-#}L$v?*k4l#4JKO;dRoQt40_M9;TZ} zE%rZh`r>iYUIUEEc;VdE|M4*O6+T=4;KVm6q~7#*DOf982+YzeJrOW#E+=()AJ9_R z)NKi@h5n0JdG0rc%EKTbs$tQE#~Y4j#(|@C%ZW!agb~5!l8?_q_^*G5I@ZVVw=&=@ zWPY^_#6hc?(n%C`J{#So2Ue-q?;73jj>h4zo8nkG{##_jOGAA9f%V*%O>|wts|V^C zHkLYD{EtHlgo10i)T)UA<4jTLBU97&Fa|$hOlDz^0{plcGlVaXi4V5}cS)tdlJ8mk z7)u;8PnXT2mRQW=yoi5V!YJ{7BH1$Nj73Y=XUGrDk5X#jFt%qmefpB-y+&qnRl#mzF$iIQ#y5U~+dfuPEx^>r5@Cj)sQ9h0Fu6O&7ZpaRV{ zv+af)0)KKdzoM~$%@gR{h7%rB7XRB ztH5^MR>=rdcHJ3~&eEt@iY&{5IA{q$2=fg{t}v<)k|R9;!5vu7D*YUe|5KQLKdF?9 zB#pgBNi3wELKuPfZ!=&8kd4&>+TTv_x`f(!;D7sd4(CDJ9It%f^l(KMBl2^6A#y)> za&x>VmbcqfGU$_>{{`#l=-o^<2tldU(a*Dp@ou666K}4`sB|W*nS#!ZxB2ZASz%8G zfrHl(lbWWPsLSUS}x7(cwA?^N}G~m`Uwt>m(D`Mnu>!u+3ZegIaZQfp^y@Nn3 zQh$jMVnu9{$)ly$`QCpYSbzM!_FsSbzZz4~mlFl_ftTgG_y&>H#E=jGeo&91sEIRz zT^wpL1aY5m342AgcigSRN*`NCc>7(?W& z0T^ikS_{XH)#qV~ys=H;#`+k$FEIQRwU0tCX%9r0u6(KnR}N#xH1|xeWOa5B4u6%* zmPgb7*|R=Uq-q&uuEoj_Whl1+Yr4>p4y&# zk^digs^<2H`9v5isKwxO7{jO*iaP&VT&~T|!c2^}2~&O(1l4xA>^)pf3AYSf!+?@4 z25s+|FL5LXc`KU2%%<{r0S?oBqJLq^6VF?wxsEIk*I8n|zS8uEX}`NS zs4zr9Z)|vkkIDb(Pn3D1nYor&lNXmhiNS1+VF*awF-b~&SqC=`%$50Qh<|hOYGlAW z>#+zZIHpv#9Vf{L%9g;866z>vF49tO0%;@Hb#CnZpMd%>I>U0-0S4b}6$JR3FB8|c zc}e>ZoIHcDak9EnX;^zWL%Vttx$6XFUKIos6r=49MH9*Y=P@&Nd-C z8U(rvhom_O+E%STH}H(641Xv{E|ypzov!wo#9 znmb}uf8mfj4q6bkGKA^*c~SlV6(JBjG>}eE{Kg>*=+*aBTzVA~N`EqNraoJUCIBS* zin_jR!$pu)kj`lWr@9n3 zwbrE48P-(CQ(cwS5SIG)2pD&`;q!dX&gxR=Em@Tuk+MWcm94*SynSrA`t024967Rl z{c8s%HvVwwuEqeRHh;bh9&n=}8}b<=w2O?>c?Gt2eK20&?f7+g3KVaYADT0q= zkZoQ(P%M2yXNtB#ThBq#3j$e(j^p)@M4rs~fvaQsBOGAOJeb!6pGqbo=J2)V`Rd37 zD7ElCx=snea2~G@FjW`S_LmU6d`tqUk*sN7+(3(L!eHpG>PuvahC>o|T*Sj$Rvp?Ex*s zK8i0iOBzZ(N24fRsLFMgt<6&kVpbOuIyXXQ3GneWFMqpzcfWamw_58oi?vv-t#$Fs z-`{TT{=2zD23L}$o{L;-Cxx+Tv`(~83Gq7|t9E;=fSpO=ZPoJt8AUD?)}@gYTC#?J zYM1z>wGvvnXnnF)m33ctEpzC{BMlnnT|O4=uG&}a5MwDaQC55%h=zG%Y9flJ*E6&+?*A8D#sX&HZAX$w}QnT8cD+Cj{`8m+Tj*wnDj zX%Mf|IoZdxl`?*_QclL#e_rIn%7Dz$AM9qeHqOLVKhOZZWLy&OyY0AA)9aV9ruCga zn}UeQO*S*TH11imAFIHR5i{qR5z%tWTmVh4gF2IjTz(2U>EVO86odlYni=qYkef4y=3yxUwU7|2LP zDoKR{?Gp$^7)Ry)>dTd|ew817$|ecFCeA%Aytv<+=v8<1ANUL01M79*y*^=n35%3a z1~iaHa5%jtBmy$LXAo>i1!hqJhVEBur{iD!?;rm6J^ZTpZrXUl-+<14x{&WmV^M5PAF#<-}{z!8DWy;I7XgqSg^-LRPFBm z_O~0FqacMr+3gQq4|g<7_%O~eI6b7{)OUf*L7M}tgJe+7CV99_R(^c@zheK{;~+#< z#h6hv;)Wq#D)|Wr)x`jVpLOCyVk~id1E)v;RtV5FBpa(1e;a}Tsp$>}Mla-4@e_d= zFc3@tc=6q>rnN0gFZnrqI|P0B*?;R9(VZ4LPoIaDbgtVTE$%sVL9;zdADf}x)_qwY znz}9OV*m-A8IdbH2MIijg)pr5K>cO(XT~?KE)D#H)JS3ZIPQ`#Rxm&K;So#QIfCP~6 zd*NR>8yz8x&!*moN=Av?TFsRVL`#zTSeG)p2rSc>f97Az@5>^&kVF++$@s&BYLI$e z;z7;AUD7sy+Qy^bdyj;-E1Sc_1cXm|>r@!vEQOQ#+*D`G!Eh?~hMOqwP7s&Smu5Z? zguc2YFM-lgO~8BY?QnZR9Bj#+0Q#Nh`|jjKB`Ej$B+~kmg#=A&0I~0mKPZJC7qG3yq5s zs;$1FmmOl+k+PT8kc0EN>PkMRdQmcqAmg;~f5!Jig~brSWzkM^TW|aXtE%hn^MJPO zSgkFygVGJ~wCA%at%Ox)MJOafj~v!Gf>uL7kMEc%IfaAHa(N0Ls=(Az9f#bSiY7kS zpb|}Oug8|Djvegi$FRV(+(5KF`7RiNPU;CqA_P_WlzA@l?5sODbPVBaQs*ES*3>U) zf9UC!#~7IdYNKRnL0z!8>k5zXHyPA8;S&85)eY1haZM=+xk0Jnm51Ul&zI)kpoxDv zFga&_NTS>~IhRh=-28gthLCZI4F$S}QYihYD{9CN;K3X2=Q1m-N#|jSp+^!71nrNy z#ULXytufm0 z&G(dkeH`irxEyJIYE=w&?LKQP(;7`3svZNqVq(oag3nrMEsVB5(@7&tqv9R*=r}^; z@*6S<`6m6@P-HF`OZa4<@nFK%Ghm&UjKWCC&wa<;AS{R7T!=CSR0%g>Pfvv5fAmnb zk^u|=l63iE`+5tIbpanzr_ZW2aJ$F&me?nN-iyJz^5^lDS5x#P270*cNA=a2dQ+L! zPQ0KO<`JKLTO;bVRWq;Xp6I$3s^@CzW21GJ35{_C(c)z2EiT?}tLz0aINS3or!}V_ zt!2-*;OV>tf8gvPjIy&$mrzQbe|&&Ik(dj;V!i>v(e{$wcy+Bx$oNxXt{{7%HR-w- ztjW^FUk>=_<8ca^dAI&H`N8n9I7VkFyb0;4zWcSLU&j4BnB?d&CKcjB@OhI+4B)I_g6vexVKS|O{`lS zqK#hx$H`==Jnpf>CppDm5bz`(uwVjO*dCy^r1#M9;tGEgNtAHx<|PsN-8Yd%AHevV z$ev;C(ig3N01z3dvI=ExlcYNkgNc{7iI)M~&PoSwS2|Tb^?G&pu<~9RZC6F%RO-#@ zZoB$9F-ohOb(ZDHj~{Pbn%uqr_~YlBwJFNv?H#w@-oE?z>HXb5KD_;W^XuI|U;j{) zm#=05>3@^Lt=6tkF0;J(7ProPo@~3#xPzVLG3@sJ&00I?5Z_vta49*RNuN7Y&S#Q2 zpsF(xmM@dr8=a@g>CrNuCG~-KYbtK{6>JI`eF&SduMVMWBXi-88w2mREMVib%C(zw zqth3;IrKFXsy2u4K)k)0cBVx(Lp3R3b*?;II)AO}Y0jzA`Nfxc=sVm5Ik#GaAP(FK z?Qp1C_!DtD(9is3s`;F1E!2b<-5OIp)U0{5r;HBWCR9y5J+-oI&j-?F28!SpZ$5z7 zrwc_oKOySiRpG-1$b)!Uq{LIh?O)iKBnmjyssN*40gJhgEvxRXYM(|?_wA(isjY8p znt%Km?gJ=4vz~h8jColNi}}Sfsam#qKyqCx*F&3)+#c%Peh|As5K347UB%wN&W-^Z zK24ZqO{!e!!SKD(g+|1}i2_)029Qcw0I0rRJrLfV+u+CnoM4c#%E3@(jk1o9_2KqA zX8Rkq1^{zJxFeZ>slFr=!H}M0OIFo|{(o{Wtjy9)wf`eV#gf{9T4m^#4JF+fM?yG@? z21cO4*DkedInmfN8L}=M>FZ%F8V(N0mE~N+hi>dQ_^#KM%op3{I<%~dKn*TnCx1|f zbK5ZNyEfe4m^8sd;~4Eq3u(Y5|3k6X_h^O+M~dW4eK^FoXHvJQlHOe=b|e$pO^75B zT^NAlSxupQaRQg*q6#);LG=QyRCLRaa3Q2CKB^Gn}fk* z5lmoL0e!2nifGP!GC#0w?j*&MA}I#e$|olaC-Woq%ej=3)OLyNFsn ziEvd|SnXd{nc`GsaO^R2NY9Nu|1fWDj0tV0$&>Z6Vnf=U7ABAZ>`z~mk@W{xf`1v-zu*V*_^Q*kCfy0&g$F1|mZt1?`Ao0P#t#8h^l*HuBy#!h6Ln!gY1z%<}@gWxm+9 zUh#|rnH7MqGn~aOY-W;O;D7r@tZ->@0n*^=R#2`Q`59xI(4xe~8;w2y2RZnn+1;KG z0&cJ@&0;woY6co4LLptIkEzbU2!}dhXKaOqeKQWcOmr6S6s$&eEDBX-#ay|%{l1gn zdTiwavQ+7Oxobd^GMJ-mHtbp;jMU4QZxN7@LPhvsNxHR2mq|;z>wj8~0xhb)hI9ay za`J^^KViQcg<1>*0DxXiBjcf4?>cc>RgCw@0~32C1Xb=-g|1JcX_oN59%s1+LZ~bC zQf%PdV|K%gieP(ox#>YPY$R}j?AW-A6kBryfssZ>M`T$dFVC{Xr{)?UCTWKnz=q|C zqZH6*F_%I0t9lM!f`4e?HasAMo6t-uXp}CCxq!FV><%Cgu!|E(zQJZh<$=%vSY-1C z-Lr0S9>8&KGs?9B9n>{!_k9O;Hjq7)=F1G6-mT-_;L^e$VLK`wIG@?#j6|ZuPi9q` zy16-${HlN`T0&Yn4nyE>A1$*emZyggqM6EGV2?eb5F2u#@_#bBK&6n;<{G5ggdtiN z43W`di;he@5#u8SZhrZ3ntB6ibQ>Ti9GK#s%^gKWvflQx6&6S8byHWY-)(3gV+jPj ztZ^Xu6D{o6!7Bqc47;YLJb&b)DO44gu}{X%QB(`h`Iis@F?V!6i^~&8_;CzV0A_u$ zSiOZWa>mHeiGLll*_9v*|8ka?pJs_jQNjg`Nce=kY{g)PTZOuf1r?eQVeKTw(2%#n zg)Fvr3EJyMoP7$BSbkzWgScHoalq2>VG|Iep6d)JX3Iz53WNl*l=j461>#VX>E%U} zc`SSs9`&2BiRBbt4c;iroMh8N;5~vafQyts28`hFS%09y<{IKUap-W@6`;+)!}AzV zL!|$Ap~=}=ByO||e;#6=(rDVU+ug%XR8WHPZG2%y+>R>j#wZ8(U<&q!}; z2&yINgfPqJC%Q0?Xd$dNmsVag>tbmLayN`on|i7d+6)i9tNAr7SYf*XJQ{dbke-Hj zRS>1?@P4NOZc*C4O{a?2GeWK_?bG#9b*D`+h{4MACgrx>1kd^$z z`+uFV0TS0fLy5;1zeSMs7%6*VOXDXai`zW2_=(^7KJtPOZ|H$nd_KpLw~zstcw&rX zNU)#Sz%mLNzlz^8dx0iCfA9N=_IX3UUwkk+Eg6A!d+^{T=(VOwG#7gj@gNmoYYPL~ z+ElXfklN%Igp(QP6!Ph(#{o^jBu;Fqfqy%ce&7s&Fbo|2+Sl`UcindKKQ1qqvQRg& zek=68I2Fy37i3Pv(oNwZPCNBT6<4l-t?RH0m~YWjBco{eF%=i;IQ-iP-lOF{gKwD{ zNe7H8!8k5J@^0MrkqK$1x+fK>XqUd@T9YUoM7&Nfmf0@)NX9&$E933l4Pou70Dr3) zusjB?e26SiP)bt-;n?9gM*=%%K>dnK;|Hm#Ahmrp3{&;cEdSIseda2_bZ>?2z0hwr z0Tqtf=JvjdXCv*zAqp_D(@e!a@RcsLr~cqr_$E_u#}~;Sb_IV53f7=AkY}pPVP`wN zAzAOy0tgpnO+2elJ_nvh+rsupTYq?9&MyRJN^5rk!%U^>qmL}%ici!<#LWA``z3!d zDM}|og-#rI?vxhhq*AFwpXCXqlE zkh1sJr+a!pz(KOKo4oO z-`97@V8O$-(>+%mt5qVINQmDqU#=`yab4}&rl?AK%3Hk3xUAa0mQu%#2r51ts}D=b z;~nl439T7WBhDlbhj_=8@$XiGNnxYa;GNsMz{RzVU!+r(W$Ea9Not_)&G`G~7#I}g zgI}aFCV#SQ{Jlxeet#sXK_zhfYd;EGZe2arR81<{2A-hW#9Ynv5C0|u?IqW0FBqyh zqD{i#<@pUKst4)GH+UrA@6V{a`_C-xZoT-t5P{G+D+hVCj5d4!Ng@ohMn`Z-jEIgd zdRY8+W`px_pjfJPWQArNbmV3I@5>*_yT7MP-G3eqN4iCC#p^rHWl6VNH-2^abX*D^ z=e1i39e+o}quW;HHZPmyN~kRUN7WL2Q#FUx>PaJ)SIU6I!s}hk87&xE z^e!LR)WyEo<;C%$AxKZrN~x3?36ToQ&6FixZ||!eoPpTgJCx}w{&YLIhOYUJg=2#l z`hPw1vTAtpCqF0gZP=vLWy`=kx)$EE{qFq@taH}e>*kcn&h zP5n{@m@S1E+vc$(Nh-X&Ny_6or{@kj7(M=!BZz(%=ce#+d6-ICuea>Nf3@)6T~#l7fOke`TP!GHmg~7kslBVAPD7>U>gv3SJYlEBopeu z%@&c#D@uY$+>8KcPEq0n^u)l?1|Apv^VCBsZzm2DsHGWyrfTnL%s=eZ%If{xfV?IxheSfG>m_k zOtKr07QI6&^q)CBfHz7^8npcNoEkEgCB33ff+N!{aZd)W7{U-<=DX?`Hpira%O;^< zQ>L`OrNo^(f=rl}{Sm&*9l8^bj35C8QHZ*IF4`Qelp9Auf0 zW5BSMC6)|pmA)#?C=1!dgP_Gi1hs!Y@5|OOrKSqdd~8FVnQn!gj|iwllgshIp{7m; zdYomWu>Tn=?2#-+eNz@~E_$$JcAf@6@=V3AXN> z_Os-&`#~ordK{s!pfJHr+Q6RjJP@|NbU_j#BT{-~Rh+3TyE-$8Ih==k?L5sHwgiX~RcD;h_NWmH+tT)a(uK7HY4ISc3w?vPZ zq_$`D!x=6*F%wsr$OBwXcnIQ%IVZOaZ!S7Hrf53}xM`$FBThTMWJ&VV#9~8!&HXdGFvh zQYWj{6D$(k2@FC2 zjixvpT&N{@23`FvgQqde&g_W7z7&Qh8^2r%!{aB_hdZ^>_xgX|Jhmxtw}E$U!)a)khsQV)0Si8Vy;SfCKZr|ia{KG#azVR; zu2fMENnm6Ef<-z8Bl6GTD9%?2c=Pt9JLDhT`SFzGG@j~5U(@T|x}QQqdH^7Xt4jUc zakF21UMTM0hKna*se_0M|FA3IvU~OP#a>50Rg2&Jn3DK+%vXOM_tk*?+`JOnEe#Zr zF$IbUeP^dy_)=O$n#B=OFwl$T%E=RGLhbIA;GFIzm4R9S)RcEsl8stX}+yp;&b@ z+1bft`o6>1M^=9a+)A(=f+$U(uoA)3d^wQ%P{)?K^|?;?b0Rswnr2#L=SLUgsRH3z z$~(9QZpB{(b-D&hl1qb^$$TR4DT&SJHpl<1@O&g>PI)z-CO`Nh0GQD<4Y)mwMjJjX z(}0CI&hw5Cs`ZpcK1j!E>|=Wjl1Zj#l?@^GH?&<_3<{H?I}(RT7z2k$7zBq%7zKw& z7zT$(7zc+)7zl?*7zu|+7z(#Y7z<<}N^-MMQsgpMi|f~mJQrG-MVeY6bG^7;FPtdT<3E-!uYdmLhcp|v_Bjhe z0e>KCD~sr7ELk8{fpYL)4cg{L|2HFvnYrcjm6EXbSuc{ANOfvZBE1CGOQS$OUaq`O z9>IT={1n>M^X%6CAvRNKUjcCAa=0=rjApVwT`Dh=%l)hWRb^YhhRd}kNS3uaDXSK4 z11~JsBF(aUF$#e~J_)ea+%XUd+ACZ$wtrBtaOSI`sfKu%q%4{xu*Mb73p|ME6}+;k zxAmr|k5Tq@Eq(%v|--tTrK zOfnUjzn$IFF}n*|{wR<*H0cyk3c|un2?Gn~iV)*^8HkQp9wL@>4nTXqE2~3Goqul# zZ*4@T=F2Ic)(B(gV|H9B37YbXHr^I>L->%Q05zobd|iF0Pz40ycTow!ji(cp4QB_|)Ovu#V%Mij2xh2uX z#smB~M0hh9+3?gL4W|y595x;I6o~d6Xza;FZL09BPjX8(xu@#xkXWwzx)*CX&Xonhh5}N&{<4aEQ_J2AaKeFgK z;n+!;qgI=Jw?49Y-0*j#nZhQ={VO#8o0D{FeY-B49Ag`;Z{&3gT2ap@1KhL8lYcOA z!EKesp-Uy4<@@{{%{3B&nbY4z& zp(V&19TgkpetRg6NHzS0GJs!pxv|+EBL%}8A83IC!6?;j7eQ9!{fB1vLC|DyB9eld zv{TUbb%T6AS#(0FAoV53qDQpJfaLY-*JRB>MshQoJjV*PegM%_8-K(Ao8f9R3*N_p zMNb0Nm7OnosFq2R19lyK2c8K)FVXU@deOEMM!7@oiq#sgbc8EV0wRc@IJ>g#Xqa_d z*c4^K5;{4Y{FN1|-%K*6N5?}Mlq`N@J?tAFgowd%!4OE{AL6e%)r zf2b)|M&?P;w#=S^5xX`@70wC_kF&^GR2IT>uV<^^KCUEaXfhBOIq=Z&0%EbT1Y}NV zo6QX$fCT}BB4BW7Ti&4KLsq+{+7k4Nde)bGUOPB139p5%Arc%3@7RO^gNFD$G}sRlh1}bylwzPVF`jpa|(p$s`#6z z2F+l1hFhu(6oc}guqWF`Fdt*)y}@j&6B+xANgn)k{0=Rkm02`qIJB-XwS}GJOPu56 zA@C6IE9r+?L1z%UoI?IGPD&S`MN>(6I`Bj>!vdO#d4EF1X(J7ji!4tUgK2+1yQPBo z-|4txac)P{ez)xC2GfklGZz7Udkzp-Sv1}^U<|X6aDFqy`rETuTW}%WUn1?k6w>;y zk+x%`fyvV3!Gj0q(N`%*w*Gqf``hr}oy+`YA|GL2wdavG;0w5W@}~YA-dtIg!Aw6@ zfp_z$pMUK96?xQuF&_PVq-XP}{~A1cfBf&qqbayteu)0IYJwMRq#`?k8SiLt6}Sf~ znT)RMrY&|Q_r@>;Cj<+M$@rU|N9eD22(Er#ue)kPZYD1nTH|11QVk+7ageAUk){I+ zO~2PwOa<)7oxwVVnjig<-y2g3DQ7y;m`Q_J5`QB8`FYxy7O5k?o6CB8*oU%0!lWCb zUA`@-Is<(-NXJt-Z=d!l3cfc&!h!cDC*jGP&Hhz^-n+NGFW?>dOVAZdPT1aDa(D$% zj+^g^yz6?O;M3v)?IF$va0UW{96|sN3{nO10E)p134YP1#RVX80-MIs7L&N(A!m$2 zDt|Mu)4o2=3I$BYH9{oj_u*Osh5OW?a=^Oz{75b22H8UdKq24)(B#)a$6?Xi=l`D9*Dg~q&4?`JdG*QApXTYXpPbnEu?dk#-McIogc)Eyz)wy2 zn14HmQYxIuMy^MFmJz(r`Ox!#%Q@jSG=EO0lV6N+;aPA7Aek@aJ^T4avOY^Hq<@k9 zI^Oe_u_M9^mhR5=yU8r%smYMed0#%rYsiJKKz9~(=bbE>@hv#Y!h&s2Hc zDCGMd6o7_#Sk3)C)N!x-nywmDiAN7Nz-btB>MB;0(P{TNdKAw2u>zT(0e6k?R)d*T zNB(!%nXq^g{yz;p)c#{yalV0FE`Rco98KSqC=V!tn{uuaU2~Sl3v1o%bRY)gN>N%? zF3Hm@iF%`ekxfhwr#(&P)&HNbH zlx*^_k@Vnc8^g-hq>j(fzn^?!w6EOH&3aTwzJ_FQC!p7CWdY<|+< z@evC?$tXf421nWZpOp6KO@K)q;LG3w^`L`FH!+fcMto8OMkSk{>5rAs0AhXORs0F8 z8_2M$A`X#e_=`COT7SinRZT312JPZf2sdX}(F+3A-Pnue8K0a@v@EKkCOiTuKrK-T z6Ays~Ewz&Zcj^WgU{jxbU(vKsI)u?JKT}DC_mjepl6RHzXmzopml~k7BOlclc75kmk34+N*VAbPL+qkiy)btypFta)=9ixZIUW`5A+O| zYeownf8aI{!S@&6-CSI~P0~P$C=;0sZoUSIh*E2VBr%c5RB*EmK7~@Pe!2O})!R%@ zXR1s`A~itBO?va^*B{>9e7s(*Rg#7(TCL;QhHw7`}gnDlMw$yC0vuZD+9 z&ax12S;{ZJC2eF5@Sakwk`zNmW| zg)O>47DwHEMKx)7X6~1ISPp#v0&J?mjgnG|05l``HSnz;PWD~jT-;ws?qDminTk?a z{$@v?C*NYz?Veq&Cf9fcWV<4SI0`$rq$zR`pM`ra0R7PxUn%%xF z>TO4_`@DN%GsG+Sw6<1;4cL`V6H-5H_|xFmzHNrzoNSa<5$y0B9VD~SX&Sy?Nh`w3 z=F7i}O<#V+P8*h+l9)I(#m4rstiO_$z~LooIky&70!a*bCfKN(Va-Y(Ct#HFe^}{% zTY*{>RmCSbRf4OxQrM{wYz%T8UE8cSb;LsD`!Yc806kG}<5>0xer1Wxow^~~je((POR_y+Ng#R+U zJ|%IjBhX0xQehx+QB;HmEL3PXRb@$HA2?pun{r=xO=AV2_y^~1v&ABz$jLr`rX)=F zBe@v?24U4w4aq;ep|3xSSbX|eZtnVmEhDd&+(`=_e;l;XJ?r`i4tuRk78$Uzwbc+& z^uu$NleT`)zB*>UjOnK}zy-mlWhEFWNHhX0cPxdO&cM;34y-3LubsC1!~G{M`iV7% zwa5TWp_iLe$`tD;HlrTWda@{Xpy$Ne;l7mr6LWnX|3TrECW;r z47JrEf5DCtn*JJ9WdoxP%I8-!1%zTY+_tx=0CvUV1QrFe9aJ$q0PCWK6-UC2#9WuG@$x6r&A z$OD7pl6JAVyG>Y^322+J%g~UrB<67%$&5cEe{FF7mXY?7&kACBzmnl_NpJ$+J3K+R zs7=$hfO&j>QPCm8K*6y!FCEx|(Tz`AcqJyX#y;+wl6mxEGFw;-^W#%3x6T(v)X89s z9U$gd_Pk){*1u&@>r_Ps!X=?AMj#{tL?+8~A&>z1>ntTih>gu%-riCy*OAGVQ!D8E ze=TiO#gI`h&!4<9I_C~ls>^)8ugXpS1>y>uF95SA2qwn_c`^KSKqZ@iPIybg<_jb< zn7a&qIt4AqpM({wD9O^rTuE3+;4^* zxJl^ULo}uY6Cf*R-SP=vA}79Nf&dgie`gnamg#BkGiL=L%EafmlGAu$*VHpeVxl-p zmOF}yK5m9NhrwrYUAc4Ouh>`}DFw+Tib_MUGaU0@S550aHX|ABy*YO7@A9r-N009W zjCDwIkdVlkr7k_++%ZuyNuaXjj?uRe1cK03(u9g9Y?@(T6+T}`rZH5;GK$kNDz8B} zbuzCo065a6*^fQSez43j1*0IXESOCD<0A}X6F-~jJ;E@VGfbzKo=giCf6bUBsVAa> zm*09@`k<^c*hM_cgHdR);cAGP`lQ%olk^|H)iz%2dWw^$E7M3y^|IqVJ~Xg~?_FVAlG1H-PS6+@A>PUigE*PWJ zc&XC(nwz1Jh0oF0Is7GQ9h99>3?d>$@6jP3n{1|t+lMjM0v1!tT`aw{qyz+2 zbuv937$2d(R6r&%XF-2QfzTR!ojg@ypDA3;ktxz!LjLP5HlB=*=^$%ry{4r#fFQ<+HDUDqo)`kDNp*W01*Y-ve1AUBO{fbRAww zL7L|isuXJ}XXjH+60eBvwjhG7H#uXIvvWlxgi+9~j2FxV0e|V`c7#XoZt)IeYraV; zYhEeLITUwr%24ybctoT&3WDRY|8>Ew1(Idg<+ms>qv<{i8VWAx_9iLdG#~Pa!uudl=%)8r|g@qEB(bNCuat4ep+hq zp=FS2lE{>T)K8klsErAXMGwsLWcD(DBfX*{i2?O%^)f5s>W~1#j%R4ZmFL~fEry-W zV-o7%yLA^6nI0R1j$KZ*6p;7rNXNNGM;gOb%UFkoJw}!ioS<1=B>%?3@w~ zJCQU_Bskfn`Gk5w3MM-T9a)E?ht@BlV4|D5yuS6Fm0#Dv3y`OX4d|dg)*!B>F4v;+y!7s{0v$632&@2 z0(#O9i;NzSumqE_#CH7K9vpFjXL_^i4+vs%y*ZS)e8GK`K6VYBlfQ9Mm`|GJKyt_H zqBso{Df-tg^3}6SVKS=}vhHS)QR)9)8^`LQ5qMo*qj8gM;%dI(E9NbKFicBu(+H!b zk;u|B{TUH(SPBXHq?~clI7BkR=K(z92Uy;u-jYeNHKySJSb%Aa5iW%|ovOv|(C^Pl z1la*+uWgBfw0Blhz$gQHDZQ_k^CByEQrX>mXz6~hvoxz`7YYuRiMgNDAWKjZkClD# zWl^P=={PSrU3w3`-!<_bM`xHv^~9AEIhgT2r{@%)c@n2?p095%{tJQwv&fU7I}^9Q zQ41b*0W_1LOemKH#|ukHQ!Bw)#xt?H{j|z5CWTr#ry0-W>UOvKm`EnrdSi@D-o9IF zp4|TM?(K*5M!Gb4eM`?5Z@z!`{)gLt|NQ#H`nTI3pMB@jx0uHZ;Q@as%-JZ({2Uzj z!5e`I)@I=pJAZyf-qDL~-cWUlig9BXQz9y- z0-Oj$gccm$p?yje^Xga!8Mj5%m78rDM+X-!bjqyB#>{4XxET3xpwP6(Eqb0m$wpx9 zs0^1xc5@={avxDZ&MJSWp5ODPi2g=_zj0%KmNB92Xcs z%QJu{os=+|rPILmfU}CeDNqg6KXq+a7lFY*Ilhn>D5SS>|Nhlkl!D%SOLFkEfg_n}M!V zecTpJ(>2t+yMd5VqjznZ{H_m?I5)4U(H`Sq3P(LM;3=Sbi=#;y~N*|Ii z18^U0xlzAW0dh}y}OTdZhpe!^W<22&xeEToDepJ~a-uv&~QWwpU%| zoL!!V1s9~PyR7YF&4V?`OeRh6?<|vIF}teRqIW;wvXF?kKWqdS{xrxVAighOM9B5J zNEPviCwhOZgIV9jS5;&oTT4iG|;B7)W2R3m$wdKBS{DwdWhHHF+ORYRLXMNZmzVo>PO%G2WC$Qnu0*bB} z=$E^fzsfd>5Q*O9%?X8wl0o1Fy`IR_Cqo$2d{2MC9~&tZvtnWzHHMVD!Z=tHFsK4j z=pBKDpq{Cv+FwMZfvG7z_eY9P*xrSAZ5?xzo@sEO`}P&p--Dy0UW>Wln;WX({Ifc9P1SAfl3Oc?|_s9ixgidTBeM1 zwbJwQYw*7lcO^yd8-klIM}mVBjXOg@EsLI(Fxwjx|6? zef%5%A2}*|;ozHmm!I0WIS-BO=s8~=GZcUxnoo%ZA0HF>ym6)t&Q~pmhvYjyzAt~b zf1q)I7^kGl7E_AzvGN<_qh}xZ9PEg- zi;I@~Fy9w6digG}=BXH+vf3Yzx0ALSb zU#iW9d@O^b&aQ~)(OeM;zF7;)lM5AVWJdTxVk$zX)ot4@Ry!=yPA+#KL@1znY<7pR znUKsln@vj%-HB@C&%P6YFNbwE%cM+46}~XcjtYU`J4GCIibKqBh=Cy{4*7p(mWuJw zFAQmB<Nh}MTLQM*LZ@SQ$YF)~Ny9!TwDwMw7lOA-KpLejClOA2CzE=9cHR4G~ zpc$k!6F(iW5u4U-EmL+ETO>c97TOPeFZEcpkg>KPD2^ddtJS zp8S|J8ti>}75Aq#5|))>roZM| zu1&(hK4E?W4Kbgf_#w~z(|85*9p`B6|Gji6>B;UQl`0Q~fq~}t70XcVg9O8hMCg!QGY1sG=YtPjZTEhP*H73|@SdQztGZZydh(wO z3`8Fh1z8p@uslif{TwFHV?C5Y*mi#N4lz|X%1rQ)l zo6)WMLVU81HABYOM6K!g58F@&@{&DpCcQ=qX{=A{DdwjvRq7FISU!i=(BBR#<)(=N zT?a@Ne5BL~mOR>sJIxh0FcC)#iREp`b?BnLYiCBO(S9D{RPTShhokDE%Ht;vBxhYr zhMZ+r9FU@2jz&lIa9Fq#>DX$&f5i4bvdl-7C#nvaG!1uT*uubu&j8msXwU?9K4eG; z*KxKF`PiU5oX%P>yZq)ph=C~sK6{S_0hK?bGVX#$3+pVxGCgj>Mvh8DIs3L}Ng>q; zX5{2kk5OVBFyDW+ijq0G{I+ZpO2ZgAW@3=VyXeV9)PfLX906~GMTPsARwxibiz`4a z5`0b%3|IJsd?=2YQ^Zn-L z^_w({loYwlRkV4EQYkX$qBONa<~rJJqsLg;)vud>ynd6L@k*T=Au|h%Xwlsd-+g?y zdAMJHt#z8kTCCPd;^Mo1{kXaR_vRiM+(<8WE^{4Nt*sWyGQWM>9J{JHu7GIcU4`E! z{zXbN{K>Sfr|oX*=4oYQ+;$*5iTA~^ZktzOWA6v`bHig3m%zJjR!YW0nC^%Aj}>h8 z5{;K=<3rJv?a=vU+re}As|*W(1c!rzt(6mhxpA!i0}*pDqKTx?QZe}#X&lY0mC!1U z)}y)N;~ky?6gSqe;xzDz8>{1O-OqH#6@80J&o)(!A696`^E3+ANh*>=O%&74o)vQw z#j0a*2#l+#Bu%!gnhGtXl2{`G3CfoPqK=7#|yT8!G{9$0bAA8rNDU{O~dCapHEWdzluo~&A;zAH{enD zQ;=IySjZ=96R@w2%Kgok8!Zx=4&Ow7RMrV+>33#CqQlJVcl$K@dwcT}`~?Al`#P{+ zkN96GB}y5Qnk2G15vfb~cyG9rz{*UdIyF>tfQbrqB zP6?ISuT7~&V@sc}r}7ucjUFLvVDG1s8*`C3OI}Q%*XSg;@jvABPvrE?+39DMwSt}C zb>U`3b;cGM`&9$7_y`arDE&=R>sE~;Epb9Zfdx=3Y+}zve=cwjyC}|IQq^(M;LsohY~IA z;_fS$@g0{Wb#@@|fwN!7GBZAVe`cV4P#`Nf_@U*3nQP&6wpc08YJ0*3;1@lG1*eES za5c>d_|&YsOEC+tN5(A7{r+M&woS_i$#q;D4i$$5hzESUnB9pVr;4xpK=?AOR9y3z ztptNz!QW+ZC`zxtvPP&xEe>twu&r2eL@Q*Srv5Ti-N`szFsFnDY^Fhqe<62nLb%L8 z{>uEi%cfYidM?`1#PS-s`3&9=&e}IHTwI+m#{OHT0+w<9R%r~W`#5+<$4g&wjy@#=x$IUQ| z!+i)A@>GRe|D9@b{is!`0L+4}-F3f9jrM3VAnUV=`?ZmAGJ9KiBT6?kpwP)uw=zxoRRR|NnLCSRDPG zZWf(R7&mj!wcvd4O-R@mxTk>N{KI<|E{}*Hmzp1r?NW???E2Mn!Ob{ZFkwsNAP4xe>p;#)8e%gr+2r22kMteD40 z4#AeFWij%xYu96!J11niI9NcFOq0hmf1lTx8R-r!*3-yff3S${qf|;ST`vjBmi(3J znazRFdp+19UBwcMq3>&+2%QF%yse(rUXH_?KtUkdVSfLnbFy z8d|c+(`g)Bf650A3w&LW(eF_h{qrL2euEDe;i&^4ww3rWsiE`YO9auS6N$1Bh_z6> zyDEC3hm1=Cu0DmT3<@3K2?h8lN&Ow_!~5@j&pv$Shn}8c6nHYvmcRA+RGpTK>sB}= zj85%^f^0;t7s7?RwmtCO&j7yBiwguFmcf9|9b*<`f0En$1uq{Ab0L7h^?a|;B6sgAY$ za)_Xqar{hZ(ksu}yoM?TgUk;(x(re^6yt>Epw7di&$NcXT7%6XdPFXZX# ztQAR?5G;pgd@H~7zy*2#KiA>@rjwyN6So803(aqrAwV7}m-$N!PXd{nmr6_wMgsb9 zw~|Z@1PFg2=s}RBGRR#JAPt0cmbdyt8tM;+g0v7B6BammmY{(*w~q(o5GS9c^UGNx znA8oU&J%7N*0XzVR2mhAYxRWweqI8Q|H2`8VL%uD@b)2QS9ZYpaQg?0Kec`9>_CPu zaHC#6jL&_lX9&3=M_0FZSC3biAAk%dMuu2_)ex5zRSZvmo(Qy_~lw&v$qIYP{3W&UTW-Tx3w zj*k^H-JZ^JDCoM_wFgnJ2a;Z2DCh&C@-YAsDH0Iiw=#nQ<`^r} z;jlFGSAhHQ0Pu3n9~wx!y(;K=VjxDa$i`Mb?Vjd;^&km`T)^pU?7d#uK4Aum3iQv| z!I2zph?|J*q$PA+N6)=don0);VvH#N??aAfbbxN1@xHQzjBB6}ho4QvON>j3;bb~! zciEus6{G5qpX!J+Fj!qsD0xapr1URgr8Me-;svrxH~KL-iFy+#H;sKgmgeF4l-Hj? zX`Fw5Sc~AwfmGQaMr3daDti`@;do>}BBn8c>TQP1CGL_1Fu48=IiF-Iu3(arzWO+g z;URl5=Fc*w|0V{UTku$@ut&fi4MGP)g%2I*0v#LBrP7V32JIOiIIvJ6jaV*ji7qRT z8I}u`aw7BY0;ps<%({g$>ZoP2qlSlc{p*Z>8+~ik3@;ML-C1pE={jw*Hw{Q*^L+fL zB57AArU7Iwz@aBh1GLqHr49-zPX;YLoQM?KDG0}~{Fj}cP{3f}pzk$)Fsbm8nreIS z-mG{0DNN1fR4Fg|kbz2;b*hOYP5Q)BQ@jhm&!#ARa{=8nOmp$GE_Yidv!myMbByDE znCiH|$4fwR|FX5aQISvf86qFR<X-?F@O|KQa*(L9iKzlcl94*7rpS4!7vn; zzE})40X37MOemKfSqwV|fNT&vI0Vp_Us((+e`7e04mldfka9YjcR15-zThnDvq~@? zccag^0QsnLdN+B<;w~PE4n=j!uIGx%k>1i@Ba;jh^kRa}9L3-wdqRgII0!hSU^qC3 z1IR|_jW1{+%7lK2T#qT^K=2QpiMgofTB1_nTu&!fqsBoDh~R=q(TvM!I0vLqEnIQVF}0qBFW(K7_8{( z;XvmN6(u#x=Zz?gI72q;Dg%3&FtBKw0A6ph(@6q@8J6xbmL}^28H(kgI4O`(f1EOK zh2>Ge;EKW@56S`+01uNh6HtLp1&*UQER^&Ln*xaTo;>MQmtSvcQyq^+<7uzDn7x}e z=dXsNPrd5tc=EoUye0L<>*{57R-L?cIHBuabzTpq=52J&s)w+%9(cf7Tf%C=yDU5h zV#nsm6H{H7>c#kSY^qc9kK5V1e_Qk&jz^X)djHe>^ivOiohefi{+GXXp=4p`Ho7?K zTp3{-;X`-IWVSxHvWs)ex}2J~4&WklZmM5@`wy1evQrEctKL7Feg1sCOSY6oRs!1s zmRON}J|0aQO`ZdZW!~pGaGe*2z!|u%IYE!=&15{ds4+WJy*Yhus>}M%f2p}%HQ1Z} zhq_mtVBmT*y#+%ys;F6R$Fs?xzHRuox%9ezKkPpp|7jXSf`OnQCIv+$z(#NS6AS@x z&83@_xW%)T=~W|T`lijaXy@0jyIH}2#ZsmgPgq0HRS}!uZ}4mjU)s8eH~q=fa(5^S z+1rT4#B(jdyYF=%J>L%Mf5FH8&9t6a{~gqjU<0`KPK}fUOp&(Zc>!i`-FHwVI-9XA z-Yb%vl?T+|b%z4c-aE+=d=3P*)uFaJ$&zO*dCa4oByT!N+RI5w;{z+xllFYEEW6-~ z7`zqu-b1E%?39b&m~wFkPPuTq6guK*^*>L)ULg#;y88F)rp51Hj<7F^R8D|SFU}(4Wl$d@r3o?~WQ^=vFXb*)I*##7W6B}tK3$D(8 zO9=H?Q@9Ow)C})1Q#kHszkBqY^5&fO+T@tmc~q0!E<--A57v}g5T_%~iQrLW%}kcu z_W0s|!w<7bZ3l?CnlNLHiZmT>jXx*tsAz|56a+!-L<15ue}TLGkahQ)SoZEw>;7Ug zp4~Lni8UDT90ppYn76VFi$SUYZ43>bn{LsA^bgrc-&PZgmmcup$EsTKfecY|dtK(N^moJZ* zuu0`E(+MiJf2qhJ#7v0Uf;xtl5n{5QT^m0O>wf0B{g$^ZyV&wlLY}hCdA7*1MpKUo z4}SgS-S27u6+8Rlbs@g(SdfA*ojl}V!EHKN>DIwKbs4An$TpjVaLyx2ZKBnHH2^M! z_%V3Yl?v3JX_sa1nHI5gPOW5-yx6pFRIrHvt*_=ze-IEiDK>UYv3{9 zG6dTl44X=TdRcQXTeQwf$(AU)p>h7ptJCA3&!7GF?1ITf*CBg)sDfZPC_`u^-XjTt zaf3u;Het6)nye5>RU{wzBl2_v&;~x@|1x!_9*EtFzAABOot9=aM*;vr_DH^VtM_F% z%Fr+Pf6!%lhHL~Ifhe^)epj8zx~QoQt;@Po$3190Zc=s7A_f+1&rGH%zsI;ZiHZ8kc9b3Nje8+>;@)X9a!!Xv&S5@teJRqkY2nu?c3P2G)B(h4F*WIHG#g=S zVQ+H@-y5E5%D{!2%^p??`erO+n}{vQWjOdaf2}9HR%tlQS3IB@5gPj#Z1IQ-6Jh@; z#sj(~ZI2J@|Ihk=zai*O7vz>U_{Hjgew#=U0hv~7oNpD7rFG9udn7(-Geb#8X5zN( zcy-)j_4S(9EIn|mfjKD>{k|FfDkZGbfW8)_Eu;YQfE9%g%AYcAc)#RsH+PY4!#lD8 ze};@PT7qo?cJ+)$#FT8vX>UiP%Pn5P~1sc zc~SJ4ze(yk0}*V2h#X;Bt6v5s#MQK;Hqk)yb$>FfJB;gtrW^VRNwHo(mho6-{TQaW zZ*%C!{_pQ+pFp8yF#a$aPKSTg>*!p(d}_s|%d+>i60)*wYu+NO)vttwIK)y!Mp?RN+s9H{}KemsI*mGpZAkR=jOa{O?h$Q`0GW!rqT zBw`hl&`dvZQ{VVj;pXMIKl*T|#b2+qplFw2Z4MKc;9(3Gm$G33DwiQ*3_mme`T_9C zy9uBC@Zh_v2VXsQ*+g*WxfheGw-d)%X6?i|&A6A7tL5b7RM^?;tG|8q*sHgbVhq{= ze*pCg_!0Q|Q}~~U|7QbSkRZXvr*1TthhfOe?(70*1_sueS*{X+p&>kzsyhI!=7*XI zV)|W+-Ip!q{Lu%M>PAZ*q!BSlA@hae<=e=A;bE~}C#5gZ|@mlwyXDyw(b`I=NOua;PF z)9;H_UZEC#1T?F003b=cKBv9`34q#~fE8wKGdX%bhwIHUE{w1PD4zX(Ml}&KuoNL9 zkda<9=Mwliq7Qyos)*50pmk@88RKpjfUbQh7R7$wkPoZ|+RJjk-RuK% zvG$Di5Q=Z|eZOg9xa6RJqXjRpe-p^$Gs#JKibW)tK9Q4yf5m#cO0~<#e#N{qqnG)f zY%gIfkEk55D$4D-GkPX=kH0`dvz3FuYpI$&n*Uq+}B8wC= zwEX_234{PTtK*i%rUF1a9E*Bdk*sU?u>feI8zK#j4x8p`%6~RZ9Jm}jw#b6TVm84{ zIXEEr1u$N#SqeY;1Qj$HBxrc6l??mBQy5*3*q)ed`1Z0Mf6BC)SY~S$d^0o`lr5QB zR70bH@0FR<;`Pw(fYpM~ytSOhO*F8z*=V^7CIWU8H*1reEdk~ZT5;)yb-MNGIJm=} zS)4;}iCMRaHZ`xmtv91Bfl#bD`F>cAfr##jahHf=nX^42T~@+!Gue_^!ARvUu4e*# z05sg>M*&=CfBoSUBooH1_UvP z2uQ`HCK$qI9$W@`qBuZs$78MlYGQy8%FFm%MuDEIlH+*E_&BOsKJHT-=X0yi5y zlLndI52>XA)%I=tlDFUpZ^ZSV!$26amDGaKLY~G2QDAkTrBfK+T5d1N0W~#unj{A} z7~a|Oe^QWQS`{MU4pfxY0?CRH>kTTW1?8%_`(bFcfbGV#SF76*F1|(IYKXq|K>w3w zSJURXzLo6E!fCR)?(WSK=1v!*gU#nO|B!Xj5f8D@4ri!l8J~-MwK%SdxB!MgdB5&a21DPodBo7n}5^T+TUCiZF-7BLsm2%h=9#? zrUzAY$^t4kUP3<{NxkE_nqtF?0%Ww9f>a)mC-z+qLK6yu|D?-D{#TdRJs31?TmH=; z6->5_mf+JJ#~#2~tNMP5?8>>fA!RP0VQBW-u-hjgRGDL9O{U!AO&1hwT;m(d;U%_Q z_s4^La(}_40m*bGipsx+C}luNc9`{caAy?gsp~Q;9pc^)2JX~}!qBxM)|@i2FDW+Q zJ@95+*X6GSi#;ekJ!c9cNGCe#0XmmjVh_-YV4@1t#88mVlvqtDFg#!{5 z&Jp7I3Sct)exIb$PlI)jt{!}NAYz4ZC>m;_Kz}wYVDDlb`gmB11a;6M+}#zEw-5e) z%K*0?^^1%{SSNtA3Cs*j`T6O0gfw4MYyyxg#O>JvH)aNQ$}JNQP+ zn!KD_X{U$H@j$6Fm2zmZ?+Ifr%=nYDf`eiTMJw)h&H-dhtva1-$)JscSfmM|?uP+= zQ-6(F(pIZgl*a9*T*maE%J6`&>-12fQOnvfp>mg(-hsNC)1r#j3Hm7t0mMNeGLTG= zFK}F=LSWwAo6pn;fP2sg2edP&%laBYa;;1_WWVRy+j5<*D$s3~5h5;~o<7ZKCn)_^ zEyyRpnq}Oh+K(CnUXghwn~0>b2mP4DHh(FkR^AgF7V(ABo`jGx9hfM|xKQKIW11k5 zsL=Hp4cpG$XezCfi}Bg~*2pP)EnB`kzENBZg>GC76`i;kF9>*jUnGigF%DDi7co^1 zmbr&%cU@LGBns%X3`_I z?cD0~Vc}Ik7%!&a*^m0aNAn4f<|KdEW9GwTkysWsjZfy#Ig3E}*Hj1lhX>@8n)-6vkC$?!Luk#K9TH!9 zwbtFa*z^C^VtYQ6mdrpyz-w-N+d(ncLykEQ#~>*KK$f=b-Tmv}b7-+i2Tcnol`-p+ zt_+?)f59P-H^v1}5}+K> z9|xBeTxpHqgNGBR`Qr}6&!9X4qPrUb>NdtC@dh>)x9QsN!%id$(BSVn;c<7l?54hC zxGd;6qAKaU7?9W7v6rWcbQZq7rn&CpJ7$^J+Y};=jt5?(J5b>ojl(0P;D$OblQ#xY zjJU_m$3+>u(-_n7eyq;zRdN(mmeOfa*;} z&w)!=SbdD2DY&KLOibyb3Q}Ef@hnO6X*&1^Qh0d`3*E~$P(djKq=5v&gduzqzUTaQ z?@>u9roV>s2q+wi#dSqN3cezv{!PVq(UjoL#_=NgSy!!SL*RH>6iOds1)VX)d_=fB zgw>SCheGN|>^anAswiF#Rr?;3iMLO@7=I@+&ZweiH}u7tCG2t0y&fN*08#$|Ae7N`3lc7u~f8ASIliM~DexF~#H%qk^j2oc6yH!rr#GAxN#WN2{*?!+;?%j-mFm*C(y$h47C$}z3*HvmF zY(#`>IYcDS{hL~BOp{b-|FS6ONR%OqesMD7*NbKDTNX}$a8rpVJl?L$g(J;`%yJ$w(&R12SwSiy4mIxv*C5V4;^cje{fH4siJ;(O^LDNnf|oT zcE!5%R5;N<2o&9^FQdAxzcVH+0a74xO^op0zpT2Nc2@%{!74?sSUVsnh!&#il~r9- zyNe=H)jQ@%FBM6Y)K!@x_Rp%#5S5BAr}wy59-2_j5(Z<~fHSODw6LRumIo^oNzD>zRYz(a zGAK5C_X^E~B$gTAof5$VQ*=C2+6?maa5PUbTr=1Q67-Y7g^Na&!yXCXLj{qbvTy@q zroPwnDQv>Y1J?s~g1Q6JOblt!W~{b+1y2De^8GI0e~cwI?a*d)5H5F{Kw-XKttvF} zOls6#HX3+4kYNVkLI55EOw1&R2qx;j{cQ~ zFxGSve?*D&fG^YPp$OG=57sXR0b0rt*l5}YgtPO@vrlKZAT(i#po$pSXv!MkT_*F@ z*_(GXS-?wx9ikaaes^70i2!Kf_}wzOIQ!YVSlvaes2L4v9N=^@0N+NYr)kBf`&tu_ zR$6P|A49q&y>98w!kE}DQP8SZx3+G@1OZ+Ie_<+s%Y(18{qP7@fs^jh5Nc`^{=n0_u6M z%)iOEJAZSPZ9O6jKxC~(qkW*WtvtGAXaE-*9M{HS&5C7g5iUOD;4dy71}_!1tI ze{VbF>~VBK)Hi}}epDIcUjg4{9DJM9<2N6N-|QRUw~l9L074jug9c5`hs6qe5%zEw zZX-DAU&5vfxIhh#wxG5>t+esLp7hfYy+8{%XF!{4(SR1PR2@JJs2AZm_*s5)`q41=idf8RG95EYa#W@8k!M^N-Rg!QntllN0WC+l+V z(FR)(zeGw~p{wxeir~>gT&cK;L)W%PYvRz=VKh=G9~7(Gbx2`^wom(fxeVj2a=_To zzIc$me@iJZi!i7~O*Et9Wltq9>%kZK{tth*zM8Vh$9%WNER6@}ncy0tF&45De?i=< ztO#NMJl|E0eJm~q_v<29n2=dusWWLK%?su+Ha zaL`GaX&{9V_&45}e-6x#e*uDho5lUx=`~gLcK^q~)I}qpzh1BJzwX^^e=5XH+$J3S z`zlZB-O052z|@y(|*MX^){ zzpQm{-0!2P^fqqX^gXCS-J2xzfKvr!dZF!FgQ(= z({~U37eCEjtkKo{jQNGUEYUf`c9+hOLQM97bhf-+S8m@2%ot-MxX-t@8!&IhcAH&? z#{I5X7Cr|eIrX96NawDq;iWLUG`45H>ZmyL#WD2l3v>j_+%*Hoe>?yB8q+14_5QYc z;VOIhrmifqR6;y{M^&s7TI5?_Y4uBk*-x6>B-Pzdy=`%_6lR238IDQ`OF3WpZ!Cpy zAjKjE2X;1j^8W1Sfe~8e=~yQO=Y)b%awCXPA(Vab{5!wakE?8v3n;QGxwJ1^)r5R= zRn--<%CTMsCA$6Kf74If{Epd&nl&>3SAs(X%Zxr0tL5A!N~(9~5cTb<$dLlj@s63R zmJ3iq8gr5o9oC5`p>-zXmtHsq6ych6OMbM|g{nI0em`VnxahI{cjb)P03b^%;&dtq z$V?tTe%uew%n-}$iTUcXZ=wf7U-!Yd0fAMcp+*T|07JIZRh*>vRD*(U_;AjLa8RIS-01HGXf?+qXg+x4LYYSh)xkU>t4ija=z;oq=T(` z*r)yK5-RIos4jx2#wW@8Q+^nCj8hkRqPlRy560G=u~lCpzN;>sJp7QlG^Q{*fiV!2 z3c_tCteN)_f3~o?wIdY@nnH<5xn5nl#EgCnuDeC6MZw3nd!rFF%?_D1`KVX|&gA4(voJ-#3)%ixq zt*a2B!uY~WaF4PE{h-ekcx1&UJfM$Tki7OL4war;@D!wqE#Wc#b@bRX?s1ug|0o?a7a&rr{=FWGkY7~xEs@8{3e=ZV-&|(rzZoummUKU6gI7mInK6y z(gq_3NvT`8+NW-^^BVtTl`SOag}D z**XT1q-WsAf#_c}QXd!n|2g12@Z-^c%e#C!U~071_fF_V^;a;qQT??XsK0VdTpz2y zf*5V*08PRtud)>xcajDj$|BV(tj%e$0jOzqez&wF+H_@=*htd0Eg+OY9g# zO0iXI=KZHssPp05MQNWFW3Sh?Q8HzF+!*0e6@w!4o!$ zb+ZpyeL8u*-}xiacjpk^K>r%a4{}b%e}r@n-rV@h>L7oGtu!i=*BQ^Aoqc+B_V}sunc%G8g~+a6XO1&(ZRVV2 zypY+|B72z&fBy5;zaKv>RDYx_6ysh4p)lyPe|_@Je-~HZU7pXR^SNZ_Gh=N2+1FoP zUH*7=i3HAgoc8fk!EK+^ObE>i?X#IF7&V4 zNiMVAu#ah&k^NCZFky8j3@e19ao?n7^~?*HRdvj*aTu&-=9pLU*|OuL`&2RspYj0{ zg1f2(QeoI{DK0gWS`B2ge0zjkp>$t3@6+V`nB4hH7@psR z-MqXydv_*)Z=Q)vI?23LnG+xnQDpPg*~_1Kwt$x)C1!=lehoucnbyo`3zv2F{n@vq zgTq|`1!4wd2;0JC@mJawJKUC`=?N=Y;Z;^Je-RDuCvLc8LcnjDftmCL7>)qLf!&h& z0WS0rt3ifi|c=qbI$9 zPr6MWY$2B-1Fbb)-?!BQ98)10b%=)WI!cBQMoOk+t6FITgl$o+ZXGo~Jye?1WQGYv zf170+mEk-&WxkmE9dRv>px9DKL#94w6Y-(@cZH&$?kBAK|^| zOmVebrzR9O%-CmNwCs1oU@G5k&QDzGv?O|_|29d*y3?r4)1u_oB}Ms{(iS;MTXafW z+z=)>!yJc&O7Lbr2JChLX1G*jFD~Gtf3!)fz*QWx*@I2{{7rQW2u9_) zc_1?^{1v@xAO1cbpsQ)~W-n(TF`mC@*H+_kF`RAP+I%dJyrPasgw#VlF zBwT+L{yw{h0mH0q8F-eMreMBkMFv!b%zp^VpQQPs)4RG=S1cTTNQ;}*0JZ2%M% zc3*&G6@h5V0ye7^%%qgh7h#(;pcx%>kIUgc;m$}AcBXw-&2KTX=F5GgQJ#x@y`!6# z|2fw@FW1+!l4?)CWz8;P(gC6T=L;HK2QFN`+7KVYgUdiK+G8XtYm$+=e_2IChgJiQ zDk$BN)wmH6U^ddZ4qH;Ki5hO>vllOdki+Y0x2HD;La?wNNWeT~L;}Hjbc@YmB)}Q~ z@TaUrI+t&6H?PW9%X+!n$0zg6YP;D*djj7ITRe5Ah`^#DT_0d|huaj))kFZ4on!LF$p~4 z*I&-1$;+sd<*ReagJMY~Kp}w4q24bS%M~!X+^pACT~|jU4OoHjcS86mV+~Gvh`ISu z81yBHiQ6UwO*!d|%i$&)8Xf8sZ$n*KptG{l69 zlwdWp;JkaIu2WGTqGHi*5Lz;-!Jae;uHJ#GIy|b@``e&U7IsUz$*!V8<-Q@e=!N5S{kN;&w^ZYrWItjk&A5-J0c$J2RDAqeB2mmk5$FHRtavB ziVW;}*-H-6U=|BravW{%I0eY1t-4dD=&QH@yFtp}X?4uRzf2|UGZ+W)Z${jOaX+T$ zapvJe&=S@p-EUL%6U>102h>l@Xm!S7!t9J|1R+$4E6X$xf15p+D&Ypoge51m!A#Ny z|A%SA&`&vkWO;7Pks2m&jpB@INdTF(wEZQb6*a>t%}`i4QWn+_rT=-bTBBq>ul9Qi zj20lG@~0Z^MLJk;)_Kq`lYE0-`8J*ee}(K|(8b^NX7%beUN7Qri|{ZW@Ow~^OMnd| z&;uKIdsr{ue;t~4i&i4wlrJJQ1KBzygYR?7+#{5ja8%c2T>(fyh{si4Znp%%GY@;B z$pMRd9B@NsN8b}@gSZ3tfGj95pHUr_XK%CS>dWM_VcAxJ<1AdJA)$7#SMyc)24I_p zg~9p+z|7k!dwurpkqM4A5+o>=^cF}HY@yWc`KRR4f1hm27$0yoF|h6KSxDD87+6;U zYOZ@~exmmn^*+W3gitxxd0mp|TLitQpoRq%dqNOKS|OtVT9i>JAR*;s5=};dbO3L( z*wrx`KvoaCBG1zJ{Oyc#23QxLo->3%ZhWU~@qTf@-(Fq?fx>EaS}tNJufpG$S4l`; z)u~dzfA4pPo7PnU#&FV4WhIKI5z~CLUMwT#fLm|@%|Zdu#mOR;!BkmK$6gJVp{?>* zfgAxRbOS6I+zWuFSRq9tYW%?Q@b3S7;oVFP?=Btw{Jq7*kCRbykSC?XqeW6?6drXS zhM7^ZBM~l7j^<-`6Iu`5sM~5wxr5~Qo5LuQf9)zVPb)~9;g=GI8ZvUaUv176&;N@F zYz2ubJq#5TE8CA&Kr!5?&W%n7^oVIAcOIr@Mt#Y_h)fh(IXuVeo*$(2bXLN`j& ze@BfLqkHPkZQ2TzPD)}7dK--G?#Y&Mm8~EcZPPdBhUc50lTKi%D?Hs*hnlY1vJsXg z)c_Zzltv8dTe4SUL-O||&mcoUZhkWGx*+=N&0$BhOEjcFcSCt@G$QzYjfS>a&ztl$ z*azXp0t=Kby-uO?+4?Ne@r@gpBz3qV({LGzV}EQJY1@N z-q~6FsQry5|6x^#0U;1pQD|@`syPoJ5~oBEgo4IMZ9;$Xo>L-N)C~=d-YJpMf0QjX z+^9nILU~efpwHfQOI6~?R$@;hArL!6M+%$U!l>S{6Poczt2v3Gz(;mX8=3aCu|eKW zb(7qoQI();t#2nI4 z{ePFHU_rd+CpbT16umxwAB&8P!mEi!0YHOCIbJ0-6=-3>wi`Ck)-f0vj^oo45tp!{ z0kJ=gVBKB84)q=dD+g--w`4jYAx!D9(uc&xr0RYOIE=3PIbavfG7&4?R4k|srfEl zc7>qr^a+Kb2T`^hqU_L{I+jzw?yKY}RX09rSJ*z;Z2-q%D>jGyu^Os%Vstpk-HH9C zRX&2$kiB?08Sutd4{cm1e}Uo^_N7RPG@y_mxE_>6+^roz?mc#xLt9Kol#h=A=vJ$? zgQ^J^hc}DOn({n>O(H)Xwgwdf0v#H!=Er5$3goMF!JJ7q9Y;d0yM}ev0serHcX@DgGOY0Q8|m z(^itwGLHZVf5Wc$z`l}GkN{(Fb|Juz+DcS~!M4D1A4?;2=m9QX#nKS^s|u?$@3){! zyS5w^5?;k*>Ui+o)vO( zD2dK0uyY`CO^MSMj0=BKU7U8Wf$^I0Q51^q6u@8X^N9(q>87FRMZT>L4U+e{jt^qFTl#hnH<o?> zW(zx6XxrEZgiLCBu->=ZBlL+8{(^W7&<58YC1?P%NE<~mm!J{DVI@-3_KR#U@(~Z<^qsTZYFf{{NnNT#TU;~F_)u? z3`KvoO;s{M43#jbBgj(hbU_0flUenACSfe;J7v3lzMAjzbqPZ#l`J#7%A}Zw|MT5` zJDo9=J7Xh zlij-9dl7)li+txK#h4_TvSBHjSc?8FSb0eVEcpzknhNb7Zq|pbXYNpD{${;#{Z)TT zq%dv}c4Dc*vurzMG^un5s4N4u8&Av^O7Q_v`T!i@H9;oWC7Ab^=^=~RDHfe@9h;Nb7mndZsA;Fq~yp8T@`k%n=uh~=Vs zY33Mv5;4}bfSCbxff>%g0+IufI%Z5`gT<>?)g=yYNqO*zGGpmp8ugCz!#r zCs%M`z5!6l-yBMRofkLj@Y4RnD)WuYa9#A%q+)&W_vJc}*&R06PtfFZFerb?VIo$I zhNa4$Qet!(>?tg6G2h-~s|>AHD%Q4IqlwgP{21^)Pu~1^@XG8E5U%rrJj6SbKv(vb zU^kJx4BbI~?QuivV!d+Gg7#ImL%nv*%g~1CJ}(!c6*1rUoO#|rg}GEQYIq`w12P{L zMZV0lvJBP&ttNSW*lgB-T|0mLS0~E{^LOj*(!cTPkonfG%fN@r2i{LuNnl>1g|+xF zFJZ<7j);Au!Gt)2vs61<|q|@+Zn=KWDm@NL8*^@3vVSW(D`R3m}Y5FQ-h?4E}u|a|phP^dqTtI*l=s3RC`eU^QEj4C-Mo zIZ#6(ns}!VjSLt^tvO!~V6AGD`5(RPfZos^Q1ajD`Tv|85nki-EBe$A=0h2>$P%5Lm@DN7i;-K*hw2+ zO*{?M?WENU0^$B4Qv!dORSo(1jGHwbZHlIea+L9y zs#j;wKzUoD*8|Ig8TG8y?{YxATC20D@JAJG>_8Z$khqz)G!VS2?%>B$11JMm3z!qs z-iJ95SRiU(C}W4C9D6FtNnv93*q8nM5z%0#b8JHRI+GA0S7^sDLkd5!F;os zgNjO;?6cBm-tNJ{Uq8-5m5=-MYg`hWt-LhR6{q$NnR9;zs3GQ3cS~GMCk_(EwnZEC zP$w;MaI5!t}}lXhQyiQ*Ou8Riy}&07G`hu|AE%GUk^VS{9iP zbg_SJr~<%1)GU>EL-<%YM*vR-T%zm=ZDY{lE?YegjXK$=<+lr@2aT&cyOV(s<7$i> zi$;DpC<=ZQpLg+_$@Dg&cynG3pU+f5%($|P8J&BK9B{boBI5zimR-2HZP2n9s2ot} z(CYTKmqi~9c5^W^D39LSL+UZ_T*V+-?7U2m5CoG(-;&T`K=imozt6ld9kM4&DtDPz z**mYOq%u6Ox(olU)~jF45Xag&0{GMTS)B-#f>Ri%<3O}Ro!5_|m2>boD)w;OkWJ^8 zC6Wv)0fm=Qk_;Pv)l_@c`uz-~M%o_nH^Q*=pkdvK`>q*wmr8^>GLGeHMMNA;IvVNdz56ZzS)Z@AcRrU{jdiUQi~$%)HGX%$ zqS2Ft^_Db*vMH>5sx{%g&|s2dvXV(>Gr#AVEjJ{R(1v37OnxdQiJyI(pd$;pK*LPIb;W` z@Q4q8C?z-4Z9R6dN`kpmScOi*orP#&6*e>@*2KWAaiyc$JP5DphzBTJX>=pEHvf># zG&>9In02nelB~PnIfoM{%LI4F22a3jO^gv!@7r)6Q2-;>?YU~M$Kh|gpyOf9+JHA7 z`mk5{V{^C0UUOU(1MJmegbmo=e-rO~upCE!I3G?-bL9Rg%7axN{S3O){q74db|h^7 z@3u#8E!j>?;v@?VjMu8hdoc(+qdWVojT(j^172Tw+e0O}9|Tpzqy{Kbry)q)9)gZH zH*s=!P?&QO1tRHj$_`O}NZ4Ou0wsj)p2c;oX|<;VIY{QO`|bMhd5Y&m{l-pO9k3FA zHthOB-$U;A6w|QjXCbNG@3V?hI9Y_xeQJZv%WB9rBU6^dyaX6)SL(uXB_tzx(%N~# z2Ie_{2gQTD+-F6R-(-G+4%+Izm?154cp6^H0<~RDw{`B9V~(Wt=wimsQwVYb>WeIl zJI~AT@ac=oZ^8szK>)kWeK5)tZUE|k(e;WVUvGV>0M}a)#nX7Ztf&OK&F~=cyj+F? zhcA@~cT#&dN3s5bM|s`Z>Uu5HWU=0c^|`4Yd3n*g%Q6T`x=?@gHa}=aMbR?}BtyNl zI#kS(-L87nsLW=w#hia1K8)ZJfPo)xKi8+*=8LIh2DmYFIX2L;( z`+O<@5AmGknC}GqJ-zqW0*`yX7SK3;bA$sw5ctl0Ebz6vf8lEvRxv)>Y*JU zf^8Fj;_kn3zpp^F1495b+Qu1QIRRX{-^Z}9+_tO+k9sym%<~`Ro!udr`?M zTl%$pJjaUqu5Vn3d2kZlqkhAG-$HP}P5gc$`WWJC|M`Oz$(!eg&3@j0uHX7|{n$|^ zSC^0d8~#%WXbBY^3zpKUBA(pjuL06gtG)-Y{xAzRZMJJ~GNEFW5+lTIjvhEF4*rqs z(Uv|w#Vvg|hjsX}fT!Gf;s|b5fO1HVN?$ys0^=vaxd{)$01{w7sq~n>UAPY~ z9{QISS#ii`i@fTz@fZgZcIo@*dLd8@BRW}8C#wY@-eK3~5*P|X_J!WYCc_m%hyMam zo95}4VQmf+m$a7*6#_6hlc7u~f2~?ucjGn=e$TJ)Tby$yEbo-MmviRUZMW0w*qgye{=Km>#J)W zvZWSR*UtIn%OAeKdHtW8*GS+(^0LQkUMEu5Rtsf`-dzcla=pC6R~K3fm-~ejS~63= zXm8`twGvu|#r0sUTvct=Y=}S`XIUdMX&xuw#{@p@%S~66b)C2JKjQ5bwk-GsX*y=ws7il)q!|7<&LJ9poI-N%=IESo|)Ld zj!Y((MVve-CFg|8@+5*^4|9UZgN#$G$gDvwya^*&Iyo|15o}lte`e`vwI~5(dA)C^ zl%kB#h73(j37SR0V+C!I1HQ(u`lE)SHUeW8G|vk%FE?tbNdl3`NI*#JEpTPK3CVo7xC`*U*4~@2|_sJKa)7Qdeer!Vp_$~u3YP)O5EJ>np zux&OnnW;mWOlDmY5z;7nZ|)eg1|9V!J3uRf(z9(6tgxw_qiJ|n#CyR?HCj&Naevn_jM1Af!^vB3^cO3Pv< zt-U6KNCp%6d!s(XgDGZ!v2SQc>+nkb+`Ov7!bdtGXk3 z0rOqO?6bqo0vr$hxU&vN&N>cQm5bnyeAiGJ5X@vge>)hmE-VNqIzI~@B|d?We!hQgoBbUgNhGl1 z=h(eue_Cjp%{8OwULvu}=wPtPegyDL2ztKfe*pYg)ir3WxM!kw{JeR9)73T`LC6_G zdUOx)r@UM207xHo{hV;M61j7la5Wg6LfjY~XpfApDH6heb<(E)UyShjyww-Z3ZL5r zdUbcY`yAl=^Hyp4Dz6E5`d?B&(9f7i>e9Y|3Ak6Z%5or zEabe%GJRMF1^yOU&7z+>ghsrT%Xe?8%`;9GrPlACQFncL zKLVYJJA7SiySb7T^;6c zAUR+pP!A+DaKH}ITXhb;qeu?0&8sku#V7qeh;o_%9$Uj(4k+_jQ4e87 zsbiwd#}u6WQ_CMIga|7%N`v>uHzl%yzqqDK{kI)e=&Rxqumjd*^nOPnxp;gFWQ?3``lfBT$k!5U@T z4vxrfycw)Iq|yR(*9bDJKJ8L5nWaF5am1^!`BNoAldXXhr=%R0+{U*&cuZ~1hy}e` z=a&adE??rhiJ!-`=s7v}1|3sLJp)NA0F~j4;8mvWb~D#E%NKA+<1xk`R9_lW;Pfq6 zI1gsF&Z*ee+cjpKN00(Ne<|1X!YD}OBKJKCeJ9f-H=L9Une}u2Hc1CEa@*IR)%!|CzcDcHB-10)9f$alp6A zOl}m`%*PLXmT)_&46Vv-X<;>3r}d8w~`6uU9_d&H695IC|z z5JeA~H0XH1S{Y})d?-G}3VELDQ`ElWKE%z?aM2UOZ>rkjghyN_4B?7Aso8YYq8k99 zGxj48xkP63*%%)zxFv#ZPrBMvT~(ye0O-%!e71!#HP#`te}l;U5qQL^7qb46W}A`w zfewqeAe;M=7_wZpv^|S3UeqhD-D6SroLjZ>P$SFpGm3t=?7_?IsDsPGmNPc*g%9(i zXu6VM8e@csJzw&^D|#0DbnG&b=W6V=)(stAiy-H6P-q1#l1X=Vp%tN61o;Ts*pRW! z2uDdHvT(*te=)kocruVk%})}WgIGLYf$hHT3f?FMgvcs$w_mLYJ3>G`a*R)%gLD@h zcUw1Thmp^MoL*bGZ;;r3m?1`AMXdm||vst0-oGsVe7Z zQ@gK<4YPD$dVCL<38#-7a?EjnamGN}id>sNQyvJj4RA;BEn|?11|6@Suyw~hyL1=h zv;6F@1DIAUPth{mQ|t*z#5RkDuD-G7IXza4e~mO4+9q8Q*NIgwD&J)l} ze;DN8pwUu^n;H$7rND~KWL_(}$+9eo|0ICYX<2(~j z8DQ_+9AgqE$MdWmd>kbY+8R#E4mW=4e=y^~7B~eE_4Q~yCsn`TVJ0ZWtYxsGJCnq$#h!@H-8!(1F3O^XcZ?M=pe<$S( zKG@~3Gab^!!L%*qnZp}x`Ny{8=7>?vv!%g$f612Qp*FUkUJ&;;)0T!p_&?gxG1sR; zxWh1T*nmIrrWwd13v34j^nyvJXZ)e4&V}ifpeG%4HgxPy6kxMIF2K5R0Tu!kxG#^l z%YeEduQ)N#;U}%5k7L_vlRlz7f6A;ZrZ)x=na%aVYOo-eG)mgeFk}2&mkBg|EW7bnPCMNLRW{%# z%m>G9x-dZs1kRyBmmp;N*KO8s(4as#U_U4*c;M-yS4DS|s?H zKA=k0(POg1$3pys2=PA+!m4kVVQmf+m%OD6em<2~3_*JE$m%`wUJ@$ioR;8;)jRfs zlSC;>V=TYg3Jtv8Y_D!bNv`6{I_FQ{*4vv+effQPD<0~viv;|m{`OneS10XdVQOemM1d=5z^`*fq?D2p?-dHT5V zag;i@@!m#prZ-QA%`aQ+cE3IS{j2XYbC(T$4sUT}#%l#jb$ad`DE!j3z$GwN}SL7%jU#AF8&Ip-=7U zRPIlC@u_OcD~NfM*Pr`RAk`&5hi<3hZ9W0Zt6KKfMb5+KDi>u&b{EmpdUcc~p4O{g zS+9Qz>m3hO-V8vWN!9jDtHM_M3_DTD3oQ*B1=8OPdDU@G3;=i}FN&^Z0zhywV-QWO z(?EiXX)un==DI)DYUoM^Q9tB|{5<5(Rb34%UrFSFGcT^i&)33W0!8&h)eI$XVKD$B z1!fv0+OLO616E~`cbV+CmmSZMt=9T2iU)r*+_q%SS6$mQCGM0=XpoBWM8H&{qD-6B z5(zv2L8<;2Sq3A72#O=f9jvo#`m#DQ!ncg@T#hV%4X;{d86chG6qel70^|w8$$T(98l4C%D*UIj(3SbViJ-|M{Of z)E2hSqQc<^x$q(bJ80c%X{f8VLxv!TI%hmZe$ERKt`rvF-RcDh@w^I@9$EvbBYpxr zOS5bSEjI2Y;__ESHS7+VEJG@S&%PZ+%XMPe>_G2UK%WL~eQEr$cWeZC(+;p80>T~nH~{4X>MwVc0_fZ^nVF7a zV^)jcq3XgqNp6U^j;UgzPC2!lMA(J7;R>$8M2XU?6`Gw7NLM^JUNe80bho5p zGF-+4Pfr_bV5cf#o6*6X8uowGBP^$hVH$dgQ`J140>-00F6$Or9GIfysr2|6=TrpR%=Xg{hr~wSJ6OphPqaM;R!dHJv&$U_AX%@w) z`_im}k($h^0vD&&Gj%M)80rjMa7&&6pnc69HOso*LO=;J+Hvd+16_%+X5JhGDMYoR zpKFX{fO^im1)hbr*`RG;Z5|-l>2TxwK0k&XF$&TLGnbU5SR^o(+56a&rLq~=8oKt( zEl3dBqsW?#lGML!&K7?rshLs~gi{A32OyO7$bvi9cM~bLHG`pa^~b=)KI8}hL$Zo% zvQBM?G7MyHYPCPPtP5qDl{p171&QEHXhqPR$0Q1;0UxbO)>=)H3HZ35a!7TD%on7( zAz**YpeT;PXH}vyxD;daE_4QjLs@)k+PdY?7JDdGgW2^W#KnKH<^endH_y4ue-`+D zThzdVT@r8EaqK}%RbuC&jGn;Q7}R1UU4bUbqzRS@-O3!zu-AE#6N%jOABdiwmpWZO6&FUTq9mRGIo8i$}9w?=yI?e7&r`j&0lS*zP##*qVGZfBjQawO7?X*=Oru-_LWeYpvknTbuqX|FP51WL`ykuf+Gb z;+W6kD;Q1;DwiB8kP$wEq=Ia_A{ou(dzT6EO)?{A+)D}rxSY(ax~Z`$F1^1`9q{xL z40yaSs}54v%G>gspm>7yuBQ`~P9rT&P3R!tipI3^BZy&AaD8DD)wz$TjH#k{ElH}@ z6;{YdE$aR?Zg#h^KWZI6v^ zWwNcW50?N1j3|kS1%bw}w2>x@M{*z>N@ic6jiSz|5qm1Ef-0YPEQ_k#q0HM64&F=O z%H~Ko?O*+hM@EMdgT7Y|Vf{1E!v`)3tcs`R7D9%(>xiE$oT_Ew!1ux%dV3qo2z~l8 z2|2V_{z<>HP-D8x_OObR9W!*lEN1Nok;E3$VYj*t*g*Ge?&;F?HU4dtGGhIpks^(% zPuq9vZGFnyj4!GCv{iw>&_^=VTpz96N=@>0Tj0k?0dqoQDoGH4oQ?anre&(Bu!M2} zBGuaKgptmPkEc8qhLvZeIHz&ADpmf>FG%*&O)Vse`!OmQ8t>Y=++z=|WEC%_$%lc( z(SQ;P5TS-jVwvbQkn7HbYDS`~82y(SowUqzc=pu@pO%~v>8E*EJw-ooxSzn~t+chp!aT4|SJ~%cus<`JWt{DNUT<}@vLyKv}H|LnsS_#gd z@}~<($3^--Y2LarzK|N4Gmid#!{lhlKV{Z%002W+p{QGiccQQ_hDXsq=Pprg&x(cJ97&1@UIZoB`5|0Y>fyQKH6vLA9WacnT zVUjQ64T6b<_Pn35HwzC|Mhs2HFe6lS*bwaPa!s$Iv z0A8oD8jQnS5pi3EB}#vdBS>7ABRfZZe~C**o*5mia~$|kS8|BE&7A+~q(A$8a(G2& zMT|Y(k^jIUbSAzPB$W??6A=KqmI+38=_@wG3TM>j6#blLGm+323b94|$JNKcKW+z* zOtw5*fS6t#&-K$NONvgJQGLHLSdWqK8E``<A#6VAMgr9`q8A$=HjsNA8L&EAJ@U1Q9k0$Lfxe;RAanTY|@XRoic4rDy6lL`9Vpz2+~Ydq%urqbn*} zKRG$vjkLLPjJzGxU7iq1>brgu031q^INwubIJcY{FJ6)V=>~kY$S>r;qtilegqu;I z1!Qdz-kwE#a(n8J(O#I0uUmxvm_7)}@9?C)@+T02Z#l-mJ03sN7E-M}OTgMoOM{EMbH zGYVN$dEkEkFlks)bT$2F*=}ZL;+r|$U`TW_sa72+6+}Ev4)iAgdy%*MdQ>-KnMQ2G z_o^&KgA^WZf!N+CAlAL&$U@Ci!&+C?Fl}Ktyr#@_+f7g(lRr@tO2N7CfY>l7oOPG1 zTaPFwDtw>ALMnw`{Fa1?SOvPhKwuC(he>9*z%i{PF`5Bu<{dti4N&Pg@&_?V=DVQ5$7_g0euqf;v!KP$yKNAM^lQ z3eb48{cAkM+?$d(FiJ%g-4Vm8}3XMAB#&jLgrq zUm0Jo=_$BIh)lTWmS9TXMET9RqS%fOVJNBx$Hx+LXh{ti)V>m7q;ebH{23a}kkX}# z{sWr&NFzlA%g$rB0UL3O1HN=I`{982P{@lS(0T{@B!%A@%&nPiRrIF2kWQcc#ifNS zrkYpYf9$TJIgtBKWxLMh#&;SlxBFyV6M-z}xxZ_0%tRNw#GIy%+aTDqnuZZ(;9bJ{ zroAhZGh3Ve!0%mF|&Dt?#uE$$z^kwO%tYW*(9 zU_BY0Hh0RxE*N$`%vKQmx_L2wxjZs}KMrAE6M;|l*1lK6B`Q_dKAAFfN$-61eDG?Y zw@T~kEqr{c^yu|}_k211n4kFTT;3#WUjz6Uir(L)qI@yS@9yY%$MNvIBbkqd4wUsj_lvFarkL$u z1f+ep#%-50$%69JOi6bubn_YCb>oR8EWRIM?=?U&3(OWkbR4Pvk2`js`d7uJXu=h- zV=5p)(JF#UA@P$STS>%3;NZ8)_LPfsaAw zYC^bQy&P~ZxZOhbAal zkZ1YGv#`rKIBB8^l`sH-&3mkHK=bd|hV>J{zYsa$0N9EloWT|#iKc`atCDS(Ujwak zCtNXmj^uU#8*Uw zN}?rhz>9QqBtw9vm;9MeUXptw^}KyhHq}xGzK3l#2YU{=ZHd*NKTY5l>aS%3-10*sJ|Z0JgHSH}MpxYbz8IJHXbYuG z551_O_FzSp+}ynxD~954|*lLf3t`3)(hP4dCm$Kw`d>p!h>-rUn`sz4Kr_JA>%iajO)`kzI z_I%{mLI3OVm;o+(pFL%$rrlGL4$ z=+g8qEA##CApCqX6#)3O8hhtK`CeGL-omm=np2-l0$<-0CWdWg>wP>`wbOAYS6!3m zIE=9U9px5cT&i;M!M@JM;q*9P79;eh|T{%qL^W?H#Y2WsT4V4)TY`o0X>5sk36|Jz9*IPW1Z+VtS> zZ>voNGTeQpfs{sdSG01Wt4_jXNqcJH6qkYv zv3NCtBA}|uEWCgOXWz|Nz0?k-*O`ud!QDZvJ=)xNTe&{o?wuby5Ef`^hSDn^t4w9u zy_tp2pRA8JyZa;?is|~!y#iR4E{k9@llk9KXRzM3#K#5K-{6p%$;>k5aeeJnKcOR= z!Y+V%c9?nOf*U8AhfXzINO_!s>ROrhT^_VE(?>?FFUT1ej|YQ0Jzl_#R$_b1Tu^}b zQ-fxSOMQx4NI(7ApefvwA*q+P^LJ(n)l7L z!`fk+&w$i}c>b=vg-Ev401fCG;{&D)|IMjYj_izLP|jI3gc)y2rqUK34tl2Lz;StX zBIpS63zW;01>UosG1lE2-!p*zqdYtH;ETUJNqGvZDm%$+ZMK5%nuD) z6%0pL#R7$EIj2rMwIJ%mzM=B4Ps4%+(A0$%xbZt<^9+%6978898_q0N-oKuw6+uJ( zhB0Y3b-Bpv@c6mL=$(&9u_g|e>6~0r6 zf#+B0G7Z|?+KQO>FcYk>+}BgJN0uR4!i9{XmK%Lm9->&QribnD&|odlQvZ1XbQ9lz56-9=@x1rm3+-g)zx%TJe8LP0gCR5@XmmUFj2C=SVploL*jy4jG zonV2`tZWb%^ffbgOQVD^1aLRk9+<=dEbu?GcqxF#7u9tlxNv43qPgiJ6wiN!!Rd~L zgQ+I?#pdVlHQnVxdC7YSH+m@g0-BM&{An|J!sx8s+$|mtC|vp^Xuz*vDRcZHR8HMS zMHWQOT9BR*K_&oDqkS+;&9pZ!)?UcQX2cxpPMDDpAvOYI2=sSzKked^;8=#OHitF~ z6`oodg4BzV7jXo7{nK9R#&vNb-c}kg3QzPGX<3LrmI^B7K4SPJier*(sjLB#In>ARx{K~yyP;H!tB2}MJX{tkfz?})q2RiCaVw>| zEn?dHrJH~#NVbD+lMOAG=d5bXXHLzoeM+eIsREmKfsPP!_?g zAceJ*aN$Y)MC5%!_G=$Jy>@D-xDgk6niuLM}(L1UTau`a$Ag1#?!EDzzOrf2i=-8q? z2U$c^=r}JO{<5Eh@edkJz)cch5RPoKqgLMSSA$`TlD7w|ZyGOLj^muGF!{FVY^*K$ zI4Z!CiEyik>kw|!7fGj&V>H7=nm%UrM}#hu7(G9#Q|NRszi4PJ_H|h@`gsLim>iZL zv6|m2x^uXvWo2^aAxi==IV?qb1{swunBDCPai-~4{b56VrkxdvvXl)Q43!g{E#vzl zy#qu|bI6~9ezTssA??F?8&tkJ&XNK-%x-{WZ)i`dtT9Dm-3@Bb%gNyEpaydG%hs^KWosCmFgc3_$1qy_7N2+e=3y`hWV8$<5Q7gX2WE3Oc8yE*3}XF zpt`wJsWEF`N-ZYy_jyI6>=E(8zfAQ(9g6jvRPC1J?5$>UC6a;^Cqrl45KI~2*A0N} zvMEN@?h7ohVf6;H3Ci=sU5MxqvA-==I#UTS-Me&+>C+Ql_4H}OO)G|*A74%Do*%+w z9-kE{zfDxeeB=1#?6c(uAz4*sjTRog%CCdn z_Pvq|h~=>j#WCdk1(!_|F1x`1eiw%aL5;|&+dPg&p_Rzr)?i^($QVj<6WsYhA=k+z zwNtaKo8GWY@cD@M%yQr8%ZG1vEIH1co1E!UEX?_$W<(#M`G|$*q|wb^IB=tonSfPH z#X3j!5~|I#l1mHQN}(3^b1`n)3nvVq%LI+2#cgf5XXro2LQ&zr)>8 zP00F@gb;R9UYl4k7NFC*G=C$9*zRsWh9m6n%wSyJyg91!kloKsE92!c#3i%+bKc@< z8)3hB`x8p!VKuiHnLqtyJr_um&Xyv{xTk|JU9}2*yuLo%aWU1%Ve|K<(IWiQ|9lM#-298TeLHd0?f3Ft`)e;e z&yLkXYm}K2HoJAh8}kShFwUIY$&orVFtXg)-0PJedX)^tIR5w5pFEvfH^7oa1?lFL zo`!*xkK|=Kcr^#3{0Bh@N+b*Q;p#^8z2O{wG76OM7s7>t1uQN~#)(LfHQi6etRJ}o zKJK|qO(I%;B7FLD%kb;dJpH)|+j-IQkr9SDNHH@K1gW;ITxS?i3w&}amBt=3CUptgpF!Q^g*CmV`t=-bnpHtx?o zJYDFA<<%gVar7ky*(?~z5tl&`xASslXSchOrG%NKouq?aXcMq0L?D!%LsW35fc<)7 z*Y`%BH`ARd1U0|;CDPH5O)>R+)4!vXJG$~@w8;qol5UCwbXZyNs0eiTH35uGS#k8a zkW+IZ51u^KsKJbv9G#yIIeiK$q>kJfTe55Z{|HB)Z>=9ewF}L1jXY{pE)U;j&!fRn zkw+3_J^Q4Bu+4wA@Vxnl8MB?^vXL`$e!+0R_&-lCh3)Rz`8{I2d+q5v{}q>nZipw& zK^`gwAZTXhle@!oH3|<7KuFPM)WZ~-#@{Znqq6%t`1hvo;5}399r_>iV+kWsKIyEn zM>$1IG37|)V`B%y!5_NaT&0W{)pQWK8?LRjcv)+kFa<$Ish}8&)gKHCm=CNc1;mj? zARgVy{ex(C2U4)8P6a z_Q+6!hv&H?0_J5))T{AiWQ243r6N50r3)46iDos-*#jTO%N{zrCU*O*GQrHE<>rp$ zzQpLJ4b>%InB*F9ST#lYJwa-oSyvZP?24BuXvN=OH;(rpgQ@^+1%Akgwg*noS79h ze~s@FZ}8tK)75II*H#uXsDGZ@KU*IGQr;WwFMs-2I92YY7;30B%PFwMAIo)*c`QU6 z)Y-jn3)BDCmgAwx13 z9oYr@mAG;@T+(s`Dq@G7Y+Ig?9t*$=-@)#DFEC ztAI&3G1gUg=CJ8(@rb5Ud7`0=rSE2dr#Rf@spWQqu`M&-rG)3P1BW6owE)geHGKSw z%X;1$B?<3ieZiQ>oQ)kDhy=ckc47D9%~|iOX%CD$Kde++%n@QrvrGSeR!aKr{;Sab zx}u@t%@|JD+4a)L~k% zA0_sMT4l#X)-8QSNo-R8=u;UlH7>LI%Y1>M=VubN`lCT?l>O6DLj#K^R7#(KWA_h% zCPlN5(A%}mGyiqgJXNixcTmKCXU5Qlwx#p(=vHwgy1lC`>|t_nh4pZI&?ps-BKRpO5YtF$=0Ie(44A z0e++=qNv9cd}|U|^mT((bZjl1rZo@5@zRUJ7t%$Q>Kyw-N?bUdZGE~NPhx)l&;s>d z0oH+(EqIU%UiW!L6FfcbpL$PLP)}o)vp+bhB^9eYZXFW^AS;l&adL*J_QAC)IMwTb z*9wQiQ;709W=x(-V!&CHiB05+pnJ4|skmZXNNcs*AfpXmH4EBin!Rk@9|?*iT>(Hdq z{*PBp56Q{}>n)~tp3tas8J!gI11lLIM>Hlma?uX6aYi+W0;YG?rdId!Hqa^>lM|E; zm>-QvK-}BbBe6Zl^0xcOyl>&Ry2cRjR)DSnYdHy(LFHP(xCoIz z!Mr?M(YOhBp&|0A+7E$GLWJLiC*RCBptLZd3^{$C6Q)hc%f+WtT0OVIf>v62x)n>O z!L*>b@)V%+>#&qdn-JJ8OqdR6%!qTNRes9R4M4DfCozUCbjcFZ+t{DPpQ)7rr!w<0 z)$z_`{Y;byx_NKKqA2_69|b+mmli}KvrBwO-x0Y*9RKf#pA|a6Xn<76EifqhV_>7u zTCVky*OqT%MF-6sz_YzObh=Dfp|Hza;gkyqWIpk)iRf&LIjx1&5~obHAj!*-ePJ=VdIo7xYld~I!sU1|UBLoN zS+%N*eJ2M7r4_etP=?jC^MO~CWos=EGoEDMCR5<9O4fu5xv?)%=LVOGOosPZv-`gQqRH$ft>KA*Fe_7es2D{Q=8;#*XRrNI>iX6(52( zSnEfa64Z?#J)=2gwq4|Gv5Z*|cXs4;f1XW;MDE$icp(}ulYU|Z#l+Rj;?v?13Q4h1mg69CXC=M}B6U((%pC0WGfDu4=7l2#V{u9V>$2d~2W+Y1TVGwYqwp z<_?(_)8t$jnUg(57Jx~bD&n<1JYs8tyM5WEUD1ptAS(w&l>X)DDGXG`{P7wKl#?&1 zS2YQytJp3Xn&eC^<3{UXqgl_WJi&Je(3myrD}h-xm2w4~4rv0RJ+2Iq?Wx2&8yGOd z1ntjtu@ngrwDHP13Z81kI}7I4)ms?STy;AGK{mJpz{^>5YH1=4%NIq(1Mfs(+ta}` zvjY$>lp7Xs{KK$54-bnqlz;RE@?G^aEyVEQ<<9yQZ-ht%+=RiPG3FZR=f_YcCLh)( z+|9EZ+zZ9u?FBCPis-u&GuSc$~Mnoo-R#M~0{izSTd&RXFE)c6b12#KS(+}0X zkA@@aO30f_>|NJv`5;9n&1*DabBf>Q`P^qKS!1o!O z1HI%zP8-3U0-#%aF^*$5}|wgD^~;jh;s zkgjly%<^i`ShGm2HWHEWujMY+x^y-JzTx=31p~-XT*F)%3Dro7Fwe0mPxZT6Z+`*a z^lFLbPD!pkueOJNRH`W3{3^&4uZ<=g7;1UOu5crfK7C9f$0iIa-uw5wT4;{N!mE1g zmqnLif-d$}6&JVh9+(w{s7XD|dDq^g-&jLf zv-DFZ>J+MbdSxt?j}=g#QC{+L+-Sfr<{y~1zdrmZCo@#2=DT>^LDH}=-Xf~(|oP2oS!CAY`CP0D_$fVC=s>hsIV6B__vP5kUh zb6-l0Z-aR^KNnwfUFh+5yszOk_|2cDn^!YOMFLVDd6dJ*Sjn;Y*WEPE!cU9Hd0tWh z)@#UOWT*j#L)76!hA#skaz_1{!elAN9q;qkk)ovF)fJbv4g0*y-;om5!$w5e*ECD= zp@2))ZGc&g@E{TtVMrgv=hzqknmZGHwrf3lY*``dG(3A3)1Z|DO{S(-U0n|voFb4K zkpMGj;ai-8%}8Vg_x8E|1d`YZWx!t=*AFMPkgBswi1b=)#DTe?Rx z+h&b!ArFM`Z5ei>7_Q}Kn6gqQWaQOl2OJi$5LmKD4d;UllXX0thrnMAm4{6hAG~xF zX+uZ5NN2l>i6?XJ*@NtHBlKd&?G=yb976V0{jT+YkA6I&7+_V?ygJo%FrjHjuM%)` zVfqni^j|*InS}-m8ME5}LXgl{XhoQRR2!kBcunw-&Cx@B-G92y!TJ|ksjN~zwmJGy zsZpXHTpN`vF&eXbch?QS`kIw2?>3+szJxvqEU)?8%`RhF#)XW-gUw&lw|>?8vU(>Hq|xPOkPzrn-s_oFF&0OWqA>6RjvZfstSdGG- z{dq4%!=*H)%XO!b-fA+?ySXC7n=CV&Z+O{~RKPW#>YOvizgPEOt)A$07jy)_K7*o?(q|w!E-JiHsq8usBl_f%G1v}}? zcY^M+?l!W7Y6Tet^>^rxcdoX3+fz>W@*`kLtsn&70G}eO@)qb8HwM@g7Exs(RPKr= zlA$P6^vG?{8x_t~ko5kf=5LF3{kR(LudO^o5}P%fs3l zmHL?d8ny4NSnSsyI~!rGU6u7ui?eRPO9q4SY{*gFEy9_oedl!a%?Eiyl}6@BBQjsi zh7jsRz%Jp@=Jc4!MziBU_nZqRbXgAs4sZJicx<@<^)ss?tmfpB{fyzYJHh7|uAQD3 z?bx}RnfBdm_GqBHD135tMF&f)jo`Z9BHs$LTIJ+?mFp6(_(tU_B?aM)N(cbxMjo2RU4~= zUpgrvsE++6{QC;94%hbG+_x6Rt<5c3%icBhyeOLjMaqn|pI)BEm=eNgmmlS^Ore0% zXrEkSQCaM==j|&Y7H!jS1xXcJ3#>~QU<$QlaCDmA!-Z2(3k%m;B6{$$D@ zUQq%wOz+|xezK6q?*_k)^0sD zXASx|ZKU0ed#}M-SOie1{rkToMwF&M>=btYZXPGkyse0mv#y>>SakYc(%ZM^`(Io? z1~)MCZBbPi2vO^TC|CpJn9inK0N}Klkt=3%1X?p-oM0JE97rj*o2JDZu}?hGfn0AZ zOe>>pktPFSn^&UkfMmmiu1{`e_dr*cQh;jaL@)x`Sj4M2D26M=EzSFGhs zYfT=SX2N^2WkWCL;aP;~DRb>mfi}ANt|aivt|v87ufPwlw_R45LTWz zvZ4I)$}09+PK*a+VXuACjk}gfxdk!^Yn6g_?bH&j&JlDWMI#1#v0>csz93cuXtiGN zWr%O=)Q8c78UiKD@!>&@prQLgGzmUuhc+IB2QD(>FIGRO%Q;d8mRN2DH0OIN2QQCD zCT2V@2B5nlVQjLKE-jEjiBJrHF80$ZhH^=P)D@OzR?12kPQHTUB~kayd$U6?znodT zp%iHeE6GK%f*652uk&AM+Nkn`unijl6Ng3Tv?IMUXfZ2MuIMirNSRN?0#pr?qjPqr z-;4$Hq&tz``0T9ub`bIaZ!6vfEVEPCgGE(>dPqx*1*Mc37COe8+&7W z3hA@JA$|Rw%VHZ|0|d?8B;mORA=z=m`)KvaCH?V66Jfwj&LwZB&cSFmGE@;~h=1Ee zgE`vzP2$9=!JFN(VXHU*qjx>$?u`;Y?ugz?+DiMcl;cKEWB=Z32_!UTjt;k-#Y4}z zE&->1t8eM{C!rqp9}UKbjCyUb_!-tN>b6qXi&n-a|4Y6d{@zA7X>nhtOij3$SHy7e zC8f;_(g`&%2U7;e2H2D7*}W3&56B=$*wX)V{sFWvA%X(lDHCD=Nka)~fNUy+DWE@r zV=9Cd;BCU0KsQxF7ElggsVbp80C(a{SOnW{6AFZMx#eR2JF-(#Q<=I!bU+bgD*EsX zMn{!cAQ2?VvI_M5@mh{8fSFqC6BTEdsW|Dt-!{BQ^Le7&q+O=!$AkUxW7mdc#`$b1 zQo1b<-U9Rt4dxAuIkU-%BxkO$$~eI;hyzW9{g4q>kZWT&)laaJA?-3O08COtUl2XU z8B;%XFb_%~%d8F`?#318+zM{caP6im^5gdIG5CrxN9rCmG3r)#YyDwrBd z97@U3(#sm6eNnv;@YX&IfU5bDet{IWqMhuBxt&>%R#VpHu1NvvIBh4D7DrOmUV3At zickXmh2B3@UHL{8!E089A*Ieud%l&FK=?g3cmaXmP+pu^B(xky$gmT2iB<_Pwl$IC zoHbz9n{Inc<~r&0>m@Yj6$;ZsbPQMjh@zhEo^OAVw+orCGzgx|1AG(&WPiQ*r9NIx zENUF_K%y1Zs~zxyls=EeEM)k*WXgB+geM$DclrkqLrKWq7(jhf37|@QK**%vRbuMj z<_g?PHHqRvwciUfe&S=Sm;0uzApCf=PM>58mJ$u-=3|fe@~~_Za9udD!G2`-i9E4^ zAI>>C@ZhP-xHACw0UiHaDgVhebChc@Ly}-ZEu-hiEbL<(U$?HS>H{u?p_k6VqBKNK;mTQq{h0n)Sg#(4?OEj}3S)%zrxe&AJ@XgE1P!UI_nv<&-S2!J@TPXyv#igwgj5r+KeH%bO*u)J{@ zD0eCPDQUiFVJHeWL!@4;dich1(yr7jFeY2qvOen=U;cWt>jAz*qv{0dzBm})q|t3IfCjAZI-s=G-1Je9 zv*W=ji2NX?G5C+K+$F(9-MtUnPx9W#g-Ku~e=%9?8Wo_7(Jw*{i9T;ox{TeYnc1*l zs19B%U2m`%1m&8x#s1}w^EGuYwaEv#8+k^EPnYSr8lMHxe_j8(a~Q%+B+wlkO98a` zdGJ@imQt*mzuY`BPR>r$awuU0J4fqnrk=16n!jN7XvUr6bOn%J!H(OhPnUi;rv;gN z-?wz(V`eRBq>;y_50E+8pg%#e@-vpCgVocr|CQ2U3&?+#{X;Ftwem+F&z(`d$fTN~ z^yEBbk1|welAyiRb(8O|Ee**h92#Kp_S?iLSU+cuWdBxpPmtT5MnrNjr)U?Y&2!*P z@ZZxmXiOTbrkzH(B;uTDmq(L_?>ygz&<=S{D$K&N{|Ru%?-H=wowkWc z1ECU8#^n)UxwX*q2u|ekFE_rQUjNvn#0em6l4)~pqFy~es){`B+Lf{?F`)kt$N`O`5 zD8!kOs}uZ|@f-63eYLAm8Nl1~3yprgJ5D|lqtdp7f zBFpRf5;Ei`PTC*ZwjTbIlv(v_6^XkN_X)3(jDp6G`(`6GpDC0=s0A3YhPZ5+UQ6^k2pcRn@)u_T_h02oKiy5>C6uXil`PcL@I8U zn}#bjjOJu+^R>&*urFdrV8DTofq_KbK+cX;UWFdTcSRWPqU^_5`cQ>iG{)MU>_G<| zvIz#sVr>-*NvdYQISHUrzxlY6FhE9zV{CNhmv@9A^yE*1Pa%r%QyR`w^>0dVI)-`x zy2h#etc?jrJ&c}=Pc=+3orUs0YFA6~1J!wl9d`zoB;_c+W>m_Tl4&Oy=u>h3$di)O zVb&)#(9jh6kp`+Qi81_p#k7PjFf?fnukUB^C8-|d@}3aegEhdZtj`Mgvv9ycEv)W- zUOYGnzP9u5VzPWVvey%UQ%91O3#wv)?~fqTnwHr8b(pivcR%oNd7l9R*GwmW%Ho% zxhs`6B#gXAmIlCT((1clyV}*XmcS~pN6UeCl@o!U-}Z=UymIiPfiqo0+p%s-bb$0y z?iX^R4*8OvC4n*)h8&f4G;jKRlzfu}vD$@%3=X#)u}5L31+m|mSS5JZFW15R`X0^N z&N=BSfs(8##qDCVCiv=@GGSA$uC3`0dm3+(HztlJ>VJSNAJ6Jlg{K}2ajzNf9Bm~#4VG^i^QWAkaxWI!8>_z- zj=Kdfe;xu#UtY#Ro4GXH1TsE2W}ydPtv@rKD$jitVd`w<$0eBK_yJCNMs#oUd%X&I;im&$N` zCs+Pa4zLJ%0PUPirvJ6potMgR%aH#a+m8z-JC;4Ge7xD4uwNV^MX4)7lAQE_U5eDZy8W-h$@ zuGbdWi*a>aI_%lfbJv`U89vyL5%v$Uw9)&#BAvCm;l0k%A!Rw1v1L*Vz?+S1bI)gw z+6S0vGtfgE1yB3(8VMFHSdTlo?nIT}o=Uki`D>{Me{ zcUIK}3dI=bZ!pw^7dc$Wi4H`B6|cZFbru;j%c%cK$)Z_qbc^5j*J}?@MyJUN=?J*a zXOtN9ci_iPl_hpcf1r6TQXJ-0=j1x4c0sv`9`~2JD@}AnshYdeLY8w7dO)q&D+8z+ zt_Xy0N$Q;$Nf4PS!C|d+6jb-zcV^r(9nE>{-ndJtGMZ~Y%W7e;6%`un@>b{JT$bd> z1^S6h+I-^^j**;H%}oM8dXqVs?7ZfziZ}y@36cp=V--qr zJTwhV^G=%PzBU?he;;L+el-t%xmv_9aA0MbyDgj`?1&V(IFuKgxVy$FmXrirXD-O; z;7UTT&lhl_Wro@2;1X;XnV)3mBkzndao8*!$N`7|8z;7u+8{_t{^-lpHK*LB9f^z+ z&_03k-5Ft-VT@dn9S5XuTxvECs{j4t8GaZ-*m@X3WQYvT#mo$>FC(G_sP^0MGb48& z({+VQUj`aek#l2+z`-VR+v=oRH*5To#VaRQ3P^o@;&T}J!>A2gcyl+^#l;lo;Ll$x zskLRdMcuJ-HcZg5pxCbT>9PRL#U-wh^H?Jw$g9!$+U4V^I4i1Ay`{8uRB61#Im#!k z-@G(h?tWbLMa8hGv*2G02$0DVdy{Fq%vBzv{`+s!$OwBEa|+@ua{-d{AZ=VsS? z%b)6A&I~ZqIo=}dKj z@+lS3$5T>qYv#Mpn)s3RI1R~GeQd3scGuVZu#_8FY}(0pOqo(y>X{yyfmSD`%gW3E z!<=TEJWqy@3=2xK93Xx7sa3$ zgy~7*;U<}AJi>xJw? z56&3$^?BW|Jt^uU^7~`A`{poDhG5yV|XEN=Cs+4bEfz)9OAl(elC_s640`4Yuw#@THNCWVCy|?e&t_!_ng?&=8ix=Sl>iZ>k&InZ9 zc_`Z#+Oa==L#r%heTLuSa@PEht*f{o zY|L^{v84a;@$W6`Vu;v0i=%EYY=aNaP5FinAE)H+W)CWliy6^F%9mL^B3(TUJvYI& zN7pOe%QL4Z&r-thn)Qz4jv21u5<7XWlrF2vCzs2SZy#50+AR2&1W4)$b?zj8Ud|eZ z!+x*>ZgaC1_$~oE|LbE;J)hF4>Te&reft=+BffvprSi1Qe5dJQ7qSfN|M~dk+sEMF zKE52i`HPt4z5^K@nkhJUbi?jj*N|oS%KRT6YqFPf|PLomV6yPY*y7GW7 zE!he0uoVK_%ZCEGI15Y8tUWF~Qt{j`E8#JD zsXkn%@xCtNKFugO>3C*YbBGX2BIlfvbP&^_Vvw|eAw~XU3ErP7(YEkuiqi8D`>rDa zF42HGj^DsBMhyf-S&#-96bTlD z|7z{dqoLj(KaN`mV<&4dbcL*A!VI!!r?M}RCBraTvM(`~8cU*E*Vb5*Ffqsw$`+C( z#+pc`WKWtHYxoV_@BQA}{e92Ul%uOUOn3C&`|Kzyg z&~e`%j&+V48~q2zA1IEa?Em6e?uX+!XQn?hM*SZgetUjvOoeQD%%5N7EY5r0&B{#4tXV%O`ue0KaS##FLA#whUuT z7A%aD>EAYUp*0;|mfiX_Y1!d^@8u%NE3RtYOJClaJ-AdMJXNj6o{sr69UA2yY#lVZ zO`I5JpZB&+^Ggd|a$UeysFf{Wx4b$r7o^*bsf$p1M034O7juva0=<& zKan2lo$(6u^ntdmN0&|!@bXJLCvOo~8dj%Ih)D{`WMV@imr(j^`{P16_oK6yXZD68 zSFK(le~G~2Lb{@D8z?Vu>s<=T7P3770`+zK^SMdIpTafgC~ck(%gVOrs?w*lS^keU z3;eCk13$G{;qPrm?Ikx*At`Of9k#jgu+6N8ZRYuRoAo8!)Kx8;z5u>ae^-^U2;=IJ zWZDZCavI~Pgx?W(H&{Iq>wGE6HqkQaOow!msKfl{awgYWkFyC@?cc>;ETFgB6G<1| zF^$z+KvoWPpso-QTyM4O{pMaU@tM6uReS8G6Z=j#E{I)Zxf1+aV8yge>Rz6%?ZPJ8 z_oNK|HT$G9h$?g>`fEG-F0H1_fV%~vYX71Z=Jsrd%X#U`b0^lkMmQD6Q;_y)PE2g9 z8SLn!28s21osQ2k1hSug3IpVp&Fe%jZT^QeYi9ps`ghV1+x6_&H88=tEtF%LW@{fa z7UgIe^`**?+AaH`EqTM)})BUzFnN-WTG?daomB zcVM;2Qnd2+LRsU4(5}-g+G%$w?&g?Dtz<8pII)wj&TZ1&yPtEEY4}f>uKeF-8ufFh z|CVW0HO}RlGozJLS{gCV33|W_Is&H(6W$EMSxF9n>^31G6QQlz^st+-beQh*yZjG= z`}^;6*dSj98Y=bqolNn~(oEq$=*V?Z-^Fk2n0|)>Zj1ja&$#*85%%FhHw-n_*TR*f zqQ{p@m+V@e8h|6`r(ZdET`0Xi{r_sZ9awhQ{)2->65Zj!< z)Pv&Ibe@mL^UXe<%MwJj|WuF>b@kGpxV})(Pvp+cd&7@&IO)n=A2nJ&`WI8IyQ`||1ci3k_&n2fTS7o z$qi+w*92IuYMBk>ZJ1qRe~*rcEo;4j^~DZWZaE1dSPb7cgA78)@w;+%G9&9r8p2&4 za;K)~h7$@ybcdU<%goL<{4PWt%n`L0T^g)vCoNGbOWV<*{@y#awd^m}eq=iJ-A$O&Ze(eay98!hsnyHjE~W_A0Ak8xA!I|5yp3|Tq@ z>5?f8kQAr-Y>{G~LY*tYxxD6%b$%8*M-?T+d)kA`Sei-wd8)=6aAonY0*{KW zbjLT>aaD4ZaCs)$a<$|8!=H$Elu{;ix!@R-(*asenb6J%3A+KgvG@5&2QP_(Vo&0; znzL_-KSDrIsY=V0be?7;J&9O17RrR4-|?9pUsbt&GkUckACWYepAL&>K-POChL2x_ zWhKn!nF86zK)|q?>Ky)cFL;JZ*LA^QI(Fq5$BiqtaqaQ5@vd$-rUON{ekYT(a>|5` z9~eQ#=I=z!#Ah4J&0J<)Giu^pdU%<6URl=Nm!M^AFWsL#gF20X-|VAI=)QlMreC`z zdT9DtoS*2wOqWtj^J-pi&-Jr7GF?>|X84pcp&zk>AdX-#SpM*$aCia#@%8AFVy1W< zNDKl|P*7A*RD?)_6(qo5iPJ!NGe7rpu0e1S$#bd@FhmLb^DUYFbs!54B@sz0xSfcp zR|s5$a@(2AsX7oJwbIk=I#4>dg+5eCTUS9vS6dmTt*i}$!gO>Mv{lY&E9*nGAu2jh z4Ux0|c#9KI-W2ZZ8H5l~Py#C-y;(VxVRY5Tnjbx|nVIpSWp8rmnbvN3Z96*Kd>If{C_ z-`WhHI=kJxnX;L)nG=~4*rK6q>)(})WRaV7Y-AKGleV}|ZQfIy%o{6Hst85-cXL)0 zjPzM&h{=9rX0l-wACHjN#u*7cr2K}O_M0|!7V})}whOgkyw1fOkz)$_3|>rji8iTY z=N*7Ogp9)4U*%}*@w13XBfj_md@!Qop@|m#LLGKnZmhM~xh0D_l8Q0azxW34(@i;#m>x2sI;UoEhRCLd_7_ zI`>?lXYDZ}L_V1mxsV`%S^&3Lvzl=`H5+4(B=^*G!^ z;sao#Ham=Qr;pr>jiV)U@KePT!7!TCI8iW*eyb!5@Uj(_MRm^wXc{lzpJV1);6l~Q z@L8xT^v>+=YjP*{g*{{7R!=7{>wxhL9o>1YM6K9{YEB-1k+^=H!iJ>NUIG(6$efsq zZoc0h0fRJP)Y=j#s}seyHpudcBsc-@N7u<#a1>ru*3-!H<`fcg4p z*R+`=X|{RIP(=~UT6Dl{XC8AjZj4iwuDF0jhQ>aIJ8;%TMY3+ey?N&njZ~{etMQ!F zqKf!~$QH^$Wr&n?7ao^u0Z1)YJJ-Q%*WgCpt5m4tZQ>DL5)(8*7shC&QYT}+oO5!= z$f|A{Ho2frASq9`Ua9n~xOuZIs>Qq{IF%UH*(vTOPk7@!Zpk%2#`E9@b@)~^sz5<@ zLF8m?ki%R@;sapq3y|eiyMz5yym}P{Yxml)(N>RQk-tH=<$>1{tnHM%Ti9Oq!I&IyP z*0YM=+NP%jbwDnRWnLkt%U`&CTjvg7J42mx(9q$ZN5AHv8D8<~N#$$*LYVj`(tUXV zfBzL&_vbtO^d)9Ih!SV^Nmrx-44ZO{}naF{SEwT%JYq0Q}ZHwMKS@Wh@ zUcZKNM(Fk^@0(tg%YZ>QKexB`R|lpC-q=5%_p@#bZf3?iyxYSlMx&egf-b#784G%R zWuW5ydiBYwi80E;OLIzMm_NegG-jSbMtDna5k_ERqj~;;gx2&XRj*pZUwtHbd!k4k zO2zHGn&Hlz8>Awt%&-*|8&v!QZq||c z)_4Ne!&geG7R%gwH@mq;qb;(54w*T@?YJ%7mXOBX=fUMTok!I7iMv!O@wJBOMV2ao z)}r#1Iwv%aQiuhEg3bJBx(G5kPNTqJP`lvv(5@d9OI$mVk?>&zX*5v82?nOdDilD6 q=m;kMhmCmLxaJ9&@@ok>pzfH83&{LJ0EqyVAW-0`Q+gNlf&T^A2BAp+ delta 70106 zcmZsBV{j%+*K};#wry^(v2EM-6?3z}72CFL+qP|O_}%aG`>UFoe_eH^X6kg$>8>aQ zq!j?-bpZKv{5cO`9vNz@+~2b`W@7Dgl5@GQMP(OG zrmv+Z=TAsrf8{?F1fxm<2ULz4hDq+ou8vYzOoKguO|Fcd8`?`bD}3O!heuBho41}d ze1O#u7%$Il|IPMm&q9|H3gZ`8IfKzZIHla4t*6|ltJ$B+*W>OV1;AFv8nirCach48 zBP|p))M%3=1lOUtj^8b;C!yD{*-Hazo;A)8 z`+hmLlWoFo3|D9?M1=mmflDbU6biIAssR4idFLtf_M;)b50nw!L*r0&Sc|4Ktoed_ zlP2w|y`7V~Pa7meo^A_gfpjV z9q~5%u1f%e2{gmRhuQ~57uUd{4N1xpl)qpn+)`M;G4Z6DNIG;$G;U!GUwh!GvP9|* z(Ib$keOm{NW+pS_u8S9frZ1~x3-EOTl0&`_4Y#$jqEJV4`+$0ur7MGx#s}?M&PGLj zxbG6u#&|-@xNusM#;E!*oW~0LDs!zWUl~zeB36)0FTu*FdK8eOB2ZwyunUX?s5KrPJKS)D^#`4hwDir2~owZBu1dwUMm+kk^ zobo4+0QC`4F-uC}p;J^PbKKX>0^`I@M@?nI`x_BSvM%fZoOp0skRQcVSDHz=LM3|c%B=p_JSG2%i%95SkTP%dVay;n{d!?d2(UTd*6C zO$67AZ{LIhdK{c_V(WskdxmQRVq;^KPRJ>u=w?cDfOU|kl^ZxC$v{1ydc-VGvDKFQ$Kn}SeSRFe^ZQN*BoJ%BiY+2 zS1M&fqMkOkM}Yr8-T?$w#ABesgrV!*3ynVeyjIp3J!dzZN&0j1jM<;PBcCSk^V2iJgX?BOTf8ZS**KaihBnrk5SxtrX8@zR9*u*%2*uvZee~S2 z5O9TI?bu{tLUM`HoFEIy%5+F_;7Bxs6Yc zb(iVgdnylhy)X(i)PHM;NJB`jEzPYZ1HgR)T1FaS<i?AGmYZEr}=cepu+?Vw($Qjn4ID zbmp-%-7WY|SK%wa>?Cwzl|O{S57#wtoy0CK`CqIo)U;t;M#oaF#nZY9z@Hu`_xN4aF`S8=FX-6hG%1tviL__*QFM@r zeZHec9FGjEgl5W5 zk+qtxa%OjHAC~_e-Vx{si;B_R7I^|_`(ShWMkYa}>Td)BeP@dbc>^BZ@D|stEIO>3 zN2UGF=!n{$mn{0!=v0@Z*S>|*N=o0Ve|(c?+G3|*g^03!NeebUD9=#m%sdhQR|hak zP63Qt8MrRT4ADe_G6n2s!2puTW2N4g5zzt-<#j=aft~Uap!JQQ3WeYRsSOX$eR1&3 zO`y%Ivu`Mebnj|r10gmF&t{Ihr!R_LtA5)UyaqeW_E#*jDc2KFEItlKZ=GV(5Enu)E;k67m6r}-uPs8uj<5{ zMsaz2HG>G@)0a%6Aax)%bkQ9A0)D@k){)cfv|u=8h$a+OGrt@+XDlaqBEIA;9;=6Q za!Mynn!~R<03E)d&XBJ<&GGKSJu#Y>4Aa=%bVE4O6P;R<`3W7+?9rX2xtmv6v_)6v z5a4Z1ySsvJ212xH3prUN;D;(3sQYKI(e*4PAQxC5b`aK;=9ezQ;R3;W3d{de{r4tx{w<9^ z-sv33KV>K+{5(EHZJImfpZI>{>(^e98@^e|pX;9p$&M_9dyE-sA_TaXW^TfN2OJp+ zX+wlG9^$=!yE^hIHc~;a`MjpGzMzefxXs6X82ISkIz!!{g28UCS*l!iNA)l~J)w}O z`OaoLSUBU*Jq;dnDDiBJXrF`bo`BZ63&6dMUYdtwe3}n0XKyv0{&K|ol;Nh|2#meO zPt5c`mA|;RcoquF@wiZp`vCTh0#vOu;j|VW8r18D2l4K(<}MXJu|NRu-In##hamY7 zsxJJtL<({Bn)Z_S9r@N)JuqAaG(PB3S>Y76f!qismY@Twj4Z<8IWl>JIwHHu>MkC* zAbt!JYYIVL{6Pg?JyDk7iG#}!|1kPitO*FhE3*GadHjuXw_9ahZ1fLi0??3YOjL!* zGxPAoiTOMZEg@f9$OjwNzCQC#v1M}B)-8INMW3+x5Z=8EcEv;0R=jSI9(<4;G_TvD zNM(+&+qZ-bnonuu*c9@>t;9B&Yh7I`3Sw#Q_1H@`T64v1Xa+kGs~`0kjGdI5pn)CQ z!mBT6^*FmPH7$!=#~3vG0Gv`)MgK9R*%Qx}A_Aoy7YQ_LQalvh7$GYrAMoC@YHK|V zGET26_pWJ(XdMU7d~JNNZijYN$kcTYi<86TQ7=(1iNT6%>vad$2HFT$I|w!O$sHhT z^d#gn2F>QCzE?NM%`u+JYE>G%`UJT*Ke22OO?5`_bNPCq`-LGh0m;_p8ktcO6@w_0 zv&P)YOk*-SG7cJCJ2i%%L{$Mc-7~D$N$kx5J5Y$!&N%8T5S`%LEg?p>{f&~pStWn# z9+a@&(0$~;yF8MqX$?Ipo9Br?*G>e`1vPBz`I~xRPD}{SlOa_hCKVmObjcI6d62a| z6U~BW03wb!hepnC0P-l?rr6?{rM1pZURwA@nuxQpr(V=J>VmK(l?}~lM~!c|tX9aL zD>@V{vn>h3m5w7ONIr=C(aKU8L}QAKP^+nFhp8;j8ZQXSUzt=Bugz}gPCY}j!Nvj| znNc%_`S*ymb4;0OS(+_go4Z9_P`KmQTY@s4V(Y(w{}#Y$Ung#9;-s0tlNs?*({XXLX)K`D zdz-^1T+Lrs_0Zy(@7s^3mwNu6(LNX$L#Srghk1X^-Uhu^b1hEs;p@@ z^H3wiL)K%^^VJL7q+?m=_6O`C7X4hDmBi6l^H$E3wI*QJ@lb^Y$V$+Ee47Gc+A*p@ z2@N=IjBt^A?LZBXT`?V;8cjUv{E zXqAKax$3Ky?Yz7!(tuHkwo^(1ljRQJlEg#^FSNm_Q!KLRS81DptP#-U>L%jt75(p= z$}>%%rkY0I;Mn?GV*jH=9la1@4%9wDFVSd5=RQ^2DA$aT)S|Wc(<4QB046f`krNDPO z^~X%J4kaTB%V}2G83ELXlfW!(%rZ=`_^Kx#Mw^wkQx$W|qg&cn)%0cTx}_#}C0)u| zXl!=~+|{CX&Mat$&^eZ;fho8f-TY-oEw0j3i;{p6lpIs4{f!LkXTVpK?Gft}rgRi& zWk`XP^jkrrtD|#z8aq$g_3M8XRvY?&v9SR2 zB@Za~P)ibDeo}1kEj(0cthqa5+Jm_?b+f4e1ex!TTG^4Js?QwUdyx}~Vgze(O$Ee& zTBsdi6x&?5>gD7q?PR`+6^tfPdg?nC=D#LhWI}*GNsFLZ6RwF2cKqLVUNw9e;B?_S zN8_kX3Aikbm!|WfqNy0K+t!Lj$(SQ&wTmsT8_QUcuw?$5D4a$?lbkR&n&=mm9!9A9 zAZZGoy~E(JnFKl+U%oQ8+OrLKWavUNw_z`?6^;0=vb4&(sp4b?TYazQ%}y&1vDzD` zQ&Iq4l>7j%JX7U?d(cV8vwqiQ*x4}aBtoa6OL%B@B=^+^0)mzL%XRRFFlYhKy97~+ zzb$`e3+y{vs?L57&YG$=RgFr`f+I+w6t$Oc z3T`8Of4+SKhWC5$vDr8dD7%BQ&P!9fFeL!PsTC5}y-TGv_LblSC~>RT;CwElNU~Uc zuA>eD60p&SOk)`{w-=F?`mg@?kpsLyf!K%whkHTr#<7Lg*4lu;(xP=Dhmf6@AXtWr zJ;{t5sPkWNI>2GhI>dYN!Z&k*4d~B*4UqYz&!ey&P!QdgLyRNBfhZtS&f?=SP6z-z zBnyCQQtody$nwzfYaiRfP65SRnv5Hq zCH*@A0^Dnr1it2s6Q%6pEo^(@Zg$L(P2QZpUhM|dA4tr2j@y+*)Gafx!1Coc^_AY@ z@?V`)A{;vu!uITSBSxNmp?0oCw)TKabZ7F;!1H4LKBebCj!7_R7!RkWD3|Yb^oKyL zU#H)fBrPbMZ6R@61x8x&+W9H>x+R9&_-J%yNNIRE zd52P_DSr4P3SP4$yaU%!smKhEO%yOUAWy}*#7~{h#E^aPCW>KZt7Vj2h9_Wj;FOxw z-piy!o?0NZ7I4xV3jL&|a^C=bDYonNQW78BOU+Zri0B7Lr<_T%6<9~PsXoK}@uG2* zz=En6DD93TtVg1B4*4&wu(p^GexFaB$DH~;N)E5qWG56qYR2>(w%C*%xYLp&;*izK z&FxxGjBcgH3gSot6A~mNr^43XGZ4irN-kC1lPU6=R*gF#JUb+I3)_GXIjb^j3ncOU zVNj0h)D!qilB||WZes$n4;Q7*>QPY^BY0ykaWMsLy=sY)RUi+R-{R4D(BAjSdT4k% zgj$d~a9jr3n8Z~x6tpBTp(P&;<_ATuxZi4L>-#gqjtXX|i#6gh!UrefOc}&iXI2mk z+foR6k2$vj{j)H4w;S+vyN!kM<^0Ny+tBBn+9fI84@nwPuv%Ldr<$a%-h?fW16|ot ziv)q*We0H0UoA(-e`ZT|NiA1nIxSbrQTax(h$j=83`U_pU4jnhx%4pVIQM~%ZO1?X zp=GVNjF1FJ;Xw6ENgGCC%%7N}GxVP%C}d8SqQ#_Tc|SZmy$3)Ia`2L<*)MedQ%?-m z(DKaoae^-IUHW@=RL3y7T<#Ks0vUFF;v%5lU>+x;A_%i2b1d1cR-7((=^RQNsj zU;QZP4Pz8mn>K=_r!owk)>~N5KvE$F=7NX8rh0IGvm^*M z_HqoeusEMWevJDx8P+nl&K7LBX)E=6H%pC>9$tCsk%Zorl8@+CRr#<4h69W-22Sho z#6-CEAfqRrn3BfYybDxR>$$k@E?HGiqHVGbiz5T+~ zhvKbf{{cAtdIbawA<6(GLK251sty>G;BP-05{Dkv^oq_jiZqM)n|@G43UDv<@1?5A zvoDIyK5DAkEf2z6F7vr;y2O7RVN~O3iV}q?A@a=KJepi8pI7`e9V66ZKrip!F*CVn<37WL z3At)C7$+a&y?j+;HM#dcL4Jt+c!niz*w~>pm3BrYzq zZ~&>WDk#H9Gp4ZQOZqY*f=ZZct@zqbVmb%3#x!g}`RIl^n^_&g%z{6sg>$W>{ORe# z3~pr%V678xF2+UG<@u4tIDc=wKb7wO9VuaUF27=CUJC+gU+TG?5Ig*dV)j4Nhe8d& zRWiGFj7y4YAAGhH^JVRg%q)bq z*jxEW51*f!fWtfZ&li#n-zyO_QqcR4rKVd$XTo1*4~jG`cZVWh=F^__5kW-%qI0K_-0_ z6@eO2;`=kJ0F1i0)U5aow&Bk&yuYPqQm|eMlo(yWQy!_wC3kZxff)|P9r2DQYjnf% z4(3)#K9b{wYbI{#QI{!XR4AJ6;Z{3WpBMoJ22T*}9qnLeG4bk7Y1)})!qun7p4Tq1 zI8pw9Q(Y*PsO~lTK;EW&b%yx?%kdb;BH5ZNC`s^N)wP`grMH6v)e$Gbz@$TGID*rW z$0tJ~F8<1}`aRKfRH2*YMo!0DgHOl{>zG<#K0i7z+a{Kr9UeD^Vdm)I<#1q_tl48f z3)Q&FLwzWs2ye%|5=tZNt#|S6{5--N=|G;csSYT{y|0f;+!#c9A zha;#xx3zQfM$0Y&ibhT{)t<*Zw8^%n^zCcBM$%Of&=752Z}+#0g4<$XD8+$Hn#TjU zGavI`&#Qv*+;Q9wC$)|}0bY;G+(Ny|Do1RZ+{n3(4|-PS4h>Uq>6e)iU(<0>XgYQ+ zGsWYhAD+AREqssT5&SZ@yUOo?yX%u(WI#j$TRxkA;XZfqj<7T^yrUXx17A+T-BWnW zcv*AT%%cOQuuqi28vuTxn zRiQu5>7OP~IqGv&UPeeO*rYxvG2Fw0WUOG%Zd|P0O>`stEt#%pPW)Sayr;+ks%Aur zhd4QU>ke|Mb@+ZQSw3hqUoPkKc){OH#Y|@S!|P{B!#`}Zm9JbjYlKcKB%#hbBy%+O zHO6O$7bCm1Tx8;pXPrENaXDsFV0oAopO2g|>!tEzAo|^g#ya%NzFt{+x3#n4DE1HH z>Q}=&6KxAaLpeHKTtv!!eloeoguRMp69#h@r!nn&a6j&tXi@%HqsW>M`}WIz8DT%1@+!@31-Hj` z48dGK!mjJZsab)_3u1dBzWgzEu!w() z!o2qnrOY)5^nD6ICWDIx^9T+it1hBtzrb{rXftQ!syN=EyMe{zNJne!T<&d9zr^^0 zMPIrMoKjiNk4dj@!L|tEeKbK^hGAfqI2jG+gxv%+%l|A9H-B+C zwXVl8_JcEt$NA16mS?)~X7pnq)TXK*g&n#*h#SZT;6>#53^?NJ`+TpzuGo*_f}H+) z{3hE)&sYq&lbkJClIh_&p1W@Yjgp?WW2(|eV8v{-6B1VhN5MGs>2QHwDzYlLA7|fZ z?_KH@5kMJYX%=ORx7(v}q7_^wlROZPA0=Z_yB?_z_KVlPd{%dTUDiqu9p=os=i;s~R4##*HrIkN4wSXT6v!oCj<7hM(+hT(`qYvl(9WhfstY9(kKyy zs^$&oRW6rv*i;lw?aY6d3-eu>Bs@=*hZ;Bm1|fhE&J_@!=#QD8b=K|I-~vZ7QrYu~ z{G7@oF4JRXQP!HEzF@%d0X@faf@)7m?IM}=IOo<{R43%jG5UBi+HxKgJUr9ZON14I zJ6<8=2Ykn95=i#!Zpw~sB4O@u*e6o10~}Lo~S$>4oy_6jYhhh%rfR z%EF(j$^a4eO2C~1N*lG7nXG_X7_lL*$A(#JsSNhg8>C!1U}jdsey9~ z*Q?bRK^02=PB=Lza*SaSA=g`3@DYkJ)w$dh4{bxUo!b8|j6l@Pi6R3eUdszRk_PrKyp`9@epq&0F#FfaPK+93*+e}Ui|G9-Evly! z_y<_vRX?wFaF3C>*ts{^=c_pb%AYi{U;Tpry&D9P82_AT7UZu&R6<3dg2G@L#*0lY zUo6YO6Gz=1HOkjwQK;Jp)T|qnJKhHEl2s-MYH>;%1kynb99HO#kv@QbZ%AEB=>*M0 zj>O?`_V_O4fiwTlAo^DF@8~4P%U=}H5~+jAO8TMDuqKSS@jllzky<&H#&HrKmi&J5 zcOQvl$4xVfSWsl3gGYYHWMrRGSNc}HIW!CCf7w9j*Dn-`0yxJ&gvt? z#gUXI_S*kEiy&5%;8g$${w~-C;pIJb;n1$<)bpaPCz= z%lUXW7#2z;CUJ5&7YrhcMV}WVZBdpEa^Po8f{MSS&E=_X=!eyRKS%t1I}7{-aX?Iv z0&h+1A~#tH^f3dtkqBS7Vg8~qWTVWx>x~p-TLGs zT)6+#KOf=F{C6J1$J|(BzwEn`u zr0T3U*4AhREtz??L`oAoS?fgwwUVE3fhR-(Uz)*+8xKNy>8jsk;AVympM7i`kLv&v zkDk{F$6vseP|1_s6_p}udcd>$*TN(f($keWdlK8l;M<=>r)wGg`*Ionn?cSFt8}Ni zn(Q%aX`rH51!9qvh_+q<$chxZaKeILx#^Hnf;@FkJy-}{0pOKGXJ6A6XFOv0f z@N_;x?L=iXd3>+J3%D-I6ii92xrs(G;Qb&P#uPv}sIQ_daq(noq#3++>Z|`2*D1>% z#%AOc6dKzEi!}R=$9h_u903Y>@p6ZX5fBM_-qNlM&(ACP3Ey< zG3o;sFOIWFcR7Wn%kOQWPF*^{PX+iXLOHeW$8?NCCw|QAW|`ifD5rh@P90PPZx5NE z417l_enPbRSC!}lmY1Sgao?IpqUmyO8EH!9g4(s|36uz&ejgmXj(8$Fe#2J{Gl2&E z?^EOyrU(|g0_+bXCgwnaP38y|6oFx9kK~Zb=(}-xCUEem8LjFyF_B2z2L6OmgL9D* z{SPJ(o(C4o-MAA@00wCA3y?4c&!nDfN2GocmC@Ibn0A2rzR;f9zRJhd)7OMveY3S| zwE8`rD*kzQs1bDCQJsidcGba7mR+n|kD;9c={$!PjBuV<`AO7YtVB$&T0NBy0b z;HQ-3D?o)7TN=+dHc?ATFpLH4cGsd0js(9FvZ*(F&mFe;u@CTiNiY-A!fWua?|eU+ zLGtMI?blJj-O-7cB7fyFHG}VqCQ2AU=iybl> z-|=yizF{B-0Svm9oJv^nXc;q8AheIi-?5b+{ny49r*&Voxg9ugRdZ-;e??yOX?sA# zaRQLVnf8G220N^YsV5WvU181J}yhkLfj3Olk z`OG7dP)4_e=;T>4fjJP`M|+93xoy*pNDft#V1JSu_)VlU@X^`g<}dtk_YJm%!gYiu z$jX;7#mp%DSKW~G>`>{1mALdGSSJ!$J}3!AhvsbWE*pT&oJkAjtF5@9?qHe>#-_vK z0!5m-+So8hp0!g^Mw{S&=+$-Cb`xSWZqgBu52@Qh4AT!4&8UOZwLL4L-T6l<0**7% z{voW}h$VbX8=PeUtO@HcVdd-~hPP?vK(Tl= zH&p-2vHZ0ib32u>uCs97-Db&w^9 zBR?Vi&Nh`%kuXv?AJai{Ke=y3K*1FL^i=auMpxUDG=haIf|$s~Uul(#Nxn8`qnfGT zVp)I~O>*!Z?KlaMe%9iKYaVGKr51?rB&FZdt_dmsPbTF@0*9#|o0XJREyZnhYhbaRQZp%S309 zYRKmDt}U?Z^CWZnJQUNSb1gQ{?Ln3t<{|)5m-->FbGPP*>liWLKA)htQ?MMiHRqs} z*#Y$IbGIEFOtq63JG+ohlif^ge&amrdF1aQTdhAFO|RA2kkG)Gw*s3lA2%aVKSG%7 zijyE%iu+xV`7PAScld}c=egjrWKy}=*_a3DGMJ#f{2;gB18yX3k=mZpgN4d8r9?mw z`}6uQ4p5SL&825^e5T;4U^fMHp6Soci?fy&NJgkDidUx~3hIwS27Asr3tEikRPOhz z*rw#x&|~U|%4VwMnYt3boQsO8@Vcem-+gU^h`Fb(9vZ2?9LFn;B9kb8RoH#GA6}ad zE?b@|oQ@?NUmEhW)Za(QdU8V1gb@JV$h`Vu!|lc_|I9J5NeUk*V`$#}cn`nV0Nsx+ zJu+=@HRH6B96fJo;Ps4;5@gW3OQ=sxUun}Y;0z_32YaH1+2Fk%>LUUqgCbsoO7`~& z2E%9?x@HWPTxy0yoxy8$ZWnFC>r7gmqF}H-Lf3TiiGd# zQw^U}8~0n^NgYf}K1_H8n&?2;r;OV%9D#>&3rtMOs~&;5?7296dNn|<>?xMP(vXrh~2)T0B5>rm2H)ecHqgQas8qN$`2gctT+W_i`Jv7 zk|n2YGDG}XV!D^LrEGGUe>bb9yX?8qIhl+5jmI(7>Bq34-(0&vI#zI+&y%w4b!5K{ zh*b1#=r#%%bS^g2Lon!SNtbFisM@GOqV8fs@;c&7z_S+*$5K`}8bN(ldiAfb)R-6v zp7PtBdsK?0Anr=BIvuVgj?@(On-2wor*;t1@lDkTEvVHmuzhzdU{y)6Bhu?Mg`|0N z>R7hdl^bU-k{Fb_HpJP|n}1pqXZ@)KbdN$mZm}hUXXHLRS^{glDF3ku%I;AN;3>SuZ-7h0~C|ae+1U+ zjN+ooHnzX69l1Gvi^7F8Y!Y`6-B_jF8D|+g`A7|x--4xwa6xN9kLP8{KRBgVeZIW_ zhQAv}MI)=CSqrU(5XE}b9?(%LVnCsUFNdOrv<}{&6cRwHQ1Wc0xHK36To9lti-bsj z493b2sKMBcNYX&=4mx$^ZFN%&9$eo%VBYTWmQ?GA3-Rx zXLjBLn^>JlK%)B>*kU7wtKQ@|eb*kmJT!)gjR)QJ>kXeWT9kKhDX!J5Xi!s^R=vjh zC)5EDSd#tMfnFCbzF#=G^4-YEf6&^SGw5+6Qz)M6nan*xs`Xa zNtp7Pr~OD^{N1Guz={hq0$(>a_g{i}))EClpZI5Ax5uNT7Ic4Psk@OIN_TZd_=n%W za{1J7>HjEe1;FA03O?2MUH#||;~b+)h=_O%Vmx9(IKG{#Dte<~A)7zl#Gdli;M_BF zk0z*)$4=44oF`pD3}#!XH2fM9*4Nm44Q1W!rL|6v|H$bRwzmzDOgCqUA?X)`MWUB^ z5d9X%!o#Li)Ag5N<{xSxH)pOUg+6s%PWDrHE-=Fq9kJdA#NI~0#gGP07G$5a?7YEh zsIP9Hg4(bD*L^IYi}ClYXDwBAU@K>$m83$6c&nTts>1_`c3D+j;GnaQA25W}g362e zklHyh<^5_zDKj<<;uuv4`Pb&-jDxJ(gRFcX^jW~Nm(c%;!&E%bKI$EOk{+Vr!?__AE4`~~F2xh}g>Y}pnrmnQX9m&4 z;wW%nHi+G0EjdTXG2U`s`B?c%$XApcoA$YIFLa#%?uwe&lIpVb`kVoRP_}zR#Ehs* z*Dc>VrcL{qx3pjHJWGg9ybGU86*yNL{wrD~oEdy-f>1ugLmA!>xy=hYIQD=Npk$il zF#pxObLiMHw;9@l{Ci{mm>x^Y-MGkMe`vL3A0IPrSVCg%$KOs-Eu_wN6mz&TG>3tl z9GF4CdAN8CMQy&_0Ac8KTk>>+(^$(c05lhby7Op~y>; zCR2ia+a|A_**4sv0zQ#fO@76Ah;xs?B!(&UrshZ*K`$#aeepvU?~K0jl-%N`Hh9h1 zv5+KFtCdF_zZdK;a9qWE%QQmKCAybnPg=mJAEK#FP{dyPHh+H@^P4J>W`Av3mh7R^ zkt{|i@SZzCaB!oTS?32N_Ps0u&fEc5#E%l!MGKLQ(O;HH-R#YnKAlbXzEU9V zvNYN~Tpirr$+#nWEPgGYpb$m2YA((+jaigYQC7_)N63}}+y$q3)%kkg*pU4rGD%J~ zY*B`S`11*o)ZqbS3efNAy>gtns?}i`D22MF6%iSnaW6^p`J15?<)x(>DcAo*I;KQW zKM$s%Q9j)ArB12E3A^b0)bm}8Q>Y}&ry@qyx?r8<~K7z{?NtM%r~+v?YgG#5SD zOG3@hvGPEG@Ss1)vE_%-7nS-X7qV21?XizlW5T9b_a?FlH_Wf|i{8q$ zkJWWnk-<{f#wzJ#@V%KonjLeIT0F`i^fMqTwD2IytCi3Lov!!VdVaTRT6{!r0oq^wr_K>Ur@Wx8;So+HuWM>+)>N= zA+I0+7fufN;r^@PANW8jFDrsVeQQ$0wj%(|$~)1t(!Nx}e@1`A+l$-*!&qd5XLp3^XBuOEu^KH|@ctzgt`>Bp4c>UTQL4!& z>j#B-;FF1S?3hECcAE}Ig&gRqAjbA#K1&XO10@h_S%MFhb$GvjJr1qQ4%B}yF6YEe zge+ZI(innauPVBrIL7}Q%V`!wDIk&tfkR)2YAWtg8giQT0S-I*7Ca z;J`iDzM}|7+jI~q(6T_1BBlC3bigvi1cV!N+s_MGsx5aR?a_&F*D-{5N3dV?^U7jB z#PUEQa>jtaI*sl6ZV2$@pmg@G&DTWCXo~U(SrJHt#?=pz+ZH5uJ|w+r+a%g*{$_a_ z1h!6LZ?4eiIFmGj-MNQ1wuMd0HNg7;b|W#F7@0)y^tRz9)X9ZS?a+0S*1QVmq4!9j z4LLeDA4U5RQ;M}2M#ZK@Ld0o`CZAZBadGyQsIUH1*VQ?W1f@pJ9ApPCkNuU%pQuWs z*KFPOMAM6_))Psg%VzdOv{Hd`O_~nZhYwBg`S-vG#->sRL(CxEignIrZSAQKm{iV_ zcwUW1>QhnZak`?u@+?$}2?7C`tF$%~aJ$r}R;Y}=C-OEjsSVAW$gY@smgbm7+gUwi zcol(hFE)rq%<`RbNK;1rZxqHY7#SW5!4yx3@0nm?`!&c3&0+z_=9-ZRUiYAq^S3V9ttt_mZ&Og7HF_NPn$ zg%2dgZYR%og>eR)XU1N#T_Z1~BmBDOwr>Wip>)A7J!rAheG6Yt=Z)O%_b~V=yKYv3 zRMOjcm5eGk_wPhQOcV)ODcMnDn62wVbFATqr#n#&)phu*`rjBPVoYiV01NCg47Lru z5ROb1@ylRRHXXb9geYngztFyEW?=Qd5aoF7M6#R1#P_!@7t%|7oB8X3vS-x*71Z!lQl zzs7KVRm@nGt#=YLIJkWE08_3`G_!rf8MMbm22@&^wm?+a+`80ioGGVDrPlhqU+832 z*+V#Z4;pRp!!JytzT1(Filk_hihnFoT5V~NwD5%Nz)Y50v-gMW{=w)4#fzjYEtBcGN?`eW8V}N87)b-=BUwC*vq>p+J zDp{w$Y82Dj2KIz>pW_d55d3|p-9g8`+1ds>pVS%it5_mkTgujCt|*~vZ_UTL?_7h1wnB^88F5QL_h+w+Ujb?FIwQ9^y~hH% zbxVB^4T-s4y#RDBk%J9On)|ndR+t7Na&tvH>rXQ?=S)j1;d}nw-v)R4h=O;~v!_}Q zSZpBZihmLiU|-PjwIoFkHSc{bnCFI0Z~cYmg&fabV~{0FDaq#?kcqk=x>wYqPE44| zh>F1M?2ScqGiA7lR5Cw>uuNOS^Y|+9(B3I9c*)}DJ$J0(-&AYNY zNvrfRjR2hBDWiSK$O({eGS<-~1pEZ2Qbwl;WrohypCsy-MJ=`A#uY>v$GIIPvwDze zH`R6B*zea=c>}DQTRWQfc177d;-FL>8RcuxJqyJzOT+~5yJm1VZ)EF(HG7Jp{djIr z4|49jS9Yb`6qkZJAo~SXQI^l+r+trBx@$z%@Brkge_?%c*4Io%Y2|AEPYINI_fZ{@ zY2Jojy+_Pc66fM&bJk`i4Nn**0|Rb{JffP;%9Qc^36Gp8+blG2Wov5#H%CwdPHX!z`{-Obpie{|;-im~3&ov}IQ5h1q>)j0 z>HAN>FhJXiK7BkDkwpUG{|o&zL(JhkJw4q&ljiuzc|9B+d^{iglg*?IZp=C*(-x*k zVYiesu^;hb0zPwjXTtUP^?si|rZN%NCbgSvnfK8!_}s_!_Ifu1dU|zhirmZ1CX-jF z{O@yfS3H08cS%Da<2G|{+Ap-L*~Y~sKRt$B%-dx$eO8;z_2*UR_ti_(x4-I>_@2d; z&cBraS~58|PNm{j(}5XIRNewZ{55i3t!Q#^nN^avpM`OHeyiaU8}$O7z+1}lpdRkU z+f6YuB@JZbgo$*(*yH&t75~RqP!Uve-+k>DjYIqMw3;enxfQ;5unoH_=i=@_CP7oG zer2~+*-Ic-R7C12w}7&slwwamIw+u*`fIz_!rJq4_%tjsbVr+atkS%}8zI+hfW_7g zEOjdKXb%`MJZs6tLku0wTa8Z7ahO*~a@lJLP~7>#?%)7y2LSSPpESC94)rwLI-eIc zKNKU1_pImNis#Nf#RcB%_3dLS^lL&V7JHWo=egb0_b-rQ@q21I11f_!oV|+?q!@yC zhO_1i@`v{x6695dQaTF+^f(~SX|-vO(-Kh1^S1P~51wCocKUH6x4VC*_Z%qphB%(< zTxkh#Mdkx`8%gk?OL!NxpCSx{VmZK`565+km*c zvI%k_nNSCAwunqYQ4&PrW&${KiV{blCkCE22)G!YPdzkHJG!3a^-KU%=o>&T9;%Id z`tA_L7-dU*ESjd?l(Y%rQH)1%0SiV41K|nHz^gTbcz+;FQ?%kGI&s~*9XJJs52!oK zEd>}{AM2Q}gGFO_2tQ=LsSn_9dq{dgDkqq5;)Y5iNcs#o6q_jbmV^YAYr)iz zo1%Z8hJO*0S$+f3qIYP8{xhcs@J5MAgO>j?r-q#6*`TPS;K+1K+_RA@#xR7J#jZZY z%`s`&N3hHcRelI0kO>pVSoB%m#xTu?!+$-)o7;Xx zu|FmT4_PMW7%;45nI*$orLRge%0e~?AZW1=L4U2!`?57msZ#|&!91~{&P~5U&PN1P zqS@tm;80V$ksfE+DC~cx3VR}pQQwqBo6BBlnXtnOM%wzNoOcj7nG_d|=$LbZenP#5 z{Em$qnqceCX+O^{yB~C7V!#m!3knn5q>Jn+&jVo_N*5#{G9qO_R>hgh^Q!}<{~ixS zZhx3dW}Kbk%TPgro{Ug86_7n-U*9K!L_C&H3R=8fBy8$#zZ(>+CJNRNVS{nran0wE zZ0rzU`Xyqz#Myz-U(aySnK^Nl6M2Bk5f4EeG3Vr#;mt)S#}sW(0`4?Yq!GuR_*p(m zeKS__ajbyiUq{8?Q|kpGqOMNa#hVxOaeqO)KKQ0CDR3i?SXoNSk6bK9wl+Z|(OYHG#I(UHF+Wq)sc z+eV8-cLIYDK%*)4dmn2FoND!tnH2{oyyiY0HPWw2NJv z1LFf4Pr&p(4xe~__2`bGDZ?BOB{eR}pjc4X*T5~cUch1KY@ngcGIt9oW4~04di~j?tf3&1-}GiPa(}P?&10K_@@wQ>+jtrp=HW38M8JYCTrU-TA`IfP zo8107xqQ^_s4G>}LlPJnfMDUKU_|~o9>s+!0dL;E_WROq4ig{~J2OFbW+Hlx8iC zc>%0m{*Iwo_cPhq$)tPVV}I-us{?Lj*bYIIW>8p(=xL!Ga3R#OrS3wm6XBdljlc75khc*}k zhc*}lhc*}mhc*}nhc*}ohc*}phc*}qhc*}rw>B6HWFbnje4(VsWv&+2uNQePv@(k{ zwL<24alKwVPqa{CxpK}Yzy7wgGP!>8+pm8tS31p-$Jg|I`1Hr$o;|t#>6gcUEMH## z{LK%k9k=8;3qk>Zxd9eO^fQ($5UW5r_^$?SbEE&8k;Kg0^7%?hSo^FO$xNg=H7Jo@ z0_&wwARjMRUMG*>ze;`zZR&Y;YyS|NskE;ExN$jLnHEMf*`F?zm&xV+)&HuptzX0C zS`#G8TAh?t3%7w6mTQq_*}WKrKp~$5SZnSWhy?8wt{Gc@C|Ef2Rnb&KJWNs+%@SDS zisuC$MDz+?+0@&5Q`E;O`#Kk(x<=Sa<8Q;ZpaH8u7<@iv_};E99i#oI}#?Dip<~6?&+A_1ucIRNF17UiYNtPVWxzEg>yxSalH&gM=TE!OF9Rj zz2BA9p{34$H-xt~B2)9_luv7fvGXxIE|mmLc|{v0iOPnv0%<`He?Qa){bSw^4 zU|&H{o)!cZtXzZEjDC?_+^vgkwQGwF@A$5(X}IBkj>Eo*NTJ-&hG4I>oG|W5baf_V z>hEO;VTat3=wjmmejFmanT%|BYLJFghf5Bdj(dv3;kZ9AR}(_gf?>E1S0fxY2e?N; zdyt+8V^cJUZqV39d@6Q1sM6|;?uSD+VyM1NV zG-&Hinol`5algY>$zGVU7>9z=2?t>b8p@EAswB zv-==uGB^=QK~35zX#2WBzMm{Qp;VCil4H>$+GIfT`t@tF<{%@vnN6N!g<3y=XsQi= zVt~zXwV4I)*I{x-IMt_N@lA`c))zXiUAHGDH*I>jNDE_$C|zzhjM&XTrH`q@B+>Px_=V z$hMhGCO^`RrLeP^?FoI3Ls6l*yrN@&pm^bzLlf|WK3&kA< zHIi9a2iFHwD7Uo4fQ#x+NgvcVvRAN$|NDB)GUu;7QMnlJvBr{9sn zeeiQs==?=V0%l%w5Og4UxpPy9EIjI$?P0S=f8rxiHg%w2(%JVxRE+#&!L3z)cH&y~ zVx%P;Nq>qI8Mr^x6e}b1q-a}a&%lUX8>I?o1%}62-+3wEI#>>%T_Yj*$i?OOpo=9-K#Cr6Ae*>*23&!+&=!^P7o$gniYXN7{fd;O@zr z`g3@5WmN_<{Zs|s&7*#Qvh!EuQUAqw^z)IP&7=Nn@aX;VzaNjL;BxsP`rE1rUa*mh z>;z`Kqrp|+9;jq8x~`kH*p=KH!w{ShEGQ=9Z+ae~zuqCZ`hC6bstviBykKaJgNaEs zh`_`_qIyJ{4lFeNURN;{up@T{>lA8!^hbVgOev(C=}2QH4Pr@ui1_E{X=7TXj`(ga z>+NA5$_fdSZisgIwxH?^^xYsGPvyLQ+NUV^-UtZ?-kY3+CvP_UR|R_S-uAwLcjPZY zS1dVUdvnR*6+}61z9;go>wSVxiwm@eI2*tj2n=!v0XQ&7703f91}h}^MV}TIfXE4K z8b@19;(~{qF$$@F%)n0j`Zy~TFd5efk(l3yYY7zYQ-jI@>*n(#wU8TR4-o)`fD1sA zUk4qBebdy@)g|{!gGzLE;(&EcDD|aT&sXICIw`>So6Wr&%-(`4S)KQ+oEHXb7Bu_M z^JXwa5*#GQjDvEa5HL5uz>EV2X_SCWvtdPoidNk%*}8sz#VN%E0hgSg5)EPc+Y z^6Jm7@;*FM<#nTw?|V=H8s=d&_xDi8z3yweYEUH}J=_4NVa%zkSW!l&-RJ00IOoR- zWP%3VHNsmBW>Ovb-(hFM;z{`bH1tsWk7>pE26nlB$VYNCeOID9pa^ctxk_}+SspK} zb+glf7?3MPX<5Z+Ei6Evt*LW+({Z4@~M)%8p)c5CVKG&V(2+k2ag13E^r{fJDAAHU}o?1NFpI{7UCu zWGWkfJ7+iZV_;LV$-|DjGsF$@0Qw=h4rvvWD|{HF-;g^Gzck<{EVX-gFIVIO!Z#KW zcV&Ixt~Gz-;~5Ar)qIAbn~ljS0JA6m8K3^(OsUiJM;i=vw36H(Xb?>DWXPP}gMGtk znKpgZkbEdm7c!stP{5nJi3f4>qQJ3qlXTsG9RkoNGU=5~39LFTTW&7nr7;cFT!Qm7xou51I5#QE- zE57XvoS$j^rgtTM4#DL97@CwU#jXL3J^PxTGu&ZCyi3nH*wD#NxbL&bb#TUEi0|8T zkr{f%m&LI8NrT5nEchg&2$dKdW$%Ad+M_oCCUt->gA3Gy4l3QmNCF!1Nevj4Y<{Lc zRz?Ge^@&&UC$MfH!>)=rM4I6*<`igu6-QPzu^bw-i%TKgoLxmP2vm1tFP3L~ayHSj zsEV5K2&4eDL?ui-1RAu|P72(q8(e@*ee!)p(?aPGMz{P-B^BOJ3O`ETRm!8)#g1NT zfYOeqo}=iGygphEpgO`|VKfMOKyox0dADp8R>mYx4HI{r){QU~( zQSQ3~{vvaUy-mwm60R3Pcs>7(g?Mke^zxeLv;_7W|R#JqiNagD0>navuV(lu9O(;^ey4kKi z1yZelx%tc0+a#UNRH+U{Vt|mF^ybg6KfJs7c)i}JI0;m^-b9fN-u(SwhwGE~y8Hj^=+0JLxiQUpKO#f)9ow`&;4diQ+_`-s!U zq;E@4rt*z_H9TB$nuf8|i^+OgMaPFzWi(fVAf<@lBO9jL^4YqvI8=#p9OM%mc&zv= zgsRKFY-(Tt65fAa4z2|}>#o>+St|isqMmI*b6Prm6NiyVjyMTzALVpQCj)fuhmr{G z3y`mhf?d(~McvaVY|#y}IO^^zs!78$Gr!csa_9pPU{eill$27eKr>dqu6*lvzf{nTv zHmvkf3`Tz`kCpznHK;{VReXX|CAfMkf$l^IHUhbhuI<}l4$>8FiU|{=ab;G=H0?h$ z?O2=^#^q`<3;mjtAc^b&YaM~%ThIFU$5@-^S;sIf2G)sAp0~QpD-xWl5kX(J)PY{_ z)G{MbPaXMX{(6c(%&^MO!0kGMZE?`BC<6=ItMaREi(TnHOW-p&o29O% zElATvDCF{W^x3{3Ri8$+`4b$5cNgdZVmI8~nxng%WO*AFfO4&k~CDQfkTgh3#-K2QE&;6h&AbrlG4lV-FjnH z@V!|3|Fir55&p~I`jo`A4nZUNOND{RMNttLuu!4lRFx%(ec*Uq=jFcen#Kx3@ej`3 zri(>Fk&}J?Oi7sTM{+X)48p3V8j^i@LtlRuk@)nn%=Cr=!fSjCvE+peRa%y8PiW2fD3|8%Std%kZ1^4?pO*loq?l69oS4} zUOR31hx<=j^b>0iYmot%LN7O`lqu3-WJW!t^`v3sK+%hr^I;89>Gi0`%5_Vk{y09B zN`)F?(?-L2SO%yL7;38ne}Ww)#!qVivv7A+P)GNz>pi4P#4%DlrFlNU?U>HKZT3ya z#0JlF3}3E08>qKN&x&Ercc9w^J9|aba)eK(zTY+_tx=0Kw00JSrFe9aJ$q0PCWK6- zUC50=WuG@$x6rH_$OD7pl6JAVyG>Y^31}O$%g~Ur#O84t@r*wse{FF7mXY?7&kACB zzm~ypNpJ$+J3K+RsJv-gz&yUcsOXSkpy1e=l@4sdXztS%UWo~10!I#Hp4a7pNj5eSI@k;(F02qZxMI!y=>Vk5uH z+FOd{IyC8WY6X42f2D1z2r|m$`IA>h=iGrxb(!t=RhegBAg-|a0x)}mV0=uF7sF2n zRMH9Pgts_ozCbdAxy#_EQ_yn!Nm!8z<1|^!m4t-^F0*a6@3RpQgLQ%e93Pj){^<0+lUyjJ|~+5QMgpCR98jZ-#wU_?_FVAlG1H-PS5}T2>PUigE*PVeXsOcpnwz1Jh0oF0IsC;*n4VD#A|gfa z(IFt4bf$>ghcVUy7E{YzEW(?}1V5GZM2*Jh1lz(}eM!|OXkwWR8anvW)DB2PU%}J* zvBcO>f5(<5+z**wo{B&s>2ip9Ovo>f6&F&vr(owSOzoR`^zcX|a}j}qJ79LKo=@=@ zB?>w@qyz+2bv!*D7$2d(R6r&+XF-2QfzTR!9Y0lKpDA3;ktx(>IlSR|G8l+F$1%B8 zIyuAcTX&g2M^b0?sN8zC^(PEJBZ?zNgS0MLe{2|xr!~4BQUn<$>GHdYoP`Bo)Gq#L zTi_)MS7(8>vlvY=;7BD%p&;FjV|#3^WJIaA9o9s6-(@cKh1UI;5Tz<``UjmeSA|_< zhNMmXz_L4kmz~c@x+uEkaoQJd1tK ze+p)xb-6-gxs@@iF!mgkc0jusT7MnT@(JGh@>J)`rF`~PgOM3^JMQdmwA1 zD>&pu*WslUq`(GE_S|C|=U3QBCGn($RurayMq7In8 zV{eiGPF{X4RoS?k5G3MIKN6!vPCoEyPx>Xir*SEqp%b1`WNbQ)ySq?qOqGjy%U~QF z`k^iS8HqTIR64_=hr1_bd?bMkROsa=Fo`7b|2dNQ3ocwM*Hj)zneSk6%D(Bkf6`x! za&l(y=BK3wA6g2jCW%ZbNd2T)jM|vMSoFX=Po^(3(knWW2vEOPFS8=94hb;qXof~y zdEVXJV%X_CCZP_#TX!*$>9H~B$mLW^0eRn!bewB+q%mBzjCE+(V`M4837X|aqHd@l zv@nN)m);FbJom(Bdr(n0!$r%se^6FLuHf>qyUI_5KDMAE9os;{=3(`{t7TE$uA`BM zzmvdW=ag{RiKKBN!O1SoC)5j4FzGqy$T}Q7w0;Q%6V30k`qp<=eq9GIK%OEppo98Y zgSd`L1)A<~g*v;n%!jHh+DV`jvlut2gko--yv$f|-I5Aju0`c}>D^(=f3LnMtZ0SU zHSo9ela{?N2~+3Y0APwIuviGPz{5c`pp<9(EcexVfE!w#lawJPb`}xg7H{lJN|7Cj<~=xz1j5#1TneJ4<#;Na37_QU4!T3Z(J1S zlV&-P-0`|7P6I`X{akDOCU}s zYOy=?`*RXOcEH(dTdW}Moz)aD%79)<@9X8f$jY5mcK04yy5H+8K+Wpeg@S`+V(uq3 z$P$#qV`X1_SyU-zI?hW@m)?W#cTN1!8Rk(vapgn~X1venIR$8*#HpL->zj-J0&vW; zO_QNJ6Su5U3m$a=HIt!CD3|lb3rk4NR)VvPXJU2xX_aM63bk@hGoH!S?QZokkxa1l z#u%HteYe&;x&7hY+YjrFbZPSXmYy%(eE;tK54ZpR`SpkOZ?``_`_AdNhQ|xx0e`7W zLCq!eb8y@TZv-Y-n}t*C{P`7mM=!Q{Lk+UFEM>-wtZmK!qP3yE)9j*YiNXK6J3rhi z#*JM}iKv_ka3T;9T5x!W_9;=!t79Ex+!j?=Znk9{9bCB3DYGUUGn?_@V&ubtLen0% z=z0Dm8-cZ>GF%eb&569reMA8{tACt&e$SgC`WprQ#*O`r_ijdpjfX~XztH;J-INtJ z!AX{JTwnk#&j6luQo?AKP6O8i&MNw*Ks8YR)U{n*1O@};_(Ec!klx1q`&VaC3VQD~ zbtR=?+1NqN`RMR2|K(-afg98s+9{VZrR_K_L11Z03ZtOXPOHV*R7by=34eW!3OW@% zo^qOO2D(o5aa%M^*HHKF20})S-nD7+yFNtX+`OhndyIoA961$^8+u4KnF&~tUaV|* zdV$g*i09)P2?(@l8x>GBa;$K50pWQqAP6@>RxPunDY_WUwFE~WEe~>yhPrk)@D?ov zhZo(KFmYR;a^L_T-mt4K_J3=UKp1VBKrl-%IuK^TW{w%mCx`0!vS`~PYA082v7)v433pW)Oog@g|_v!(3s)D2eMStIyUN=;5@}A2L z`lot4wD~(&vq%;2wwH6Z3GY4aqI688YH6mIhaKEsn)}ddtMXV!@tZulAjmWO)HvAA zHb3RtUUivsc6k~WT#&ZzvbK*k57s0znKZ$_vrLM`?5bjm-u-~fLL%P&un}DN(;$z4 z_`ZA*A=m37Rm2~j=zp;eW_=f5RgvK^35I_u9RSb)%UOxU+!N1D%&VRBzl)OCln$|27wp!dLmPw3}I07J%0gzY@}4oiiv5|7*g^I z<6up|pbAK#cLWxKdZw0Ye-V)erl$DZA1OXzdl%ldb<9zErV*;-q0zu0c)0lpcw%9J z;C8P@C*~iclb17y!E(mW=;Yl1C)`I4uYJ#d1Af5Y4;aw1h)-Arh))=am=P>j9q?N!@f{*tW&fHDk+q{15y$! zQhcFknKI7RVw2x?-a?S7Ph^4E_U|n$6LE(T9BjfYg}f9uo|#h=0jr3Sihwsw2DE0@G8pG>8IZ zd*Un$_EKJoEy$CLDx?H(*j8UzS&2afIP^uPh)3}4n>cLnb|ivGo<}-^ft$b<0>0Df z*qw(u)&L>(@pAxtd`#r?#+f!aU$q<_ zlJETZzJJ{QfyM!r_msLM=h<)6d_Ji-NX{T3g7CmBIL*LAy78!w;2FWaomU}MaOWyI zZI$|SQZ7OG>-%ANNzPwI9}f*O`4g#BrWm)^U_!q+#bFz%0c>hwoRTVAOexOC%5RX5 zo_*kRup`zkE?Vxxd|%M;El8j$pGlBqLLQWXRez6;oXQ4Tx>zTL+(g+c$^qSuz@4N0 z#m2Y*fIWbHsWuz(u?&tnyCR}Tb44WhW-TmFE>x_M8Q}|wsR*4`w{5#v?XXNcx!i#e zp@8PG*&V`WLNeoQHZ3)DC#sP@`%XY48P?q_lQJDu_`)zbDg=V>6mirk4iUp428Nh8 zDC_v+2nVmaSt+sU^klJP z0WxNEzg8*^f(X8X1y2Acu`F;3H7V@9=|XF&btx0>Dm>|_Q2KgLdeC8h-oa*0dUT!o zTImDVh$kU|W{}oQ{B*!ZY+AduOxa0&VSj-uRo0C{5A{5!hc6}CL0*qP1@%?odFjZgj^xODD9wyro?cs-6Lw}G- zSXPRe{+esKHVFs&g!u_H#C(F{hdlRB;}y(zoTIh>_tK@LC%cPOsyq}12AbbjEJL*q z5)3PrJAJG^=0o%1D|DuRu43z_*}m7ePd+~pp+j=b98d(G4?c9Y-TNtCKVeJ3dxGAs z>SFcj$$u^|7~F-Ay$araS!hbE(tl>b!~Xo{+mKxOayaFt21`wHaoK3Iqr*{I@~>!8 z2139UK!7}LMz`t<@yR~c3>jk+wWi}gY(pK$OZLE-^cpFou|BP*n4hv#sYk3~`5an9 ze><#{nw_BmbW3-p^N&iof)M@`+0~{y?^r_j;f0) zkDoY@oOLl7a+Y0jK#F!b8XeWcVc|}sW2^oC5!?UBG9Oi*s5)fQG~AJ43j-TI16=2z zK@-^dkRc&l$JsvQV}tT=I%~n~@|*i02Br-7>^&L;RQ{04xC=Fo4 zR(+MB88MU>u#(20eI9lRFJZ;=3|GRnXUvPSvCW%VS-~r$&OB~+?ja_?ID=6a;0v4@ zM+p9y?U{VLGTohjkTxl9J*vV2X=?hsNjM483e55ZAgagZO$riYu3!^1ASz*}dDEfj zzg5C(`CUx5<9z_#wYOB#3nu{t)q~aAlc75kmy**97XdMop-d=$l~~<#+cpw^_g}#` zQ!^2QFM{N~wLT}!UD8V(b>?o|KCl$+@G6lCDR$cb{ua9niDaZ4JswNY0@z*b=MRw4 zb0nj;H{Wk=UcX6gq@>7YuA?AFtH85i+yD z2p8S`@ZHCEn}_><)mo=nti@`bBrd-D*N>a~e{b$RgB!_W=Q7uU)!J&IEb-g7&9SSR z;|hp2-c|nF#J@;shChk6W!mnxZk|>~#%%|}lXzbo>$Z6nHugNIpBoyRxCGvHvr;l1 z!gN2>f2?3LOEfIe#)qOS+o9uS+krX!RfYvXg2TbV*2;;0+&EJI0f{*n(L_>csfhfG zG!AFhN@$fv>(N~C@y?$D6gSq8;y7T%jn(nC?q|CD6@BZKo^7fc9#-Ct=V>ooC#gsh zHBn4EdsfU%6swNKAuz6{k~G?~YASCbmBhjmZTxiAF%UM&gg?Yvt&tp>F^Z>Hk8?r5@DX&IB3OIk#ykY;=eoY1d` zZs=#l>{(vf?)Mc@wnftiUPL*P)RU~D-qC_BFZfV^K47c5x)eB%qiOhj#rY&v{;QZ| z(fs>F|wzL}i_D7Jp|(Bs$Ezez#AfzqdC( z!Cw#%xUU2I^@#t4QlgX*sYxQM6Op=vj`xO239QUSs#AkC2bic(N8YX0PRDQY^Jn~g zfKL_QpEiy}7$YDqc&CPvBYgzIXF{fi5OHL{J4#qT^f~OmDMjJfTICp>`cjT^bn1DI zdJZ9fpTG=?BV2zY!u2Gb&d$Xs^n-B}`rkh6Xu*j@iG48+z$&N%{3|eb2jm&)A=F+a z@&ufXwO8g;dz}(0vtOH1jm8!~Ur*&Pk{dlj*udUTCpYFIaTdLpK(EnBaN~c_>7UT) zo3qo;Dr*Hh!Rx}!it3Clm|o>*`mpdGS%1KP&wCW`>5>4WMe55oZTfJ5HH08^x3Gf3 z^!*Bsw?VG&E~>5Jl&3IXqb0w)#9NC@XJOsjld?Mgk!?cu`DRB|32=cktPL2!u~*+? zwS(@}|J5HUa1|t>x*q)LG;sHU!5;h1?pUvjXXr11E%hN7ZNHMFl1yZoTAF8PQIHdt zO4kbue}QKX(3WY`QM1j2O|!*W*tf$L8{=5tL)ZT1EBB6SLoRe~U#@f2a0Bx-Rp!$I z56$DZx+xmWZH?&V2yLqdLzY;fRO~x|mGQ9#xTDLOnwno+g-!gsEOzCvqfLMfj@&u$ z2m`lzDACd`?!JN;-%&|YX9ofwIQw-hGvl)-e+HU^!n1;dA6hDyxfV`miGT zzvw9}I7Q@vs%ehErzYK9>a*~AWX!^x_ZP#lZCW}=uH)ixs3`weBRk-d0 z;mfd6aZO{g5)5_)eV4_dC|Q4HjZleN9NNTTTan@(t)O+9`pZytC*yR%oDv$anFgs3 zf4OrL^2-e5ugtHzY>H*8=VAvj&~O9WR-wVQ-L`ES)MDD$FgiHyaVOIse<=RD9e$rS z#9}AD0!DBUK(%mZf%pK}s7Eyqw>+PDwm~lja()Pzf8g|Cn?sC#3nR}C%$}0MsATab z*5_}NVh$^fg{qx>1$@)#x<_aszJ~xHf1yyV5n+0M_{He>n^TV-S>f>L#fUDv6DfeA) zv!5-9Fe$Vi`l*+SLx%xJ@MMWOe}b9Q`fI13-rWKo@VCLb^C`zw(eJ~gBTb<)?}+LH zlH@_?nNwlrD$t*5SEz+>`8~f}cml;trcZr;Nz0&HG8RhzzJkT^bJ0AHOECp64El4; z+zxdQE9NnhL$D=kS&V$_+V$Aw&Iy?=4i?ZP)9CSx@AEn{Bi*6JG7Sv|e~ZXIN~QSH z^^%}$(O;RK*&GNx>p>RjDw0?XeP7c==rmr*+v;h}a-6?8&R#gMLgR}Apk9AjHOhI$ zW!3kHeSuuI&rN--pXk@9&cJcQyMdo+uy(_suLf?&n2^xToidAWy(|>YQ;|9Zm_2PA zB&cxFpvjSyhL&vdbQ%YjfAWFC!oM!a==Ug${&|shzwr+j;i&^4wv`4NuZGU+Um}n$ zok$d1iJY)dyt^uT!iS7Y0Z+_Mj#>7l157zLipv*mAn zoT}4uaoq~1gwd(JP>_wt^+LFC*R}_``x(GDdU1iE!x9+Kxg*Ttf0^Vqf5FQK!CVL+ za6R2Cw8&llB53QN67hh+zB`wA7*De^cb5?GXzPGkou{++!LWuj(f-CiyU8&Y;cPSK z(hX;lRxig4m_rD~jH72llU{k&rZrfJaS`SrP>PywM=A&Gn3|;*wSuV0;GYE&hl1&NJGuxP>>cPW5NO_&k{87=JxSm z9AdL4>HKn*2qtyIsPlvyhxP298KQ_A$kEm9-PPk2<_B`3CPs!>f7K9|0aXl7e{2wQ8xuIc z{5D21=7~@idzSnh#Ie?Ch*RrDKdxp^>3L;Z(f4SPk#8hcWZ|H8Feo;{mUs9N=6ezkG&nX3Dr!0a_W$LsL z@sw@e5F0Vef#()7yej4M%=mU)wP_D2cjJav^gz8-Q61vXUBi1GT9v}u)+HYUAdw;g z0e&kpC}57UG93;}Gk*oR4-Wt@*ZiS@#M`Tao+k!k1dD8J_0#TYe_jugV8{iW&c@#B zmF*K|pr}Cqj2#@w;fA=0*iKqP*LC#VE7jS>vMk1k^8Y^Mct!{4)*0_BOUSqe3UT<^ zM7+efq!>=7gLan<>RvIb4*98$I0J*#1%;BQbVN%35>`s1E+}3gyL6);lar`7fpXK> z*JEiOo=tU(!+^Jp`C(o49kDn=?Mi477qGe z;|G%pFR7`v2k*^#$DhK~Tuznpq7NCUWLc-0IMSp~JT=9;@cV3v!Z#PtO~W)7KkIV0 zWimT@9yrH1e~zh+3w*o;B=;{{yBihxWS=4O0bC9}YGAR6{SyTgM-WgfI{dbhKQIYG z<>U1q0Wo7h#FvPj@H~_<{zfU|^H9q8n}?51zQKc{;r;%w|7_n z0h&0x>XX4R6ql)33^oBalc7u~mk3!5I|rcIAb4;HfR{>H3@m?rIFAQ%G>#$Vv^Vc? zrjPl8v#jqb!Fb$_KH~!9qsr;sr?fHMk)gL625Y;@lEf)=7ocrKCaF=ZSG{$X@17)J1FfuMckAn6@8T(mm}ohjfl zy7UCq*x{{C`@??#Jqvnw#rVL~gP0GnOoV8}OCkhf)JGByimJ-Sr-UmEy)yd7AaI~> zWLTdtSZ0{tMbhD5J_`4k9W;c2V?mqs4lV@MH$eA@Ct(AfGYliJ30a2@ONNYYa4}@q zmSKlbV0w;hA_eL(P9np29H6t;LE;yqJ{@o?>6M6ElJ$Q8S2A7yF83HlNCE0{iYCRP2apkhHTbZ2KF+cW6?GNyxwG|g9HXMEZt*FP1XT2 z6w5(zQXqe$IA!1p)1!dF6@@=ulm#jP9wuidpaKsSIF915P|_=G3Lx5h@}yT?e!Zzp zbvzo4r@iW8_HNpozZ#A{^{S`i$@_Zpmed=stC!VTb@JBXgsywlc|Dk#x6wJP9>UIg z-~n@O39|+7vhW;;9h)alOm$(Z7vsyZsZPy5ZfAe*Zt?7JJhE)@^gqo{KlSj}nKC8e zfB9P%N*0E0ql=@?l@YcPK6Ix{X6th+yEwP3%c*(m04_4;ruy}_|6sZ;JHiwhH z=g-%>WJ_seC9o}Ei51!BxBj>1HNw@or^$)kvAXX)`U_`St5=W-wr~l&QrN<`8sM#3uL~Jln#T zwoc+re=@b)9g0HsHezbMiX>;{0X2Btp+L0vPI3gF1A%RIsI5-2c zz{>EXJzp%#F8Cq_Zw0>hkSQKJ}F5GTL*Ytfz)3T|lJVw(c3d54#P*#6Q zhfx*|m)rw!tW6r1iMONy9Tf72?gIQY%Jy|4cQt^Doqh2-5#M$!NWqs*9&)hYHXW>V>tLR`j8lDNn@vJE z=Mkkg(Q3dP0GC4i7(D7s1!~W<%d+=Oi`Y4*RG99!&whJ$!Q`UrkUc$A zK`tmHylNCD)bJBEBv{*-CM z`z3d~xQlcf-jRO|aPk@{T=}}2k*cyt(4}Q17xJ%V&b!t zmzU4ZnW%I~#V*7qoL_`34h$hN9}>3gvG>Ax8~Gta=g@yjHb{au@cw+Z3eZ*XuEO;( zq*c-e6iav%6nBzVUKD-iJ4sz^W2Nc)!T{VEVFjvoMzn1$<=c5aw_cX_0`|LdYt*U zePRsS0e=SS74ReQ^QZ7X5C6{wxFA7-jZfWZE)TAR@y}ZSz!!11dDezGAU#)t z84*?gVzXYOiWv!p?0^L|T?S#Zh8Zb3>O5pZ;(xc|73Z=li4vJ3B=z#*SXE{9?mAzS z>gClE3vT*-vC1pd!jE9hYC8ZXNxVL%zQGazwKW-5n6=I1==mJ3H_Ny%vK>J2?DsRO zi6#R}(PRWN(re~iGJcLI1=F!9_hl6g3-Jk-aIB(nTGuu=Xy-_q8z0l-cHLERS~11KsqoI)kQPbmbL=pj=sYV3gO-4zJFUVnh zwjA@HRC9wf1nlq=nge_$b6^12W}~GvY8w*B4~K{y-{2M`?5e z{qpS1p{UB(WSN*H!1-zCx{^f3RkKX{(Yv-kt`7Mj#0n-#l9I3xqUFHl`|%>MD3k0V zwcuQ_uqz@t3ytk~kG+QtDu~3Ypm|lB^_yL+TgGRbr77%#6&%ZbP_5AYrXYqk#W8=z zh7JL8&&CcYca0@>QKL0@0%(IF2puo8{C2C!l5c%Oeoy>Yepd|wI#@%KutT_ zf{H9s%+SL9O%n(KbXLbLi%kUv?QksWX+^TC-N%AK6WtJLXmr>#S5yA8Y2pBK@Yo^? z9*fyzX3D_<0T;k{t!63w=o3`XWRM_0%%+tL`@&NgU60tFm~8m=vK~sbnpl5kYZiPn zG#8XDnOamsqX72GOlt9ZXm`MCL1^AuP6HDSY;87L?t+Ve9mUPsB(NpG-9alZ-LOu# zJ{<>l*fWcB=q*uqn`l$>`rCRl+7bxGnv?H`ovBhqChEH{%anH7vw z?&5kT00W@oCL9IkI_rlI0Ze}@t{VOWpa8^mFPx73LzEE+B`RF{@li(dG6fLPw4yT(V&6i5V_R$f;aY}qX~aJ%WhZ<)wP&L z*9JihS_GuxQWFeeGY^n~o+u6w-0_$zfSMQ}gz_>zSNR(V2FiVYC^yxh_y`CkWNrRF z4jDHakx65j-Vdpz0oC?x`z3F&BfJsUe-0bMkgcQ^j27}VE{FoF11+7x_||fJNe-x~ zxzi*$0AP4$$4fzqX;puSggd69tQJUCgjjD-IV~tx&D{?}s|DL`OnbGu9pM02K&QXr zTf|mF^sNW_pLDyLHqZ6FWM>vmlh<{3Z=NuBLX3`WKBxJItc#BJ5DV>ahH94axyV(^15dR zP1~1$Ge`xKEu$s)bl})CFxIL*PRY7*?!n=MzYx$cH2ZBB_DKj;=9pNMDff62f`X50 zu(2FoVhg%I9^{h?kOm}w)0rqL{~Dr{0VUaC*4x3IQJ|-;%dB*WdqWtwQzr^T*Na$l z%EZ2;*Z_Lq&A6(|UkM(2PauOL*%^E-&PK&t34sb+hbH?-7;|C9pPUsO6jLZ#aW^;zBV%gS>10a=eH_Fh zO$c>A8_;*vm?dq0wOU1K+-}NcOb@CI4+y)UhZ2oi*1&|yU0!+z>TXVpDq1J#rzivv z2ZhK$GC{roxJHG*yt_A_sS#lAK_eW{&Y&*qYXr%)GU1T@o@;N*b-JoRw^>GrxO95@ zG^d@Q^jozcOn^1ZxJNaP8Zx{h^Gr4oN#hUtF^O$bNUgkoCjb`lh0>m6A!Rx+QIc_? z#-GPDK_XG1>oXd*ox9OgS|=Cdv-z!&Q}$Z6usps|TnvS7TnrVRxEL=8czs_>6ystX zrra-LsvJCX&!*jVS?Q1{AZW)Y`7J4Xpgjz7c0802To$`@$SuGI^=|kiQI2Pj8OyjF z9q-f8e;eF?>hodYRX`Xorr_C+`ah!i1fn?!4|~jfm@E>@!lv=b96D#w5dJmQ!9MwW z*8L^bIq<>i^XYv(oXut&eIAUFML5H6$%=rJICb}LBr`g@@a^z`oKjO?Zu{|4j&x{R zvu20HmtL(^cP{q)zqQyNhSHK5XbAS28{Y=zdI*?3^KcB3G63Xh%ii6;j(rX-HtC>g z0i`l#ebSY|vxsOx=}dNNz9UY3@m>EkjES2 zFnNe{vV_M0vI3OWXngQ+;xs(&K>Q%g5C9{MyAhynV@winU}JHcuKhmjM516C{9PwJ z?k>n~>Pv>pf{r7qlFp04@_IY=@>G$|!nfBn*L{4)Ec1GsLZs30z>9PTDqN#+c!U&S zsN*uBF^FQsJ#IcO%CNvdpdf#et32IE;y|jG%uFW|O4<6ljA~tf^I(X%yGJHIbnhhH z<2?$IGHn`WP$d zj49?L!sQ{XraV3rQb%IXp(ayB@p7oz_n1t)ed5LVJCSil6+OG5FV-w!kBjd0`1k}9 z^&cP~;8B-72SQ;!Wa5#WRPbY#%rh zJ*G#Z#(E@ie*Ja>-~$vnqQ-V5yN8h|5Cj{I#@AnWV`y@n(B#?Kcb8|6zPD+@C`oC` zlFRo=ni9@LVvHm-<;mqDc{AaJk?Bk+J$dTL@KQG*GuY{j1~HOZb$=$|UD9`2W-xThCW~Ub%gT8^ zWoklv)AM}mFMgYHHCY$iq6`hU`QEkV<*wahmv8*f?LJ)W%iLeg@fJ*Kmlbu>GTUy$ z_3gTPBHQiY3WO-GrksCH_V64`A`vlLm3qR@+jY5cLo;b$ zWab27cL)^j1zq5Z!}5)NsCsF(@fXVnMcIhD+2)n7;dZ_c9czD+aKGMCMg8kFCB}{y z`qMt!73L9`H6 zudM3A+FcZps@^eAdPk8&NnLfh>J=v%1t(bbiuR`_TG!(I%Xg7B!UdA*G~tve8|=(F zz0za`4_oWo{#}2yaU#Ni>)rTgOgPP$THsB2FRW0l&q2<8}MOw2& zTGi27XBbqQy?cdbLK4di$WDpiK`1(2C@ltMdN^997_J%Y0}1xY;ERjKl*1kg;6nwG zRN)1@GxfcmPhk^I9=IN`6Vx4$W@1Q-He&x<1kH+ z>Av0sw3XHt_{WfKNpD-avoIp|OBAfC)vaw?@qm8-FM==?K<0KpFkB)@gkZ-!?^oDB z;32@Q+LQd)tI=iy>Cry=*p9ugRy**`C1Jd+3Xru6+}fd(HrejuVMgcAz2#P+3upZWS@_s7l zWL?fZ+F%Rfmq=+VbQL~b5g0ARm5Q4L}R>axs1*7e_2nX{QU?mofP?SS?>IZu~qn^+xDZ z;zlv3z(JBS(=Zf5;NN&>{y8u|{sjp3Z5H=$XV+BK+x;H{Qx}bb{(8H<|GIy7v#Ag_ zahq`9_f?+M)5)}Y^`YNLa#z*xQkY#D`!ip4RGj(Z7;^gpiC~$#W{`h)=WkzQx@5E7 z-&QwVWe@Mvl|`0Hh{x}!igiMZeCsQ%erYiKNt2tTy85ZNEl!rgj4&(1(I{al=PUn= zr4SCJSj6BU&L&UZpZz@WhE{nxwh6&GpUk8b;YbA*2`c-w;z9e`e~crG5b)nW(MF&aEM@;(T8faoV!Fxb?+RazFieLQUE&M zadXvj0ZK??PBNmyIuRwb&Sd=33&(&WT(fS;Pj|XdRY%?Lhnx%-J-+|0oG}{!WNAg5 zP6Y*-$>Yb5`{9`xVwpWLUtRW1^x)OkeK2l7Vb$BvqJ%JjA=`g^<&Fcmcw1CPh4*|P z%$43|F5ZW>^Zk5TEQ&mcp=mUsR1?*#+v~m=fe>I(f^@0|ozoUXrwE&MFXAOR-}Nfe z!B+j+r~T#5J27CeOgS}q;%mK{Q9&X z0Fe}+mazi@)D=dicL$qb7edyGe{-C${W(x6>gL671`7Pb+IE8PUEgs2h3(-QWEe>0 zV&<&hYFbI^dH;T_48~M-dAB<2kop6fl_IKBZu?V8ty_Qlp`d{uRiEfppXu&%E_qv4 z$Bj^0S0O@!@r9W{kFo~+pwAUBvSJf{Ajd68UbBfqqvsZwf>g05JSM-6<&XC!a48z? zO(4U9$_9>N)8r6NsN;!{!*sqF%ue@)eo;217cLMP(DgX zE7%wgvB`fH*+VcF{Ir8)hg)9}Yfmi&7k$%$kcXx1^#BT8Gum)SPXbf(-#0UVaKsRU zrQDG@)>TkK19o*1`mkIi0gP%r-wa?H9%i^3)8PG0G|R>qj`L1W26it!1|XvEu>aNSI2SOT;&`RUB&#;{cO@ zp?J2AK_uw~_;Dcm7me1(MgM;eum^rT`fstzrvt7=Yklv8UQ~YtXB*XD%YphU$Heuq z`YWi>b`H=KeFaz!rlIatfNi~t6N^z0rgvman!2(}X~xu!t@5&0r;l&?YscEiFRvYI zus45n0?OXYcXoY`3|MD6lJjnfH64qs@fm=wQ(?l!rA$f(6KkWXZ!CQLO5;$ZZ;Elq zho=-2mJUhF;zh;W9TanC$M{=s4vi6d5GY5IQY2T+>Jl;8;) z#k$#ttUjGQ-|zg9=(}?WZ=ipTM|k`*%Ae~WU0p>fyN7)8O_;~Bn$Ua=bjMdd-B3KW#4r{+sWO~0X!FR{iK}&FT@BJ@-kX|udA2&>QeR>0KFi_cS$S;5O zb^N%mli$kMnU1RyO-`xvUk5FvqMrTXTLF5A&mZ&=J=WCb;pvaKW}do$Ran&t9~NHe z3Fky}dyh)HP}P5Z2$VSwO>=kdA^ceE{dbFUy57ydB0%u0PvGJ8$ut}l5Rmy00LRRC zzO|O5MWOQ&&ys}p0G=El4M#9Ix@Q;OA$Kr~+lH-Sy_Uo^f$$BO`d?HY@tl`NdkiW8 zGM90C3@m@#HWGfKD|DB^33W)aF+92B-d{e%b7DKu~svl%j9~I zyh??;`1$(ZPoBBFKT_t3ai@WhAN1M3KKbU0>+gTAE@slYRI-a%mYMXkufMv!`tkY- z37qjT?UQGM%RZ@@5Sr!MC9^DN+!~@-(5yJtvfOe%G-Eo;-~(oqrD5-yVZ(JWiF0(r zKBi$XNpF-8Oc?ksW4TZ??wi=GjyVCds*bre4ujRqEORP6TXvjupGqd-Q$Aoqa5Y#G zH+g?ELUhyj_%uwx1I09u^22_MajBWqY9O2CyCdXsrTfCUi<5I>au+j^@$}a3=GFDt z`!fN2^F$=lO6DZ2N`O2>p3GNguYTsq0$zfYm=PlR)el`ITC+?WxU7=z&%Pxc9PSDz z5X&6U0?Ekyue2>@xGh7|6GpP!sU&A27~X$RJmZoH0l#quX42(gI06g@c1!9^7Lt8DtG3BCBhhPYAf@I8Kq-09Asuedt*cR35)=}frL#2sLmNB7dvrK=Z zGMq=J%$YK>nes8$Y6i>?r$o=HlKZ2E)}vuM1*VaQK{BdqnrTq@SvTzCBfRIGDXy05 z*o4CP8T;&umi=xROy%0m`H4#%mqgF>-$khydm5E_T$J3{s3`wZ+9D@ui;iiF8o~sd zF$)?rlimo~tlsvj>}Y>DzJ#2u7v9f{BBL*JaF8OAWXI z{0hJaE!DtR%f?rq;c%eNK#Tw~vSbI~XY}a1=K+5SmS>stYg9RCg_hvaaSqYQ1DRp| zujpMn|M$rNT}_)ec{LNjar&}dU(Pb4Np|2Ad8=?=U(Ae5KkwwwF$*?l(ba!Jd~bxH zOoZSJmjcSt4ZQM1U%+31*P7=r13tQn&7)xf00;s@o@IP(GufY88hvbTb8PNU{PkD< z@AC&3$e1xL1J5JVl<20xu@TvVwWN5fKdMPB@M1T7dBmOh!5ewWuO=BF%pzD%1G6$qM?67s{u#lmF~!D z?1%_38|d8lEh*PT4Y%?6%a>lr{&iXJ>CJ%yGN_*UyC?-UVOG^DEotnRR@@z%7ALW-!AIf2od$TiXm zxAk(h#dn|?j3)SOv#x)aVP`YN1O$VrjPEW4YSErZAJ!#}R2EDaqA>|P;MZR+WR?~| zCyUn?l6%FHN`OKDnM1W-E|x1`bh%lt%c?4mL>jOH;qHa-QO4?>^bm8KqcG@AbS4cE zNZp%9JqQ&kW$s(X0=PCr&7*P6&Hp)2C0L9*N(WYPx?%anTSHE@Fb!%z*Rm zjk<|NeTa&AyFqBls0Mq|B)EDHuIg}JuJ=2yPzH8O+R3h>MCHCDO5hnrIyTTlPs2GT z^OLL<=y@-L7i2jK@=cXZrU774u4#4%*cs%TBlk2X=Z~&jVdy*>truE zh=W-yc*$|Jz2_7lmp1BNnWC@a0_=uc1hYzqO#I7O!ajp>5dUVxogepOj2>qWA{=N5 zYm)AFvHA&SK>7pfCt|d^%wWQ7GusG4s2Ep45u_hVc|#_SVNHh=iX{FCDVDi-&0^T01=ft z)o?H4!Gg2SgMJz18}!Px@x=QpWCw#T{;oHx*Sm1N2)`}D!+5|SKt(nJHjqFMY~bx- zy?lRY-YtJxiGWkS4A2Z@YnAlA&#LSJp@f8^yeX;@KmtNMuF_(=B?z84*b_|-SUli> z8!|P0PoNFr4%`E>pul`aby%LfOPZ@MqtAwATLq3Yw~2;?+5-BRulzRv+cYc;)+YdF z?#krN*|$d~INC^%pjgrwAWg7^QnweMl1qQOEkb{Mz}1AnRzI+iu5&Q3t{l`{_tgAE z?=k9qj1dT-a<0>=Aknu7dQU+O3o7P>AdIwpMgg=aqfkIX%E=^}i~{KZ-e|F_WyXW7 z9(F~ZW#-a%Gs+oYU3_}Z-~)N)I%Ny@iv#}l@+t@vMyu0uk&N;x{0(`Pg!EONDh2$$ zKHPt{t_m=Qm2N64K|GC^=9~3m888Rjf(vLC3WzRD7O)Ja%6dBXYOoA#mBR|;2soh| zVA0^712n}7DH>7Z2abn#_vZ`m+0^iE!BO9E4P$8NdB-nj3QZGl6ij`LD~$zlrU6~k<#&ezYp~!4uEJh(-9FZ!kR>X2a`ygWeWoHGaO3Vki`bIn6!5_ied^jK%;J09pg1 zp*W2}M0UNMVMYo5psQm}*sTxiC{M`I2W%;pb+8h+GV$^{_AjGci8vy(qcnZgXvu$c zPu;nRTcOfXNsK{ngR$)c*)p!O6$GPg`sO0z>Bj4%6PwVnf(N$YE!m^|q;G&q) zh(Ucz_G)ZM{(df~SA2=m~MP1*}=$#T7P1#bzjmkwYlt%>z`s{4CR3(mVCH6ET z0+B;>q_DXyGSxeFLNgv|HAgWN_{gqlBjdg{G|1bjZlgOisxmaYJtQ{R?AsVaM5Z^J z`27}kuW)R@D?soSy7g!f9quk>f9TGlhqWA2SS|wH$=jx@=0`%5FXa!0cCGHgjys>^ zqT-sPimN!70HkSC6gT*S<9gc?gWQi>-4l{W%svg(|95E$7Q}mgg7YIr(d+Z~vB=0M zoSJA905o`z<8@S1ffg2QyI})u9fOhKI6gfQQ4uQ|5c|^z*4`KFQ14N&e{!Jqe~YFQ z62g=oE4@!_vRK_u0f)g=zX0r_olEykrNHtb`jUbl14qUDZAo7@Wt{bVxh<;j(3eo~ zVsXBCy)W0x9o_osoByhyh8z_FM4mHkrb9nS&FgsC6@s?aClrPrMcHzQvVCvrSWW@E zucD_^?f9r&Zu(@m9vu6vf7l%M$7-m?iqYXDdoT89t@071hU~@3$$&SsdT8T(2^6oe zFGWhE0fl(M^`I=`ZfyZ_@3F(|+hRJPe0&T*w_LRyR86?pzgcY7l;;U-61nNH)vFK? z=+JmMKQ6OYAYa7`=1kh@I1+N*HLSDp@ZT?jI)A9i%e61|GJGM^e;?Linf0bR%t9sr zhy~r^AfF{%s+>*dRRMX7Pj2xR({zO!v8=b7^|F;Az9WUftLhv~ZMiBsG7DcY#z6xU zlNO>3&rdTkPiF`4c^R&0mrLTTIxni5O>=bP?W#$5EV-gl#)HRwT)HammMhQ`;uoKI ze(@wwvEO4wqo&8`f5Wcsw&i^9$If2*4}a@@c;ydLT+$7&NEg2>>F`#f083)=Y8cxL#&3RZZ$L$#!K<`~=04@i;u#?TSJrbe0NWZ9v zT76#J;Q1L(`TVdBlN4LX5y@A>UC0v!tcHgQ_LK=}ogt-ee_GZ^uM88CTZfBn5p-QC z2N+Kbm0;-dnakDWwhgHxI5RLP-+Fi05dmXd`+OhIU{I?Chpd+2a7!k9E%JoYFnoPW zr3#qf_E5h+Q1RDrCZw3kZlL5#Wk3grEUUmimHINnGCAGlfdg2yt1Av+`R$=8@981; z=W=qz&DXYpf7sq1e9Z88ZBY^3ryt%}3jam|ewZoIw3VbZ%wew@!>;(izLHgt0B7*T z`T##_D^V2&+XBmdD2>pe2e^10N<-+cO03e{Z9$jnwjAJF8`zN=h*RXCzrlNm2R2;v z^Pg)d_z(n|*i}vB0;p*tn8st0;^YA4wHN6TFWNGbo5@1khBhE%QqzO=zTF<7PmJ)F z#H)ukxb`5y3^1#7lezY7BHCNgoO|p}jfY?G2jc$&1PA;8m*j~IDVMRA3@Vp5iVQp! z!@*!M_~rtTPHrZ2^8DiQ^~D#@Y&w^TiVQ`6vdvX8K@649sbk1eB)XtMjLEEeK9kUw z^hDWipReZod|g5pN+rt-uQDm-;s1QM-%e*tCHcE4r^%s9s*>VCR_?Nudt8Lqm)R~5 zt;-U{m~fdAWkEQ?X&CWOTub+C#y}lh7<-ZBMU% zFFswc&>&Jnn35C2HPJ8?ipf!~prBe#Zgt-)qq$)^% z_42Wwu}ABl{SAsgVQ;bQXDI3sMYLYW^$W(rD7^cu1_W4US~}BDRG>tq zD>(T3bf$UoFZgBdmnZ*hK%^0zD`L5*UYa?#x5| zXA*$y$?Phdm%H#!aoFuEMVB|e@h2FUnv4t`fAgebeen0?I--)Ic@k!Uo0s6$C>)c>Kse?zS` zj5`gCmd*S|+S928W3R~&rdh8A3!-J+AGKyAA*(<#Q6j2Q) z#rmcmx-)4-v~*4iXCEO+fSX*-mN@|W4&sV3EBv!;x0x?8fB&>Xz+g35KVERoh!C}j zE&3#xIY$Kd5!gCEh}d6*kg8m-f8K4gILr#}Zx=uqn_fXeLn?Hx-J(D{rcU<4E}u|O9;M+^dqTtHjOco3Pb*OU^ZKk4BBBY zJ1|2bns}!VjSMV~+H$@ez*^N9^FMmU0lT3+#?`@oTxOyJ{qaI%V5Ojce;4h_$U4A< zwe1GSZ`k(G7&q0jEX85 zgK$@N4!-NG?D#wu{o|16>Q16ZGh?j55g2xa-EBq)A=1sYqguH|Lm@DN7i;}O*hw2% zMcfaEs9M14GG|N|9DIXBf9Je}If0-GFxF05R1gUFcbO8vtg6e;XWXpmSW`4i)T501 zRJ}HX3zWAtdOa{bm{HGKbe99_)moiJg+HolV+X=0g~ZLYwSnMWk%J#k4WJBMEnrSC zdmrXNV1cNCp^P1la_p%nCxwC4eP8y=M?`~}&anyM>r6t3WHqnKe@xD~!V|yPEM&S? zjLzUICYWi!^CMF-$w`4PEfMyMRi_&+M6xO)J=MvdYYcpQ$vaxX5eoagt;APsVVMLZ3-V)WA%T$`MP~S zoC{f7WmaQdD+OZ!MpBL6-LH7`Bw@WJ4dG-y{p6%+>563=`L7uM>x+NpFg7`sG68$Bvjt4d}id4Vqb1cjsF2Kz} z!7B*q>D2;!DBQK8roA-?{264CGkjg2DKG2$mzM?53L=?03)s!}+L-Ecw~7s+6wGjS z?C3U0dgp)?o?@Z~t|`&rXcnuBH*e`=e+dskB!`0kcBHF`#1lvuQx%if7vK9MQSnqk zhy*fexB}iQ^~ahlYRAr2$4P3ewrR9uZcXGVZCoyIuQph$gyJb8#Ryg69ueo=f)_4{ z&t=f}27!BF)wt?~gf2dg5 zRcu?i0>nxUu78)#DhU3Jle5VoJ6MHBe84EVp>FH3gH;m7TwxVD4R;ozg;h9=rpKBY zxHYbHRGSAz4Iv(&Y^BkS-rDj*Hq-1Zuw&7=f|X?51*b3G1!+XWpDYt{z5fBDddy}}=xyEXQT(RxGo#znre+1kYS?G73m*L^l7nk3J0l0zylFfZE$`oz@>e2OzB42NP zsQ~9&5yjJZyR4`by3Oz)@w{Ay0*5b^2zOF@H%GPpf=7AX+3I>O(`2#UhWWXv9(jGy znq(OiC0(dLdYd0CqvGg=1eT#*S{*88$!=FYYD{Lc*<#MW4ByXM{Hv4)1xqj=<^&_E7t}Y+@H~gm%&=M*-7FJ59igP zwArrxk_i=~lo%mybM(M*aqy39kGAyjDQ@YzF|5Ov1w7@>S`$ZbvjUVuc2xS}DHVXl ze#P!{z?qxyFmxaRmP(K5+lBk^;-PEvaXh?fj00x>a@p-d=$ty)==<2DX{?_c4!c2|uV z@s_ej)m+(RGdZUpwrVmDYFpE-iEMc+IXn6H3lN|VTiub9edrcVQXmL`zz0yt;(n1V z{&n@@_UenTyjjRZWJx9$w;vWh5vg+v?~O<@wYc3Z-Yr#f^ZxeVUwoBj!;K%oMgms}FMGP?Dw0|mC8QyGn*u0h zYPrEz7fK16xkVyW!c6_5oC!lKDU|e!)nKexmrYskh(Hr&87&Z1>HLw99Ao zyACYcRA{S_#YA~#KhbWu|H06aUOln^j%`_$P2LvA18<=&+o5b^B0=e9ws5Lq-GX>` z#et@1uY~a#%ylBuj+xlOj&zzV^Duc*O4bSsM#UodwV&ff=94hRh}3B0!f8K}rEw#( z72f!N#bB12R*Musme>1sN-0tcrOD9bl%QD@JeJcINeCGrGia6t!BvHGP*JIu6>!vArm8MtMja=4| zp`xYd`XOJVPQ!MTNky7PjkCx%*LYs=HRZs6#&*E~lx0hRJ`z)AL}twVqg-$IC=4R; z=~xocqG_q8+^u1$8wsN0_4YLyI*wJ>L)*8>Qc(Y!8|jkeb$#~{E&0O<#SPQh=AX+I z@df_L!|56s_?Fe@vSo#K;8NZo$R;s*wjB&M2&C4>ZP-givPy)LPmuCsQcUB%s_*iD zsvw@&w2(_4?p%zZ-*0kWx?(f-0xWPg|MW5o9Tsd4_3^hG6evrF^bd`uu8z?cpwm}E zr*3RRIruIOE^2?M$t+2taIj4}GMTAE8BJze5E0TSdT;I+vj!d2H9J5fyi~Jo60EQ) z?+#40-nZp8uV^U<8?nE-E^c(P{JgJ!>I3RD4|@+oza%!HL0TU~A=L%p?bya0gm8ol zlbdbWBWv)R+Kml%cv2b`GieQ8U)L47^`30pL%I8ajMf-C@Jz%_;49lb=|{nGcD}`I zK}dOdC_oC5HN}b=h^%Z+bd1c9j>Ay&Es^1Ub>vB? zZ|u3x0oI^i1ogA0#Hd8mcvE?Gj}`$=q--Cy1)H6esW7QN0Rq7Vvz>@Uo7r~K5SFA3 zV2k~$5z5W?uT6d2;E_ZEYkrP@-Ak&3GU;413hpHkyABQpo9ssb&v>uqd;SN&k7ZSX z#`1e6y5Z;T1Dvj=-U&j^5YmHtfIsQ%;s8MUr0eI1tEI@SRfMa-=p^FC;6OWMw4z7| z|K(Ym{(mvT>+@D$SR-6!7UkhDF!a5W1N-C`H;MvQ6o0U_`ufzZ6 z@Lw*!5AEB~{wB2FbpPe@htU3h@c%i+@#Vil8}br3)yr4>qXSF5v?Lg<_dYIvJHa2o z(jHF{XSC{f;s1}}|JxBa6AM`<(p2^9pupcEt7-6aeXiBNq97ad9oCDsL6-712RB^q z?x**ZfJg@r5FGB4nV9x}fTR)11G@+ou&ez14MYcw1ge3A1`gOldMhu%cNED1wmIp?k^H;oFOV?%O|Wuw zD8DXZ>)Xk`*SA-nt|WJ4fPdA(LQ-t(V!fpo9AsPIxM3v;5$sSbK3x6u2>C~AD~JI! zKnt6IrvX?`{JwdAL8su$U5qPWmxp4viQzBj24)%fY>s7HJh44mfAMbRom+kbxYT5t zu&@D$b(galf?0TTi5Vi9CSVCfYTU8sz-N?^lT6CP8JB^O1pq_!GC4fqQLGUW#cH5cj z>*WhLq~RFj52`N>DRBDc4V(u(Tjx}4>+Bjc&J##~fu5A>dU_NjGLg9sg}zJTBsrXv zi6rgj{B4pBWaOr+Kg;)z)H<`i92$I;0Z*0_0@zZrWdqA|Q=0Y=y+_7cn_0JtD?AdRAVhG$)N16*L0LIdI-xx!Z^oACOfLTxD*iZCG`1=|gG70|V< zsRpIk{&5WL@%|Wlm2iUkIlpIc31<3~-#oh`VXs6aRYD}`z<38HI=|krj;NZbjY>{f zRVFYRc9{>RCPDau^ksB^;@04R?_yDbLZitbQlzu(^bAkWITB|+ zbk0j%?IS;o!QUf}42Qsx<$@?W(4*DK;#ms$!24Gu;7LWHXZ45SGHvyLj#~cW3t&6#?)AU zhtT#Sbtm8vt6s?ZOPZ}m?gu&y-h%Az3u4G{+0t|@!gyY-xpt35-AiuOioQmcWfv6v zaM|9=9H@iK!iF<8=Y;d~qiDR6U>akDi9KKPzDs)M$9U{gk!5o1wYD`KUV|X#dQfNu zERrVn;zBDzu>kTBw6P&$n-LC#L6Q1@3vPWO1~>Kw$o;IP}OMmr3A=JXumydYrKgvKJbK7kx3ClW6Q^(6;#GHimq zrZSxoflW5N>oyQ}8r+0?JXM1!Ih zWJ%FBEf`*wjxGE#O9D!K4Fauyc#|D|(2YhIA6&vSZztQi|AFtlB~ zBCaBvp>?I#b z#TA?53CtY&A^aWd2U4Q7$k}WB*}4#2g)w&{y-j@!vyzZ*C{WJISqOW4ZSi6y9lg<-=(2F1B;Gk9?!e3~*fj2#2KuIC{tUF%G;Ixyk!<_3}^FR2T zr*!O>Ge~_AFQ$y~(;(6W!~lngPy0V(AA}5Wm-G6v$LIY$_IY9B2ZD3oo!j^2)R$m* zJ?9)8L&)XbnPLD{dGBzZ4yO#TccxD)Awh3-&WXDLR{4aJWk=j5{~H@q*T=Df1Z$0lcU+CqgijLP)f~TnOU}6J;zyZVAv9 zf&gBbynCN4HZTT?6;8^-{BLO8VN^ zae6`A-%MK?3gQ1~OQ&3)3t@v{px=Nx^QI}tBlTJ@z&sQJq!Y*6JDTY!o7C0AkKXLBqG6W|xmp$C9cXqkvVRD#Fg zf_q^Ah4mQ+Edi(Q!Yq#rMJSO8nay8hT_D4JApuknc@s5UWxubva6-MKdpb^}6QRfW zxhf)P`dGA3on|NV@6bPI{gDydei1s>gxUP5WV&C$w$iUh6STxEX z)0OLpVJ-OM3%)&wfHaTrGk!o7t)s(a{l`N5gb49JO+Bf*myV?jl0H@C8G`iSk=1+V zy(Co3IW55xt9R@NCy7#&##nx}6&iTG*l3c}?bS7b~aF zw_c_U69ELDgs&-wb*Busb*Bw1%75ITIxZT?{tlHGNd@8=^$z_Fp40J|U;E=f;VW4P z%|r*eG)JEuiV7O(Mt%`%gAJu()2Klf*;4zVuMjrWIuP7&;TAKj1w9+3T~?1b(Jtv-;3S=LsT>`yKHxiU-m$y$x?)xIr4hv%G>@Je zN9vV%-={B`JXwpkF>($V0RfV?N^=h62n1yz3geeCat;%hMSKnym+`0`DVK$O4oM|@ zyHRnJ#hKbXecbpsN}bzyZ=*QVo2SF(m#ucY-=6;d)%R(3m-BoMZ+~HBain9tq)|m$ z!Dm-v#1|{aGExvaX>^CEK^${}G97#}3|-ZBupJdkCKuZL@NJZ!FVQD$U! z5k0L}M_J-&z3P?qs(-NF@lfT>0Q8wuZO^nSY_-p@6P3Ks(y&n={mqb99rwflfJgG8 z=vpQK1Sc~F(Zo6pB&e7M@Tv;#&NCEes}5 zR6kVBQ1TWQ129rxrct8(dYCj|RTg=d$&P#3@f_J|t>2<}K!3w+OXhslwM|pvPRWD@ zsTfZLOeHGHv{@~YzylDJ>W`6SFhYo+ID*{4I?JXnt0N`SUWZR33GnF7Y1{8R$>>jHC<}lj> z>~mWdCSW}6!+%a~Vf!p99Da}sFEX%$)~%L?x@tRQ2!g0{##7|yyb$3^VFBK)UVsqK zt3c_YHIO>uC&05b%VyAG<6a^ze??To?vTkcq$2q2+flS!Czj0)^d5L?z>qktUS_wS zvzxcMZqH}o0fIVrEg8$Q?<-V^S}>5Q1e8b@a3(C8#eW+ISiEuWL{)pr(P%P$ts)!S zrBJ~(#;}fuUwt1P(Gmk zaz`nE&K;AP=_od4wFn-nF07N}hKTE!DkkcblM5pNMuAI&U6>oL;3`a%D7{*t+4+EU z#dG5|lYdEfODZPAWlZq&w6O+usv@=-9n7g=|3^K-a+(;Xp_e#SO*L5#+-X`3*r&{9 zYe~x)oEQR#5|m9=3;SdTkRX9{iwl7XF0$6%Fo7D>|6J!oIs1}FEG-#l9Lt`ClJwG~ zV9D*vR>3W-0)e>FbO2mS?^%3HU(OMV1pwJ`>wm3rQS~JB1endarDd-LaLAACk@Wi27b}q{335<@Xxc2;@V79a2SDa1oPLlM@)958$IoSZrix8r=Ueg3_KTULVgj;?*N`ePfJfUDRb6UAAr8 zwr$(C>nXdd%eHOX)n(hZt*)tW-bp6&W0Jd)`{yM0COhZsz0Y22nXEuoPA0^W+u*$M z@Q-+p>y%)+9OmIWWz@`c(YN^(EnN1vYEx13Z=DE}{w8UN?m~HwTp-+c>F&=LKh9ad zrIWE4sSqY;HUYpvVh^6UKq1evYON-EoqjTc_q+Qwy$*2HN;yJh3GE}(TX6!yuVeHy z$UxPmW@dPE9yW-4QxQbiV(!m}QiAaVDQ+u&zd5ac3QkpZ4d*cpc}@%pnkqRz{J8_R zvEU7rMT^()69uQK=tvTs5~EkQX+1X2p~>*v7XshYb5;PS4(Z3!z#lsvU^$T>6GM%_ zzQRjiM@nvjVdQ!oK*#a1L^TB9l8KUH_F)FoFl!}g4l4Q&jqy9t?&*RDy@cQqHrLqj zv9b{mZ5=bo){@RFTA@<>i+zL=G7@wZWDuhSrMv88VRRjPj|h4I4P>nSXrmqfd{-Wf zZx!n5{gb8T`7I!>>2HoVbkBDsZO*6iC%b=4Qr1DPXjy`gYq0Ex$j}Xt3)Hxnm>8xI z&k(n&QU{2YlkOFR(AJiw)Tm524Y}=9k2AsH3hUx@Qo2PW-!l-bpC|_fte z-{oeDg*~a_^~1-xe&A>XS`F3{h1z=FVL_nzsl{d|Ma>+rAggMKQAV27a(dPQ*QR)5 z@Eb*S_e*xJAcMY`83zC~$&iB3(@2f&*p;YEmoqBK6M2uAWjn1!n?*hPx`bJdrk5#3H98|$)X|^EL8YxmoDrav-2NEfTgjXvq6V0)f_2@Xr zlt?qKniZlK?^+?lFM`+l}=gt1WTErC4%SB%nrU^6jZ5W`ci!?6OfDxteiFdIy-9Lpl z;cQZS&GGmf6rKY}-S#|E_6GEUj?IW@KC6~NS6i{6&DDF8COkJH$N>QN?V8+hSM71&|U%ytt&ZO z4Ib_U?&-+LGQxd}K_<&$D{e5KI8*O%N?L&-9ql^WxJR5gUHb zD;}_i+UqhPutqMUVkxl(jJqih=^e5Nm!^Rw5HtRBi8pZz?prjk9Ij^IT8m|wW;#Hf z&p?7jfChb3HbxBZ_~m-|RF8Is!=D{O2!6c&@z_%s*Md(aj_d`wdJ`R0??j5gC&zPm zxgaAn6Hft)4PxYk4H&DG4b@}OGYzQF~KNtTfi!LF=z9bF~pvbS-%6AQ(Fxxw3{yAaZJ=8>`jBeI$W@B zwF~Ae*ZFCxGXuA+T@lsBo*FECYcyu!TsYeKuT^69d~xi^2n|IoqTiCwGVdvm13Y|6 zjn=ozuNKViL#5&B5u8^uLt7Z~u8}H+{W8q?kuL$W-y}vryy&*$9*_5)Fxk;0;mE^? zKjX@RA;k!}Ay_d7Iv}qPX#D>G65j%I0)$Je2dD0^u3e?yxr3-M*67ECO}PWpyM2Ea zh##MF^iS2TCg)y6(J%k`y#LRN&)g4=I33}C5JCXcu3l-~X5jFjVVi}o*Qdm+(5xH~ zAnpLhXJ=AGcvQ!jMr)GN{B_dS1D33gM?m5#bibxSed*x5zM~tBe8xCEC)%pXDvX1@ zr}8bw1~T5B!BRW}dMJ;;m>Ec_nP9Za(J}_roupqI2<6EOy26uo4D)%p&Mhor^k1rv zgb>9&MgGX7`^s`CCQ>q427P)hU)GS{SRub^%CXuVT9{?qIH?lAX`F>`)8@Q({{V!l z7hnLB$yV9IO;azi(q3c@j7)I*=n$)xrPluHp&rE}!(irIdcLPo(#Luy2WI^!k!H7~ z6x{G2W*Iu`T|o&wH%dP@6TLUImNkz@~^hz>B=9^GCl%EIuO6F#tf)>qCa$2 z)|kXoZCe!{n-H6G3^4W=qjH^dhoq0*>(gy?l%CQOfLH#PsDWJe>=z(EMna*e!4u4f zxVCgRVhr~y$lgx~`ndVM zv0UFCi&!iCvE_ZVgW!nDJZpWAg1?6Zi_yJx|KA@K4w?x61B4$S{s8F*$Ui{&0ctB~ zA{zAEe=66lY^Lb#Ah-kL_r`5kRQBfwtnjVJwHlF?%=+<3OukxtC>=f z5|i{OIoPC-7O)fSeIl0ZI+&3=MxBsqEEG!V8X_l&Y{_MR7#(LDFv>zbuz%bIM;r(j zNa47rk(q5C#tip@v&CssbIfzjJm$xadumq=sH@+aa*cJnp2I9!qrL^`Gdz|69fFeTh?T4kw98wxjYyV$|K3qHdXXw< zM6T)=ftF`c59r{z1e-jdXMoojhEP2x@qN3rXp1(gDS_}rs+oQ?Tf6jYNW0YJ_d@;1 zndU*$)Oc#;t1xpA$hJ2Nk@m#wn|4CV7vj@PIKT%Y3&@G(*o5>mnC(os%arE^s8un{ znb-=%eH4qR5eQ@n?r5lg3)D8bWp!np{FI48ZJF7S9ngdHc3KP=jD#B_AYWW=6G?VS zlTU6n6YE5R(XcAQ+$Pi#@Gbd8gZlNK-a-HA=VQh(W^5S;Jv=C z$MPD`E@@nZH>7jsN)Nmc&$0UZ8_-gFK}Jh=z@QgnJDALsC1elUjl>(I{yg=q49pG; z4pW_>;Aa-L;DFvZU@{?Xcp()AWgj^g_ewN(o0A~coVDq&NR0|N`zT|T(}kF-oxh}9 z1&im_kgC#w_?)5VQO^y7)Y0-5$$*Up5;2-BYaf1E8rFC`un8_t z=+O|V5aaN1MDxOGiI6kzOU&&kZ##*Q#8+4aaJtAU-m^DippvIU&?OvCKhUI?hbcPU zUIzWvy5N=O?1&XN6H*_edaQ}6w#`5RNs+Mk7xnPLO6<#Om^nrGP>)Jr%bY|U@Z4y& zx5ExGI6%Ap@bp1pwynQPBEbLC(XajfW<~%mU4Q~a>>Rb{qd<;%-CE08FcSxgbkF^G zg+IH(ZXz~>dZWAqH~zP3eXQc8YgBixxK9cBw7$3N4QmCD)#82k{OQ`!`+H&|{U&{D z%<{M^=5`965jBZVr;b3FviWZUfaeQCGy{sy9e}{6(KCebXJiK{@1=%{Q4x@=;_8Q!wbI^$Xv(-M_V?>%FTd4)JMkg24WRF=fo{ zJM2>Pg$(Pc=Ky-dFPpSeKH-Dw%=~EU9-JPGO zExxKcU^q_6hxZyemI-GVm#3XE4T*V}m1;f(z#9o6WBI3%#__kspzubnurU?EKmhOn zH9SCV8qb|;yllI{z>_4tB45+IkqJh!zaPE)9bIluAUB(JxR>hKvwA}>{XDm=fl!7= zHIcNuYg2eHpp_cd`QOQ-X@?#|x^3rLbe6qa@0N}&745bVX@|-3rI%4 z2CDJE=~0@lzYJS|X6PS|m*rA`WBKE(ro#=wF3+?Ys!KK<1#ANkxtIkjZxb+oG8d^Z zVuNO3-rQrq#))}k*YVEkaRkieGX)zbP6+AO?#Hd33;7NRi_Lb_+)*9vsk*Q>NLbxl z6`sy}CySkG3DU>fWb?dWt+srj1sjqmNuEl0t4H7~%ftLtY22w#c&h*7FN7vh9tj7& zClC$Nvhbw{@VXqgCIC}OvS4=OMkun-$a)7!3K*?xFsa6Q_SuJN0c}~TJ;qATp3=#p zws?RUPb~ttlY75*cjjxgE4AwbM!y19!XGATPNe-%9<`wmbc5W_%YCRTuz6%rDSbT+ z0*Za9HT$9uVrc|g&E9P8G6W2dUP%xfUS0QP$#$0lD_37-N(eYu+~F+DtpcWghAkU& z#wN<`p;D%L&9f7hPO5#k}NhRNL5?f%Mny0i6U8YD`{U`?x2}<^K2#uUR^UVguwv0m^ufzdyKj@x0y3BYu4{}V6AB=gM@K@;A?}| zzn8@I?rihNeMZq0blH99IZmiTk3~-c=L>3(1%c;%sN}*nLXRtxAzEj5T;G}nXV>z& z9taG$0&ZkX3})#8;teZ730z{@tuRI;#Y|GX3wQyxm(RXHUHk3Xln8Snw#;07VA~Av z+t47vbMA|I$K1?7+N=@W4V;7(0*|hI@O`R>6b={W{mcf3GMp5iEs-BP;_{)XiVG87 z+lr$oQH18_n(Z6sn%#|LMu=}>?PWrR2(Bos1dx?_6=o5+IHM z`MWef!JeHwGO$i47R`1C4UvAeLBx$#hg=Io_}w@njE{$^KhYPz$lrarOZCTvGcHud zLalBMmE2_075tJa9q2xS3`Y@z=43_2_IvR%1b$U2fXRUSjaLcPaJ22ub+5fVMc^uH z=phh|kU@-Gvecy;wh0&L!_|h5be|$93FRTlrZiSxp?s2yKfQG)&g}pS#-%gOdRAs* zdEfH&&;W3a!svwxy$&(m-H&g`a8UZ)cKuZq*XNWvj4uwYz7tBY?(r<6XTLhsEl!hd z7q?+Nt&TY2%m8MtTz{$UxJaH^tt6DwpHNN6;^+d+v(J@yJh{2-S@}@eP|>haX3hdV zM{knrJ9!bik;-Lp}^>X$L6&a{pPnj@b#~6DSzh!)=XyhyPg@ZQ~VPvqxULzQ; zvsjDmh_IK<*tv(sO?jyN{M-kH$vXhzTLclaui%^nvVeHi&W!->c4R5>qWr zP~}C1N1W+^?KKQ zW6V0KUob=rK{~c3973z7kEEt-{E1ZRx>wf@kyDtB{cpiY7^n5*)DZfMxIma}4Y8YDrjD(fia)ud@Kdv1?=)y=b7(`uG zba>O!l!=myAB7#Tm!V$*e!&cu@o0+tJutZp;o@u5#JvkZ2-JkUlI_D}7+QtGYb6?H ziHwmfEB1{a6l#@xOg9PBuHhB)D7%|@$0Ey}JnPrC0+@E&NEO z?-#9pE(>!VukKz8?-P;Fo^$eVXe)t|F~%#){+MUGU?p;OBon+53ufzg6pE7(2U+h& z<* z+QC*RCu_?^V-T4F#dl^%9Sk;UmhF*EeNJ+fd!rvR$j$bq3;3ViEm;hkYfqPD9&#rI z8P#lD#@|Tnr*1mD?4lg@Z-k)v-nKK!Q23L+b~6*GlUU;<=~i*laDn`+vqy&AN&Utf z6biqUSEWZsXn#YSj5!ZoZ~C~;%epY@fS{vY49nbV2KzS*U-Vx})ar%) z8<22qN;%J%+}Wwn$%-a@p4}`SQog>0*69(5F0H#Gs)XD1wX)H9f#A`5Ss)q0y+%c3 zCo6a(g4bD$L#A4A$5Or3)8pskS7?je)mD77;{bSndb&7q(AUjm^z^4w!`$rte1ZdQ zALHmcO5XALxPLQfqnglwVFGP z%u38W7Yeqd9+bz9!|U^W$!ld5+VCFIqTLGPhkgqQ2KUK){@>pthob*qlS>_y4Ng@5 z6OC88=}GOSse~^~O1rZu;4!AaE^5iN&5ARKWPFLr> z$nPE{LE*}quc)+Sb`CP}_on#;rC6p@tY%+m`tS?Pz3&Hrg)ot#;5#M|hH1rBEBI&S ztxk4~z(U7N34qA=;P`vglXH3+`MylQ!Qcckc507bu>GYxdp6o$z-NQUw$#~w8{-l`$dn>cLX6}M2z#l0 z@)SFzRbH3kX}Y)A!9cy*#S6Y{^ z7u0NKRbX#)@_iQr#?5j%BfFZbspkksjH6^H>wh)h^cO zk>e~Jk%{%|5>8bvM;?-~OZN(9%8^_sjE1m>h4G|-}{6uTxfA&#H& z1JVWDVV(K^Qc74=to>*<1(V78LEr^2^aITIxOY_v;US>`UNDa=e z3?ORv4{Hp89aQm~kPT|*TQ9gCI@moVx- zx8%feO#Lj7lskvL6}&#+=2b?=R0s@gsZ|l^S^#hN3Lf6YWj*hWlDN;meZlC6?2R28 zuz0@RR$-50&6%H@Nl)xM@K&lV=CCoPS*3^nDj|J$p9{QSPq4Vn-rl|Se{(8efeDS$ z?DI+GK26covT*~#<^BEf!8D(M+aJ7=sdj7aI8 z`~q-L}s~>?BF-T9S3csMO#sMH1zH`!|NlHC+0*dya;bh2Yw~$#p&O{um z$Ej=EbIxvyI#UpLRZpk3-$(bHkO|opxAcLHu{dV1-QTc_qLnTIygGosU~ka5HFw} z)9n}fy@BIAM%~VLiDS++NL3Rv8rLIEz-_p`e#nNPa~MYp zDT!oH&StA_VqM;560Eft2ZgptDY6LlF66PI&`9%MTOi4uK>$15a9o@mnwHpDb!sVE zR$Qnu_c-8dnIH05gH|=sXJ~gI8y21ORe0=R*WahAGkD4BX_n!WD>UlDCKWRpof7$ax^;ok3w=}G&&wYptD1GdyeU_2Y5*yL7hc{IUhaIdFZU{ zpv+IpY}-x`^JcWI#N(Tw$6Zv>pGfK{5m_vOYbKMxQz?*kI1PZFp01Ik@(|^0^$xkC zTo%%Bx2y9DtEYHXgBBqaTlXfTw8GC-D}p$&@DFd<-+#g5(KJm8nPShn)ES*Ei8V4b zOlv6LxY_zWE3AVQ>GR7Iy2KeHg~Lr1c>1T^IiOEZDE+G%o!`Lhe3;(c z0{9rfjQ``ooXR4IFGDIj#Tm`E;OqX)IiZ4g!wx5!x!YlwrT&w(8ze+e$w9&$+8V za-o_}wX$T32N;^pBxOdEHcevEBA!t~`&RQ7;M&4ju(WZV(j;n*jZUUNo_+%KOHe2- z34?2ONm8eUHm(iE|#*6T@w30Qi(a zUUpWZ-0t(nHaV`&hGwHgB%J-N3j?Cga%ZRqLC%vPjgC&U$}y2D?Z|Qd^b9=-+E2tX&&k&8jvO-f0=I{EXjww*mCUwAhb044GN7Cgmd#8mirLk*RjnN>1di z7RpU58gpFRK(#p&0dh!nV+p78fXI4!gzN2o!cF-ocQYMYSijBXcKSSC%qDh4H<25I zXm^pq=4x9V+OsxyP~a9PAV_7C7JU`O36=8jD3FtIWM?MuI!<7`)ne09mQPrQ&xr|{ z+9K=#P>=0kgAyzccHZ0%=~n1OpaVE0a$VluEn^^ex zyrD)0x-pTq4j!RkQbC0^Z*UiK-;P5Wf)93k7I@BYy>)qA&*ri^=W_RHZzRLyQ8j}%J07Bt;#6=X)alaf@nHUK?#FEg1|m#`kLEA z!0Tt?!AV*C1-685HBli~0!j%YG-ojxDq@Q^c-J8An;b5jE$kqqJiQI2mOEWHW(2hY z9EAUvN-w^BTd9X6VA@U#-DapoXB&ULBU(iTxYx^CyE9vsjmTrt3!`vnt-&$fNl)%g zBtnN5iNquAIg(titA!Sc*awR_W0+Qo&zv_!p!?gn1>tEg^w%k>u3=UMJbey<{Ynfn z#-qx~ZoB$&$QwM@t8ge0tbLS24WSM}9^x$y)s0DiP=sgD%S&C3L*+mqEkl_rO%whjx%`(7W6M8$~fb~fg+{Csa6qv$$Ks6 zbF7=BG8=k4z!01EF#i2RHie~BzR1gM++t}=g;LB%6OKiT@2@FW{z1$lY`RiaC*`Rh zl-0sfGoT5$ti%zFd>VbPx)#*(FRCng@ zkVU0ODudYwNBoz+d(yHZPiL&Ej!es5U%Rr%*UKSAuYcM9oQpW5orf4z@Q$E>Vn%cm zKTi*X0C=-8mim`sXXaGH_oH%8u*@2{5yUGxv~^5CA*g()k%*AOW?rRvm~{E)5g)a~ z8jwIVi5G~#){QS*2yAuRCkJZ$`s>&O&7-Vl`nV1$^gekz3zj&@^DI#XiD*xqfRZCM zOj7djxl|~0X@S;;W8=vulJBs)+|Yx%V<}as0f}?0T*_&($a`GHABJ zu*C)LFvvR_ZfNv^{$S)cx_0NO`iuBz*WQ0?OU~O(p4k~mlKOVnk&o6@lW(RRb4Ixm z2PviZx@vE3dH5{rx;>k>u0EW@sKGQd-CDFY5P#84J^e;4fa^x2)B5+K$ST%b!jjv~ z2LSqogOG=~sn7x=#%_u6%Nipx&?nz(1w5$KL1CWyquVBkOpzG%;>^5omfD;P(BItu zFVLoN_P7UZa z!aYuScQm5j!vW1eaLjunwQ{Uew*JF425@ye;v-q`7aXvVJ>B@lA9#*F;wBSGk#;y$Ozi+@$DjnJ7YsztAc}(b?~FM9rRzbmgYOL7hw+OCpGTC+S?qiscpZ{% zqqbrEU%VY6-wRE{&fHTcQn1ZK@x!}0v%nkP=7t)ah_7!d?Qqdti4>+IL}hyh#3`IY zOEPFy2CoCQD?{1R=DQWbcx(%000Af(X4xO>nW(~R+Z39WoTMP}>o;~jt zS0;1e`tpz{hfUI{?wM(=5tS7(v$Zw~#GKTZ+JBHZdSU7RWqL=_KbhWtH0JD78HWu{ z#DJSOH16@fs3Q^N;q9zNyLpJ!0=QXQ0!TGuTYA&6x+1wmgI*CQORb6vy5~c$IFf`1 zC(e?7?!5lY=!wo~j9L~nMrl+v+BP4kB;u?{!SOW3;owf@S7iNd-R91q7w?U~a=YMv{o;|Qsks#0-clF_sLBo3nFz5?$dG5_~Pu_82uWL90bYO5_udV8L zIT~=WIiu~@55XraT6MRoWg7>z1#Q>@zxKii>&gkGQC6W4h*OZslmVNAL7fcuL-2Ij zjt-AOWJDPT!LY$u_sQgo_^6-EjL@1BnO{Nuhg;20E2};RB36n(Wilh4>Q0Ssr@gdAb)1~{c zK9{^cQAlJgy$wvwNE-DDNY<_74pjifVFh_q@%5@>=h}TMQzU&9i;*stNgvt-$1~pM z7Fa%Rv(}YjD(;0vc?TXyi9^JsDATwm=%8$U0bt3jKcbq&zJQ8{YN{&YfoiyUjqJ#b z8^DF};drdAd-ht!$)?|u_Pcw)NYq7_gHB@Mk43rbm%hnt$>Ne!LI`SzS-HSv=6l4F z;|0fogk0HRwa(^cvgFU2dxq73QDXIwTzf##sd_`%{g!d<(B<(;k;RI3w92Yn7%s7|Ft^(cg6$ufK(N<) zPE&=5P9)Ari5b{=LY1HX@|_*Nb&pUD^VvH0HL%SZrNfv+;M?|NOfjzwXZ~qchQRDf z;nAO|H3GJt1jTrwJE%WUCbR~`^J3kPX^OiTy^h=uOyn`Uxw`y9`G@5a5ZTrJbH@jx z!3ZR$7ZRigRJf>6;#QAk^AS&|#!%y`VhQhN)+|C{({1W#L zk)+0jzZlB|;hKyOK0tGR8RF=Nt#@;?OP@@@gTS|>URy*GG~msw!*f$BvTOHjTxI$O zV2HP;PSWR2*9iXHKn3}`DwTWGL9cKN{gcklrPrb*fffMjey?AUJZoaWPLf!#s|&bv zPX;(eX-LRO3`NU|Ylj#E-3O+kg=Hpyjc9Iv4l0E1Y{ z>6$}sKZjaDhVsp=!p~vGxG2>S zA+l!lUW6x1sLb-3(_>RJ8n?_j(c1%*(-CHfii(5C0A-4il}rzA8Jxu={=M(3YeuO=Mi**m1(gBuCZ@8NzlRxjetv!5RV5PT_PuYSVl5Q?SMT9 zp3d|S#2kga&LXbAx?ciTPfqjQ1TUCsbkjDJ?Yf@*&VF;zMk)0uCsPHv>aJ2##E#ih!ui1p0nZLS2*Z6>rOWZtavJ4^iIy}?mCZno}OP$ z2LgJWsitF=kUd>f_Qd-6icNam@}F#W-j3h0)fB1sqZfvh4HrE)am!-&r%{z^kUdFk z0u$u~9sJ=rmd+XYj-aD|nc17VxH_8|+5OL3>Z&q6?*Flk$5X*i0v1edP{FSN<@C%- z{n5K;0_IAMSH-sn7~bWFFs`@30f|=Gtqq4)VCm{<(6kDS$bgN7oyQ}#lnHth1C!2a zg1%nuWtsySDMoyfG4~lrlZ}3@!u!?VCaa7(r)uJyZ%to$wf(MxhphSlWGkzS!k5{6Kbt?Zj!ZS znV0(Rn&m_ky0>q+*HD3C3|ydldrQLI+tLfGT2B%oCmE3}5F)MWq9oesf-kmR3{+0F zFlsf)sVuAIW<+XeZ*K{p-kjVW#Y;K^CTq<6W{Y0{Io@e;_aD?(8_DIhlPZXlt>sFP9ioEg%xZAe`JS5x;L{9eqwq=2jRlt)7wRR?wq(rbcg^4Ovcl z|Gi%Vu!^oeUm66G5fe;=Pt%ZYoEdk#>{g||rjMKe0+o*WPCNETi%NnAwN4w6z>I z#?kywYSTtBL&F;xLY;uUIL;%78XcGMZN=NQ(mYGhL^AOgY<4%XjMYyGmPt=(gCFcnlP)GITXn=1j#3 z4ow0pipS)XbxD!3vw{=4>B9pYq`P|B%NouvG^jISv0Zb{B;9JxVwDsVTISPaIWi3k+A!6VmXR=aZ+KZmzX)j= z(wpiUOsol%^H*NfnLHH07@Ad7Fo)E+3NkRjL>@aH3-Oe+1e@>#RTZwQKR^?P%uyS; z9Xl9`Jm3$!ZXeKs5QGy5=sYFuxo*(iBsc>zjC0rQP;UV693AaL(HuI%(=Tf?BGg(U z8X90(-R9Y}T}LybFhnTKUl{%Ru*48TX6n$`0bhpds6(WPSdsLAzky=NfoKq*hNXmE zW#HR2mV{O*=5tioKqE*W`E8gXwk9MKGd*(BDK8@j71Q?%ZztS(1n{H17H{Tk-L$2- zU{xymhaYc3J)tmTLt-hpNnaU$3G|K#&FbJ zJLV4WgB^#77jv$FF9fQA_n*6>wo4kNs(J(evD;RVs;2!}N_1Wd7-jBTAX1aV*_BZz zZqD@&og2TKB1m#-syu0Xq*A-*v4Gz}*>#_T0{lVhRW(@Z#~|0Hyad1L@q>pep_eoN z;*okmy0>DlK_4yFHJ#_C{1|SGZgkQL z(DW{YslkrZxjUs6j<%&A>3EuO)%@s=-gi+9wQ8~DK2K;lXPhl>+&gR9svBtSv-6H~3DEYB-8M1=JUEZoEQ#HH=<1KQ9Un;3rlfxBPHtb- zI9|;Au0D}CmJq9fW-H+CC55!(Gq1vLBzy1BQ%OohRT1bkn08YWE?qqRA-BgrCJ)bD z9!(y;^E?{@JEZB!j+f&M8)UlE)z#HB^hoshO%peO7Izzc8FPw$YVbeaE}?{`TMI4E zpg1nqIX^r$eG8KkCV;d_Mo%+SrMi^~bwm-jzO+4w8D$U(plhCYh{E>rK#eWD?(lq3 zLS`_ZuXTiMZ=Q7t~=cQvyh@$wMT>jVZ^=n_4 zFt!#O4-E~0vVoKoqr3_=isuGj;-u(ZSN=kaO+4Pro8m@27SS07b*QYE4MenJvn>^% zAbR)pAg*tS_=~>L**9y~koWdZl781}_oE<6p#0O)X+8&M(cp)+3*OC)|D&Z&FQ6Wj zNb1L4$1+o0a0$~koEt~WaVl9<9%sd0 zif(ciJ!ox8tZrcLvL?MEeM0!}^l}YAboj3oq#>y{e8e9A(N zD~9Hoqyf(Wv0vl1>#QB7xTofnL4rp3hnXE??c%OsXJ?4mDo!dPajRrSXCHSFJ#~hB z9l9H1FU~%7ieUISe2)+`W8O7~l#&vLY zaaFM4;KffV+elcUA5_8mhpj?$ePQAo&rg48GUT31{7&(V zhcAP9^zs{`xIOQ}HaXLVYLw&SRO_nt-vPa1k+r4c|J z_ruMOZqAA8CUqv$EnM1oMY!FCJv;Rg=v{Ggd?+``VGrLe?im0LO0gqe?3DMLRW#s zEwfkD{h;c#BV5d9emJvuaro<-!=K)dlD~3#ROD}MNWjhO(=@ib@4x4R>Ft6rJN3c| zi}Mt(F#BQvgOI?tMPf}GYEgGlFyho2oVR#4swCrS2}}>TrM!O8AWhn_&XUiVLh1Q? zujHDqrZW3SU4!3Bl(KGZ%vg61RakQ=7-(mfH}95f zrOjFz48^E|yaYrHM&7dMxw;)31aUGHNnCS3=YNy{Rix4JV`DrsdR~8~H3q9MRuu9e z%qxtQk#$~`vRWGI7qzVJ;6xpO070y3pv7;7E*)jf|}vhNrRr#!QsILOJ->FE3_ z>cf*|rx<1OcPxclO^sBJxr&}w08sU+FqeLrJJZfjlT`+VXLbx{C~lj_dxTw&`Qu_b z25vjRB6(e#af3*~-B{BF{6+L7N;dmDz(*Z)CF|kQfakN+#J0SBm9;Wc+~O4-JKINw z4zID$C4`v`Pjw;7etn*odo~5WIenJZ^kVV!7N1Sz1T6HGA5gAUiSO|K@wp*x!7>M8 z8TY89S=d%`5O*frK}moM&i!PLZJE)rNqP&h%LbE!)_)sMp1;5ezNOT0YkfwQ;2FX7C!<>S@XS$EdGA{cU6LFM6z z*s#rXxUg<3+_ls5&642}{}Qk?+o^b`0pkViKQH~uX}STxcVx~g5MMWA^ILv-OU2RA zx9rXuv|Ln{_#>X{OH0MN>6uSKcYGM0w-39jx>TJvJb(-VJAUIYtoBZ{ZWna_S*#H zR;B54xM}%+*;YfXsmsgFCAB!bgkC*AWx}B=(+?XrYC*Kn*Dfb1aN&-wh8Q2=5};L* zaV_`Z{OmB!lyaI_?;%FWi6o+LaSxRTt3#KH zShx|4^ul6?l1d~~ZQ@bbXX|dmv1|Yi4tJ!?`pSL39VLXn$wv0G3a6%|j}|=d4MMcp z5t!Ly=@TLw$~;{l?C_<1G~sillb+PnLusTRja-Zml~4_658+53y-pN$2M#SOrRLx7 zc}isV?`c^XJkxBB&YZa@JoYUv6g8zEI)HJe}z{k(9)8mg=MdzuVd0a<=QrdW0=MSA4SF zW`*Rz9&>coiWg)5_YMcY>c~;Uj*Aihh)aI#oi3*fxrYmAs!@&B2-83A0Gg?-O=kee z+t6OY9#fm{&i9%pu9ef0oB$aIufrPNW771&4eu@WOB-iMw`%nT3>sEkV(* zRBUBer&sftIT4NgLZZgv3&b;=extI@|GtUB!vWaaUmBZ_)CZPqseyV*7~)B^=gof*vHl z1U@LVBVp=K>pP!rMW#5t!RWrH{&pLyTBno-SHZkTc~Oav*U3HQm~r-p9*?muG~cah zR1!uH^F;nWW5w@ql!T!Vn9v$FuakIF!2-lESP~RNEt~|;1(PPZB!j7Nu#gE;hHyB~ z(#71dqPsqNZH}?5@R*X3i!iktcCQoRTRnkF?etA7uTQnR#F=hYrFPUADK%s%ALXTf zpF6U2_b{zqI$3=^3|bMshj$HA+#RS}>QC?RXcnv0^J~cZhBj_a7m>Y8)1`;&%XGvX znDg)cdJW@qeolmI=Qv*CjP+qQGGA|j7c@-jZa)mhK^{}n*E~e>cx`7)ynR9rxnGle zLpgvP>-p=!fL$u>*rDv=9^?DFRZ2rtF&})u`rGf()adH_FV=1}vFo2J`ZR{;ANeg2 z8oX)8nJhGR`Yn%^t<^rG>KPIg@5rnlefqW9aiC)b-pQC%6u?QBCmb2qCa^X5tr7h8 z7qXwn>r_rSC&oGz<7SIu_I#P`j3W}TN+#Kf8YO|Uy1kS<@!Tb%%GLVxJks4x{jWm8 z;T_J4*AN=TSHIN9=o*f2O~N%#8QAD+uh9zw0)8v6^@OoTI9VZfoaghl9oLsr)O5xks*{Lk=TBq#YeeK zCf_y4)763@y8{QM{%>u9Wkpa>jvv*s4eq@Fa=tAD_*_y?OjWr07M6TLs-SFCV5qEUpt&5j_6IgN z5m{>151m(li-?>}N4WhRrq~}CTn~0`cy!s7e9nA&ya5BwNGRKi>BRj4a^k_tCEK!t z;PyJZ&N|6A;1S|~sT4hG^0tZpoBf5-(3?{&RxOo|4O7YiZ@^Ofvr(0<&B{CcW}}?F zyv-|{od;W;iI3uNb595R^ac}?svJYL?Szf-R;PfU&Qnhu$UReCBhkj7r{Z2y@vceT zQHH-k<|)QA4-lE{E>?YC!a4k(wtv-+Pj|zSHjlNZQXcQ&<%2A*X|<_FZEe(gTQ9u9 zZolv_UfNR2*jd;on7Qc0L8sp) z;8cHhJA3U_X;f({ePh$)lY{Lc4_9(V=9h^&cT;lSE))EwGk~f7Ya5HF8$J%nxwx7i z{@t~besvP~8Cd~(F*&PuosX)bbdAsV^+Mv?F;RM!q8W>Pbd~NMtX_zY4{jlMs42G| z8s&%I;O*AwUZhqermeyY#NRNQ|B*(q%b8{6@{I71_t7k9`2#xoB0($GB2OnkKhvjYG=~K&SS(}(JbjjWb{loZ zn^?T8mCVJ>kkMsn8z&eX%$ZPA^D2#sCZrMT~tMpUk(m7!EnbTu4`S=g6{tVJ@{ zlCAY5AB$W@=;Qw_P8{O@1t+&J`?pJ+Q|?axhO_byPU=-HR#%ljIL~Yy6zb0M!TBIC z^Z^6{k+@tSmy7Iw-pVDl)q;8Xp>j|d1R@177ow2RbG6s7aYu3qYA8rTBxN9y|9+Ts zG3(t(R+3B51bLrJ*U=Nnbx|NGVUP5b&_LQc+PDaDIl0-mqp(ODE>GkWtfQ+77ff7A zT>7%4F18MwKyRceD=Du9(}bw2!JrT=b%YuM0h5xEk=9g~LC8p{OR8`w{f{M<6cV~f z7YBC~7gPo+`RCgMU!sMxsU}0HL;BDY`GNv5u`r`ml_h^CN1=oGA;wAK=d-JL9mh3g z{>SPx-n23*;dHh^;l!*GC-B4N?t?WW*mxf&ss4KZJyxCD&=@_o-hEtWW8Y4LD zqCT#TCI93*QrgRx*esLyU|8oOX=q^>L9XU`N=R#yLXhM?rOfD=D@m63IJ~i-qa3q1ho}-kcG};6_5mXD- z8@_|_%v_s-#THkfQ3RM@5Mv$*njK%R`&g}J=oHt=iB-0+FVe4IItVIzKcCIopaxn6 z_$@#CQfQ;fOp35(X1@Yx<91_?3b|T)JO(K?jK)VEGVOVFTVNC-Z-IQQX5cgB|9kja99^0D(@9+ zW;s>@v1!F+eThrQ9=f#fwH)82^`x4BerpK0)?UHaLNF=~TlK!KENij1;*$}8gOcb| z^FOy1?m?eN7yOc@6V6pC%N3b9kZZ(4QVG|D6X$PC; zZdEg)qOM!>Wy;l$W|E86jo1m`M%Gv>h@44B58kB~GOnggQ@Ja+5Gl}m#Gv1ee6>9c zVtFzmUGOT#FMg_jgn4L(h!x?G&j0mMQJvLzBWvlX5mUJk8qny40Bk+10Bu%|EWJFu zS&;WK-DHHMLG*m!=ys!;>T1nFchMlM#v!Pdm>((3Td`g^v zrP2*%;cC$hFw{|1$r_K|bBoN&u!ty9ZHa_E=Wn;+RRy4jh8$LObOMxmqJC$_gA?W+ z4d$5TjdWL0aJG9>tw7U0xi!=9sx>2{ImX$T`;8&xguk3L6TO!i01Xw}N&V2+UaOSQ8b#+kt6<4nh^B^ zc8e_2=At%{_^vs&WT7%51Np$=D`?+CK>7~3JKtXHNt|lTu7^ zKPB6!=S_K}13?|98I`o$;kC5GT;9aKh%d1F;33(Sz-0Hqhr4TZ3hFD|Rg=07XSz9) z=&m_^RiW)+COB{#N1$CD8>b`$h-;q1Df3R}Xqd?Xt!kNlMHiF9wY6IXnMt$*w{OzJ rW!r=xz(=A!=WwMN*T*xzKkDHS|D|mrdJs%UAtftA!N;d*phfWy!Y-%M diff --git a/main.synctex.gz b/main.synctex.gz new file mode 100644 index 0000000000000000000000000000000000000000..7a311682951fbea3e0906ccae9067eef8fd0be89 GIT binary patch literal 237361 zcmbTd1yt10*DkCQij=gpv^2=j2+|GG4a(3U0z;>Q(k0zUcXu}<(hbs#ba%u34gU4L z_kQ2^zH5D~wHIslex7}vea`uvbLPwpdC-#w_kTEjP0|&*k|Q+L(3166SNNDFV(hCA zb)c%$GfGL=Bv}2j@ZGHouIq(i>;4}B|BT>cQ^bVR|jj{sX~oc$FRH0l@_nt z*L$GGt7m% z%hS;_4urc)=t8sC)xl_N%}2o-sSD4`{i_D@WLq+?i<4LlA+Ph>FR4PFu0D6Srzew6 zihC{B*DIs3P%quPo$*0~`P%Dq6LvMp{c*ve?)R151&kSQbLK)aR=$ok5&H3CbMxK)=~Y2Z zi@>dOPmpJ_%eYYdfnmR-2TRmc)6nDta(urX9d(Lb;09m zAHPV#>uMjhXM8-y!HbI=x43uhZuG_Z*&R&iX7X<5?wa-PKny{4`AG0Y#i+5a4)o|D0r|qo$e)8%E#XH z>!n4nbCPL~|HT*5XY)V3b)?iPFg-rtPTwyYPy1L%$fGhdE7c-gs&amtxOjDx{i!|bhpl#d=(k#eZ78o=;bKm ze0jQ5Hn_G*Q66_JR?M|`k?8m2?01RcSd5V8-A>MtDeLT9$8y9pWe)%2uJFjcih78u zu2oD9x6sY$Mt@B+|GALyb>qe9PQTEl?(X(wtS&HbMn@l)qg@8?Zg2SM9UPC+T@0@| zm@3{~F75aY8Z5X+lI!U&4W!(zPVTX}oev7;@L##c;@nhCK3y8npI4s3&32tS@bEdX|kgpRS`!+H_YnxHdn3b2|yiOcJ_m+}(lc>6bJ&ySTcI zqwa36MG!jl?(LbGG(>Cv_b(5HSSGmPV+p}D9bsQR zp%$Q;n{8Ta1jb|o!F+|Kg5T}z%FEs3Zuck==w5=u0kYds#g^vFlm3~H2F)$$ZBA$Ka#z3Ym>4gK-kosJSog1h;f?POIa$Fnt_pL+y*#;I38C0fX0cn8NP{jlhu zoq=x$y0=12E?2gFxB7HLHHzmogG&PepZJy-ar7O8(L7=7JxP=y3FM zsgY&3g{tASV(?(Kj54%lu;u!Udm%S!Mii1+NbXM@#H+fwtf7dX0pwG8IL%!GIe(~czhLK6& z<$1K#u%)28OY$u0awm2{P^c$Nu%zK}C#)eS2U9WiM=LO+k6``-O~1)=(0@GKZ!22KKFMOiLhn3DJud+& za!m~x#S=vYl|?0q(7j_Dz>hVV-Iiey6bS0IEzZAGL=4-sSS*zltKcH_+3+kj4ewSz#BQ8Om#H}DJc9r1`N=q8 ztYPDCzLKv-KLTJR!VztnsU?k=8xd{Q-Fn5-Q@>mrAt2;Sp0VU4lxlk=a_cbHz*TpY z>h7V?eLpQG~=+aKdxP-@X zfx;a$k4xZ3HtZLXiCuq%u~V@nUMz-_tF~q>cKCZ8CeF&U21%D#3H=QoU~&A`C|}hb zcUg-ms=N34E#$7{W!TB%yMiB|eB9wQk5FE>ia(GfsMh76F<=KWGZNJD;P8 zg@`Vdt6D67S(YHc?oVpSE&=&;BJ5$8Sh|RQP~5cj&dmK*X_8=6!~RyCWsp8M1uSD5 zF*5PSH$pl;ZcLrxirovrToi$NwCiEQ5-Z5EXW8xkU%5KcEy zST9mjiutF6^}hV{*A9&3!+olsc;9XIffo|5>@ZYk!*SKj>`LpV@L%g-kMQpHLB15s z%N-{~J<@td}1pZ5zG64BV{F=&8=u%XU zqJNnu(ylS0BHkIK*3FaiBI7epudY$79_m4`c& zpQtIp?ik-UK%+u9-6U+|i*M)4bK7-&xgu6^ z?}LI@*_3T__KJqnF~y0y!jEA|=^{7dY&p3|K33f?EZg~?i$dObpAfA+cr1Z$f_BL* zc?rsSVV6Ml%LYvPkD-55fWfO#-xD8wnHF35cg)A5{ZzjlZJp^s%RR~=|J;AsrI@+ z)q||?6SKbb2N*{WhGEhq`742S#@Nsc+@1(W$O9#iD99or^3kK4J1M14HIbv=FsNWG z(wzzeBug|76-k|MdR|CX@T_b0;;I&x%LrGq^&@(q2FtuB!b8mp&wJ8j8!Pj=G5B~U zNP_NzhqE{h--+@obG}sZHbgURU*;Db9c0-L%1tl_9Z;QTk>t7(GC$FsEM#lu(7HDA zzgvpK^V?)Bf02PA*q<`}AdnPu(CSk`G8#2ZK-<00F<`pyGez6NuDlJ$qZ6$}iLCj6 z`3>k|UWsEGnn=@UxsO46?1)UCl&(gfp+aA-kKh~~j|ubY6O(_#waTqtaj!;_pESbt zJG}{3B_oG$qhgA)z;xOffcql5Y~K#*|p z3A>FLveg&p?KT0omEjQxmR3v`8Z&(TmgU@f8b|9d(uftwxkpkrQkdr#`1gn$tT-?& zWxpgjF9;M1kCT>Ni3jSIz2(UH@y%EI01}vFo^%|6vcg#EzlIBv^S`(hfBG1mpilMn zD>bQ;@Z zZ*WshWL~9>kfg?$1{^tAv~Nbnw?t;^ z$N#BB!i1Fg!MoZ6EVch}3vo2Kzyw6;MoF9*+OWf5L>VbaZCQT)R?@TUpyuKk|C9MI z?1>)}*T=kGey%Egfq?#PjZ$5ia$glQ0!t75?Ng?Bxe25#^qmH|iOd7(i{Jy5?7LqV zZrNPfibqZn1Q?qKZYc4E1I5{bc9=!fjzM?z;WuhxT^h>nSn7}I-0IL>G`${^+P#%e zq!nlk;&qDfm_d9e`se`9FS?&2tYs%X!Y&6(7}w;1NA2sjG@n=!Cc4W!W={;My6^*a6JDGotNC@ih`6zWIhg@D6&gI}n8;u*r??LpJedeZ(hds3 zr;$#!WO46;AL)9D>FsJ2l#kS8i-^jNcE6{DG^Z zu|A%ex0ON`x#k zsp5#hq-+rdF2xAZ*P#nJ(uei4^7J->qn#B+%jSW&Ki+rq!xUYm0)+a*hSCn zDE+$4Relfs{18ZDQ78Wm%hT5Mg#IwNHi42(HHA`=*Eo;_!6=YqC5m1-QYHI`mO}g= zXBRz+&KuZlUiSkJvT$4BQ3|eV`TXXP;4T3+sag9=P|E6ebFybj#cQ9p9*$Suygj9V zr!4+~?#XG_mppB~GzlslH)TqB{veEj_#Y_-^!w_EN=MluXzZ6dc;|#3iTxj*CV>l{ zHj2CJf3DE%3oNY3jyX-%!s1HDZ`50R zurx?%lXRkQ?`xDv1sBM)i0i6j2zPS9cK*lmk$$mzXiLRsQ`=}95rA#84 z-{!~2rMJBwUUg!%l6=&}=c6j7?Ldj9>rhgNM}t0yM*|*=S7X1UdTe4;Xz!hpwnnc= zRDKxoWST`Tg{RB=f}h48`G)c%tw+9B@ks0`1?Gtnj|>ye49H~dyZI~J6PFnyo){0b zDMjBb*N>}Y3#*M~&0%>T#zm@PaPgTDqk~nCbl;niaSZ1rW7AFeFuotAp9$vj=XlIy z(uNO`GZ-7r$jh4NLA3*x3+{ezY@I};2LE0yHlj;6R{X28TG}4HZco&J(w4=IlAz@&p7iSxEB=c9~mlFIgsIBlT_ zN`3b3Y8rN$NjCm-1l1}QEJ(V-_>KZO71Q{oCa%&jWG&H0qH6xQP2R$T=?b@0ZR>48 z)kK~)X%%Q7Ymjn86WVQ$c&@BG$C|Ao8bN|xCdB(mFL*#6riEMlgH)r&^s6jWQPAWYQcs5KU4D#Qa^$5KkFA9Y z=;yWR!N-ojO@6FS%SiVa>EO4c-J@fhIy<}cQl7TlklA4LE7}eHXqywMQUCe8DOIAr zG$_KXLh6Nw#D|e%LACY;amqF0(hKB~H*h$1dgoz_wZ>4};D7W~!iIprR^O|Jr$+hrtxS`YY2h0y z^vA7ZmOKr$g9aTJx?;Oz+DXOFDPI_Lm|tFT`}Gxn{BeEta+J^XEBu}A(Z`_<<8D8*&!X1c1}SUzUK?+}HXd~=ZdRPE-TcZ2o^ubeB+%Fp8Wf9K zRbIC5s=Xl6*nE`E3VG5A)mF8Hus@uP`T}7k>B{b`HQ|u9*JbmbNvxHW$4ZcvOOtQY z7t!LKy-vnEtSEY;^7JI4<Cy#Szog44Kj=N4%^1t6b8R@}tflg#5=e$aKk0{%5p6-4vhx~WnK(#qc0}n;CE~L zJE-jUCtUHorNiFwl#S;eH>9AAN7tK&Yg|6iG?~ghf#B2~#KQA2sSozzM0&z~8Lka3 z$Hyx>K`qpG%c=V$Jj=66sf#??#IdogL7#aM<)|PyETWJv7K@{cNsoM8cC*`UD~V0( zw|)$wE77cG^?dXni(OfrX6dx9?wHH1OWm;{cDiXxbe*U0CdH%ocIv{Pi+i?I1@j|vs7M|+0-Br(=0E=HY2VmkGhm^M_d@UHl5DH zya#+;nNXKf9DEHrzc&YL2KA^A|6J^f<{GQb`rJ4Gy5Z>+k>bXp(6@W-okzE}t=d)q z@_X0pfbWbazs=O@kIFvlvaRZs ze(M`P-m*l~+YY`e%`_@9)(nr;wZsWWWEvr9chESA63M-BX=8fyfp!jS0VL%?l9OX(@ev-tH;>}WB#ci8hJfwUvDu?YB{1`*3b~f8G z;W;XcU)EFU+gdSe(h}phHnLV43)SkgZDA5MupR# zl+Ms5EZz+)lD*ynu_fTkw^k?M#bq!F6x*q_+2l`&kp>enPgEjPeYSn2hg4CET8@v` zhuqHSfZ*F^5P}iIq0Z@0D#K-DYsoeeRtwQZ>Nq4}*&l9Y)A&ll{qABa|GUq(yd^vJ zHXCW{T~1X@-w&Y9Bvm#N(=aVHX6S2g+G%vnwhGpKWZT+?$2vTq(nTAZNo}8JS$VZW z!jGv?=Oom8c(&JgSe_VokDD!`j^4b+sKen;wM9g2*r7GV`Q!bA zixA|t+dY`R3lbCWh>QN@3wlJuuf*uyPMD!kGEk`q-{1xjHL)o2q=e5i_M?FfB1Bbw zUOk`RQf1o`K5s=Ax!#bX5y)!k`QQqe8gUV2zVpc(=9Frich9g4kFZH{$q=>~FVCw- z&g)(?vqx8!_8Y4oeo^Z_t)b^L<`3ztN2bUk*PAvA4|HkFCF&%e^;TC!7m2j*#807_ zGG4P#=k%u=?bn;$<1Y>D~Cm({L-L!)@uri&{0d>0FCv7Mde2N&x$mCIN7UY)R9~ zF~s91ALn4ngiho9w8m=iaf!lS>Qp2+u2qCQnGi4;7+Pm3cEUsiFFq|JP~5{xtRsXr z+jL!A_0fT0k=S{#NG0bCBVa-|FGVVyXBzF>Qq6FtetDOt8mmoODz*P9THjFm^I91j zw<3M6Hx4a~9w}zpki3u(!z)Z-M>UR*$h8VpOHEu#8}A_4jR6E=ZxGYUh^=Br)JrK%|aT- z2SWmZbdQy}?Ypvt6ocj|F1rz4u%3sTWh*RDas%vP4lN{y7rM z($W4j?__HX^x!!XSvE(A*}Lu4w`~gCP@-*n$jf)jt8ZI7`^z5Ta0C2$UR3~Ih6v{N zL+<`V`Nzit;QulsfdBAw-{mBV!e3wm*@>Pv{NW=5?ZRJxfadNg@b=;SLZID$nc-sY z`EZ$kl?}B{^_Ri<3{xAc_b>j6`NxOL{FnLvC=>qTuNXN0|4*6!uif;2l*t@9+W#3$ z`?U#$j)OqMuX9(Abj~*(V^c}S!PX(z-8FjkS@DOj4(lC8vkt`=Ni2sq3jSl)7^d4n z#NgX`H+GX?=2KTXv54en5p7WP?Gpw6<9b>}hT_UEYfiE%8Lv}3NbIaW(3~v2hzw#n zCduF}LJN81sGU_z!)vk-F(+YwwM;WPg-H1@2-s5U#%V(@pxQb8$@pPLQL+kkE}sV0 zW3!W52K5VOU5a>=C+F)pQQ_-&^MUrtYSl8DptVuMsrE_`=Fj6S*WV8kw$qQ8KdIEd z-M<`7@|MG!1+(Woac(PG<$Zy)+*GZwVr`8ULqGno#tz|wy=n%FEhV9 zXC{o9_&w`YUouSen+ORxdOiXe3XE?kCXyTpIU}2`J21b_fl)jFrqus;0DAPU|KLDC z6XMR)ng9bv@E#h57sn%r`=1<|-evzo;1O-;KL#+x&+i!zaiD&G1HhDV|L+5z17Mp>M0h)gUKxJY7ZQ!3=IOA^}a7M~w%h0g3wgedX;RFde(SMQ`wssz*efrZX z@C1?G2jDsW1%Mp?SKv#R4DDnnW=ChSl5yy?%F68UJo3>?8op#QgwHxzyoi>5VA2D{ z=R!B1A0~W<;i<7qiI7|P-nezQ>@SKxpcj3NG4bP-LSXv9ur!adWK3_5xkvBS`V1f&7KFmR}l6o(yce(6rYvuNQ0-poaCJjW|E^9 zpvQbg+#l7R=v`6lsIY`Co!CF-U33MxDQWQ9sBfuMoxEv)OguVJ?u^(K`q>}_d7`d! z#@drRyg4ix|^RjOaqUy*E|7kYZEpSnW}T z^PkzX*Jdd$k*weUDOC-6e zdpL6Uwl)Iz5Bw9#Ke*+)zI!+$+xvfT?BMAtI2@k+5sqUK0H0`$(_?Lm(}TnD0O(&h zTnrrkR}35ui1`Z_!km~oIeHhO$J$DC?}YO;0EO@zfVwT3KLx@0u1J3hg7fvg|Kxxd zqw!9Jl{#{36lbp5`Bcd{&1x~}JF6KV;fJeEYz7dRwW+oqwz$vrve z{n|n}y5}7XMka**G$0?a1IVZStD?;*X5y!BX1W#!n7&xmZ5QsdfiCjg&uFHTtjwZE zWEA}mGZ0T(H7_cX-2c6jJ2UEC!;p;C4uoo@`a015L*bR2m>;n5}cHjg#2hh=$ z`OtwJIURV)e{l95arh14{p8o_A{InF{idGY-js_CM0AyxB z84EyW1<3CJWrh>`okoDVex_-w4<=#+8Cq8L17?DhrAV_wWIEFd6)u(f>Uk{K+GU9c zG0vrJeGNP_!{FH(83^sEv-I`o3;i#Rq1~i0)*Okp7W~>o19e@S{U*KYMV1#r+alks zlKAH(6VTRRhEvbpUCBt0*Lbgn2YvIY^H)w{|MpZ!wvGQ)XpS59ou-{8>s)NO9uIIxHVXjQO4g!BAM%Wp0zH=k*TPVNNW9$8aozFxm+n8hf zMm~$!is~HpjSvS>D1R*XgmYtX+VTOPZy*$vkhl>-1NXiH!pz_hXb>R-S&2!mb7N9j z8~-GnBrMtm!wcagV>Ji7*=exkAR^+%$b~#-xieLPY&S~*S_i&)=}r5_$gvOm3d%Xl z?0G5&abrb*Z8+2AjUFf{*K~vb`R_5X3hQ1~)1#ZvYoR=P^k1|7rP z+@f727q#c$R>r&AX*`0a1+qg$t9JK(Ihmb5o}`;Ar6Tq4Bib7a&#qSlSPe{Z#5VZ7 z&AWWoty+!S<)$V>?PQMcx0KfSCcewG7pA2-IJXml1m2?)rIh+|l2st%g=zWt(iOfB zpeq#az?t~I%xsX%q=44~hrgIE)biQVM38JUM4yx8ARPfSPWBK<$T~%D zQtG);Bipv09C@HlEb>QQ$=Mu!7XRVRc^=AXQAgXAQD$Lyl1(JeD|89941OFS+vT=y`a6lYncRhOeZT{Y#6dV<$b0J! zHR1`1)jRD|*M>i?MJRwnRXVN*REeP&7wk(XC((gJThR@s%~)oS9Nz!Z!*uYa2N*6R zV5nYM0!Ea=Mu@H$*M|+j1EF#fUEn|PC6s?~p!4qGKNW*f!IdYP7C9G7C%r3D&_ z9~8nkehm$O;%Mtp!lyu`#o?pvD02n=HMhf;%oOmeH1MmWVQA5Jgvhj)-)I{jp+f^` zzAy=vJH$||yftGM`x%9<&qf~vI&pUEV3A>LsLtflkb8%vRBm=v#}j0Hw?lRu$)qoH z`hi;m)oto)k)-n3R4NO~-m!%lX-YU5x*|`wFCB?l2vkn83!Dv9GOD$VG75~$22hDT zFwO0Ok>LXhPo_y>BLzk#6c`m#IJ^TV8|dbHIM8<+KtJ8Xfv)}s2m0wA4)oJMIE;N1 z4u@w4`pF&+XCUK92D+6L4hOo`^e-GP27vz+1Bb)K{DmWxu#*CojHeHV`{48eVA{Y_ zz*_@m*?m25Ixx%bTZGerS#}QxYPp9SRR}fX@KtvVmrN;~A-?lFo%-omo;>_ZKdWA9 zN+07m6U97KTR?n)bRmFLg$I@jGQlc;!CywjPf|d~UzP#-vl5o$@RxVouY}>rup9O` zp=4Z2@?HHwo~aIsi0tKBF1ciSd76&IY#48UnWJc+6UzsVBw(zJOtH#(z2zAO45z;~ ztoc&IA0Df>Rq(TfE1QENR$hdE+4vBj0DHBn0N1_6I2&4-|Hq0Brlynu8`Og=>xH@SrH!!1PH@;-ctLl15f4)kn zdiCfOk?ECMe$bbln->Nim190BP0+DyM{!I#o}FCrAIqU>W4q5IG@GiTe=|_@+`h*t zt?R>m=H3$i`@oiQ*{0>0xNwJGl*z+FZWdp|XUj`6vsR44a((#}QhkD6M4li7#}hhE zVi_8>HQ>0{CeG)~>WtCGKT>SjOj|_qZZ(~ys7kz(1y%4@pB5AJ&1NVU1b(cTKAOAV zWrpMU=;|n?Arah_v3gwFPc;sv@QRnq7~KQtifzypb(e@zYTGBB|EIuGz<)6q$Qte$ zllz@`#eV`~@?GHh>}B?kIxu$r(fu#U=FaT{z^onsrll#cPgMXr2?x63RJz?tSpq zfW^=hNC9tcnkZ~K9!Mbv(5pNDq=3_bbHzOzctY>t5&f>CikoqSz=mWpmy!NwL+X=A z^t#`Wf6D*oKz;gq1c+G9PFgqNJVXf+z--v52Xvq#8p+9)+MaEPe2o{`C{p`vfVOO%AKc; zz4W`rzrNmXpP;IBYgLK%rU!ns|OQ{a6e&?|*}^F5jl zfUo!Mf^t)-2YD~T!b%Ouv!OBcYP<$AE9s&@P_}2gJa+#{tANo{|f!zMKCAcxEfj) zs-+v0P3(FMV(gSf@3d>vbyvtZd@Fs8N@V6ssa3rdKt6aLa%}6}OU7(cQt+I^ZAbq} zv?jsnPaY9v-g7*dz_y`)D7b+@l=qT@hRBR*9xY1=M+Qgz%ZrL%N+0<*!zeT7D8Y3C zjn5-u<2-I3(oHt$Jt0V^(Kk%6;cAGmlqCvyi}g^OI)nzX(91fKw(+s{CMEqi)xn$q zH8ZohLiE%-ed?@fe=!9e)dbS8F^ICMAt(EcX~f90W4!?FCl1+h%OYf(xa(JgKSX!F z5@_JZXTLLZ3<#3L>CYED|8NsNPm7OrlN5#P0DQM8L|{MYE&C_~J-u@(Ofb)kSmCl3 z-DT>lKu(%WP8!c;wC`rhJG{(T<}$OFgk$)~-#&JD%G@S^OMhA@*7<&VYU|m|((F{} z5SDJ|t%AZVf}fx7)~{xG5?Xy~RbX^iiZb=#^E#P*w1A+sI7uZ=iv8DOTfGZvL}Fr$ z0#gi%$VCH776Y)zg{vM+QRFQJh}s3vVE_S=2nI5K>ifWcf~azvB}NaPK13@2 zKghaH9KTH8{~#>|9RK^k*nTI)KL%bHX#|{M;k+FFFAxaDkp6E2vI4gM#pnk6Wx^S7 znSTRt{dNCs;ICXb<6j+chW8{%#8aTSVtA!pe{%QGr0`I4{-;r0SlOQdkm6qep5p%s z40-gJG=0r>R87s$SGw7y|BbOx7QK~uNpgBFmoLgUN%l6oF4no0evr)CX6~vZX|a7; zgMy>NX%68hb^-XbxG9opTFB~A%E9aU-G|2a>!qb**vnW13>W-mfkhkYP@cNd)7NC| zVF(yXhNl--7G?^YUm|~@Pw)_ELn=XO5#u{Ml=9Ro3Ptr&I0u`roI6x8F2fHlKkH*q zsvB8+HRd&&!sh2fjSr-NG5pe^HDsI;`n?J=5{-}_;%>_PEHYrM1e}9B>|K-`V;3!K zT)xp_@pj1PLx|2r^fa^5<|Xkeq(o&V6LdQmsI`?Vt&4B6-`AlQH}Vzu>){u=@dia( z+s%n4MRfJXG?~rgpO&KPSV&L49?p|(*RBIDM+$+fhza0Y!fb^m#muVAb8CVYh?=eB==9VNX`InSj6qC zTZP-Sn65}DO(b7AMjw~=HZNm)5Bb6OilzmTX%*J>z`E!!W0dlc9mGUD0xxqIMTJ^%z^xeS|3GJQD~={|j{B{8vCt z;Pih8EOMq+dQZBKFgC760v9v)xm^S31kKU^lN)#%|A)X*fZ{&}1KCw3oB`bL&=vg+ z;DvdN{I`Kv=jH!mgg^z&;S6|T{{le8uK#V|uUt6eUmb7;QprUo9#EYA5>RR2e{#zL z9cKBTMw`09lKz2EPb_rBYf2#CY396q^Y&oss;nEl=NAZ_$3$0bB!>s2xZL(6&r*uS zS`Iz!Sw|hs!iubN&U`w^Ql){tsmYZ1N2?#oxSNofMN9MGT-|_+4K2?NQBe5Tvkr%O z%upbr3$eSuXp%FXM9zm_#S!f^EnyJlo6kB!!Wr%zf!%r#!kyWWM z2t(>M}KeN=y-Yf#> z@S_5u_clfp#{3egMCraAvvhSY+_hdXE0lFenE=`K& z+lc;o@6GvBs|vS|!VV?&e>xqBb0D%nl>lXCCWRVPUYZ_+%J5A+Z`IX!fg9lqPEz$S zrO|ug3lm}K!g=pQ!$uMHN^BEl+4fokwy+gjznkI+vGuWaydglsQ`#MFIU=spL@}V(+3JrV$?U zf*zA^zD$YOk$RcZ&P$^!;t661qf_NlDtQjUI?`_fsZ2E1B=9|0&?_t3@Udu6hmRw= z?0!4$r2w~$G;YI3_m_=DqV)vB_R<5q-$l3foiW}X{IRLUWXPvbr-6qIWP=Ya1yT@H z0Rt`$G5#L`ZLTZc?{0WS8@Zks>+{@O@LNMu+uV9VLmw5U!zhyfe|Tv-+hR9cL?w!KoTeG}|&FYb$&b&)5~*yFnE$=LWctbmCh;NM0(eo zo2K49vKPv?OsI$!vd?#A#n45H#jzr6*XtB&ke;Rpaep}>NFHjyn8&A)vlo-?Zjcvg z(+cHykk1xnsO@2>o}Y&28Yn!j$BFL*Ae<+JxRxONfCk?4%Wy)!N_ zZdvo6kL;?0=(^Ygwjl$*VaR+dB=t9OnXNtNloevwhfe8tQ|J61P|q-uhEI|x-P}()i=(ZH)6XUP%7x#NkFe zId!Xb#L7dr?&S!p)*yTfeF73aJ*h2M!(#tx2-f%oTwRDIf(t$1H-8^EVWA$P2yBFU z5n@!)t9!svluDmab|GD-Q3{>>2HlxZXRHIGWLZ8h7q$Y1N{t+pLt~B`+j^nNF)PT8 zs%(?O!qN59E-2j!V0hap2MSQehlz|g7_8GWG>s22`BQZpAC?$asQp@0OU!fD zf>K#h8td3M%3MO@>pOfH%2$yaWsF9(@dlMLZKj?x3ZxH#a)?Tnc}1xaX;VUD_m{X!>YI;ZS7;3Vk(8(~WzNBvvZ09xb(5f}cbvsY56GRpL23&@_4-wu% zA%ef7xG%fvq#SJ~PJ$p%40gR19ul!U^9%#2lSM~T_$%R`(vz&J0tqLh5-mK6k&G~)z#e#4o9vmki(S+p-10W( z>G0fl1G5yg?%iY5`1aafS!i!P?c7eEzd`sSo0rt*(rFdMAp@mH3K8$rT$aKZR(lF zk>k1j$#^|ljaJh^=+p4~B+Hv`{8!46V%xpG6R_VWY zS%HB0(w(kgD{jjIAN0xMh3)imlFh^trA*zMtUwtn@Oj<6O*&f{SU4Nr8_PjwGw>$o z-gK7(?vvj)&<(y%4sQf_^LC#c-pC43*bY3oEK{ye%Z3L*D}*n%_Q)fyOuCU*wL@+F1ALBh;6qQ5Bq|=vIxBLSa0LQ9QMpg zs-FZm&<%`VQ>nLXeHLVLdVRb=&_D)!8MWAt@p+rxwrlO}ZJ;XEi85W#g=^Y?=9M0O zN3$Qx?$H%lO3;s;tgel~vIW!W9qH$X0>#CrOZ7y&L+^!L(dL@=@!c?$tvO}!<=Rm$ z@Fyeij4SejLC{5XbW=91QzHdwwwfCf8xa|t{&3bD>~N9&>QfrFDSMeOK};Auhoupg z<;;Tp?MAY^ES)4v`OiSI>g&iTt0K>kew6a+Jjlt-9vLHA#pN_A_q~qtmbkQa=ptFf zv5KN4)_BB)93d#79-q$yog9vj>%gosbkm+h2EDLZ)uAKW>p_VK$3c+SglM;6N@~F`mO3hv1)uef^2a9&1h38TW>3Dud;j?R z*4`+!yH49c9JgF&@A=?a)*7YVsJU(Z@+^pu+#{Pbyfyx3u zqH@aV+0V^A>0?u>{D=Bg<}l_sj?<}u@7vnpPGbvaZ{){LuX>}>>ON^JpPibPP_D;M zy%fq~OF>En>q;xka(}zfrFLeBKnyP5ddOOb@a!tv#ShITfg5}rcU#)eM13%*KXAF5 zZV=)lZuSHtv56Tk)At90=w<{}b9n%te-AiS2PT&WdaqK}?$0IWL_0J~&?uJE$#Z_; zLnpGa7kmTNGQug!5o|3BtJ|kxj3^TcWm6#Ruzf4iK2YCQB!fKYY6E-#qeN`;O4e88 zyVC54?Wfl)F6=746BF1)J))sn*nAPx_Dk4rAiUS#AP-LHh!#-^)APzMd3v8MMRiHU z%jYRmq>yeV?bjx-cG52)_(iVMkfDTR)S^N!2){)U1Y3qqX+JXk^5cyH2M>MBCzw)0 z0wdPxSUQaS8jaO_qTzBvh5AYo#1- zr50qb#+Z4tov%(tO~NQFv7?y6FiAo-o+pH5$xcvm^gzM=g)Ucb(N>#IP(kPFtr$Kp zdaA|hCs^fp`j)F=WXRD>wS`K1oDG_DQyjAxx3vkXx-GABq>M^v5;{Kc)BOV;e24aN z!%3_XqMuCE0h8!j+f1FeFMG73^m;n!KIAmZzTo9EHig&}J&~U#9LeuueH78=Uv4M1EXs)uWCa6x zl*-|94cyeR=>d*3Ch1f&Cv?CtaaSA%lIMb(YL+;DB~x)Or=wXlLz8Od1T`Q7Xjx9` z=mlF$CvkD_Wtzlo<%BCFmU)O^(IJJj9U}A%7nO%OA=wexn!2kXQe82^+dtL=4tK^= z)Mx9#<096W>Y^S{L0E;HYTr}6QU8pSxTG{oGA3M$TdE|a`7HP@yeU;*u#N~Dw#L(gItb9|!!BKpt z!k5)i?9&dl)rBJQIu(O$9-+=6QZrQQB(p3eAas6Q1k_kOl+qYLGaD&6w+P61JTlAy z5xzVZXQ_@`5-|$L6zN#?U0{$wXfmw&q3*C7+8i!GrShO`>oIEQ+JkwIQ)H8zxQE!R zDhX>|I0QPUC^%wMrWB6m5t<>)9Dvo`ocxrPa~Uqxxmulv84zc2-Eo6an8l?|lW|68 zPC=DQTzw%f*%VYOW=@d|)+^`kfsGi^#JNq$luTqQpQ1o@mkQ8?~&Vuc7B;HVA|4YJ$O~7mF z4g{U#?Mo~Mm!<-U|EAbUCMuwSH|JU8raoxT|VM)^GB0ZanPBkA6`_ z)Cohic9Fu94U0WNQ$Qev2O2X4By8*H`avUzq*%#X8l-OaQHB-@b8C>+@;EB72&lO1 zzRUpyj~lO~u4LolsOA{pvC|M2AB<9y8A)`bZo9R|eN$yIz&tK_sQVG-agldSv+zx^ zj5&@0O9O;vM_ewqJG00J2+kj&z0c^)pvbs7d-_iLfaYqGTL2Gx6EhA0Hee6c>>7!~ zTz)Ju4mr}qkDViPjvijonui=a|5L#3y+uNj#Fdz?5_;dw*JQd_U*?2lk5+hcc8DcG zhnP8GK?`EO*%Y3Jekhn%hr-j`jV&c^%F!7-H50-vM+c>WtvFMT4j0V#Im$qV)2)|* zp+cnTtZVC`)Wo}psM@)nX~HIByp*ks31{fL@Q&sjcaqHkh1Bgqu12U*D^2KrD0Go3pBPo?dTXieL1JoF>9O20y zj~r+eBNggmMxrxDt~g|#x@?YN*xWg!YIAAH*pA2n+I#v=^B0*47_9>iwierCqr)1A z1|r_bh%F)81L&00#}PmiXWBassAq@6H*-ep=3JVcJT6C>#HHEW9c?m-8*x*<-PVi6 zO%l~C%8A|R^!IH-8Vc4dYRHx8lNQ zl+xVD*f<$W**3DLb}~3vJ(klO4<}_*%jt>4)!^!)oVfCz?7O+WJ<(~G_wC)Z!kBOj zSgMOmRE>1~?%_FB7c{s%rs*@fs0*^%@ZN}Ib2y_#`4t?^hY8Ec#9n}Tljr=%&MH2J zZH8FkBJ;L|i&}u?c64rNB2q_NwSZfZ@OC^$?9>8wk_Svcqqx|bbGXB%0307qc`gT* zNpRYtE63Gh7fr6|@S}}6!kc<*Um9~vp7cwI+nPz360lQh%#pYNj)^P+up=Bk>M@{% z<9(4?23*1$x_TI8z}X+?;*{{Xr!3dJONa+cF4M#y%JC{)>^q)lIoyJ0=N{x}ykPPe zpgK=&I<91$bHwXYX%zskqbYL*(g>p7Warj*xQKD=#ktc!aITgxH!?ghxHxxPPq^>M z`?>qmq~(u?jkNAhm2=trK6ihpH)2e6k_x-mv73J<6|S)%3urD54KwWD+ZKn|Q8u`> zcPtLIAx)d0-b!&mqp;z7RUGQ}C6~E-H@!HDI|eKjCaDi78qL1KE!0pYXLMQGhZzoM zCuf7L1VPwc#ZhX3b!71%TvWFYK zE)vpI2#4A>VX9ah?-GidaHMDbB1fWGxOI7#BQAgrcr|BC=-{fgHg6N!7~!&>M8O7} zh&A^qNA!ef;&qO(vskF~wajI}{&b?1`!e8uW*jt}5@O*AC(CvTp|>3DfT3?VZ0&Vm zIZEe%+M$2SqIY;~V#u&Q2C1-ki=%@(sW8Md9@?4`x{xlOz)eCAd$cc(1keldEnC|v;>Itq zpfh5>DaeE<{0p_rb$yhh{7SZ92Pwj&px=k@0z&Y8M_K*QAi26 z(v_j<34=8;HwVz{l@6!mCcztlY8jS09V~}{QK8VQ@XRs5Z#r1zUdm#a1aa)lB)hGR$ z*3TsATpcK7a;y6U4vqeP6a8udjEOCWioVYYFK@I|@Wr%k0G0R(9{rYNY^TvIfxTGA z4P|8iZS6F|#4aC>*RQMxk2QlyB3$ugYKM6dsi}7wNXLqSBy60yUaE2m!0rfF=K#_) z@Q}>`9`E|aGz;fT!t)io`m!MqKQDI95tm`%!h$&f9iW;yAP3?)*n~UEkgJbi6FV6& zG;=m{Lh8PmEZJ?%Lb2dE#z~YKYQU^N2H*;?O0;o;!iy=+kiEH)k-Bk$=9+Un-nd=V z6K>zc2EsYO;}X^|*AuqTgxYnKQHB7QGOObqNIo^nj_IOBYkBQy^m5Wy@rkpcPLO0}$EzowEbA_hG02y+w*jYeK zS<9L1i;N3f7C9HP#eHZwiw?)miQ>CjP0@Cg+2F=p^v*G%#hcnTUMGy5S2LynBX2B= zJ#GMYMlB4M?0o9#vzwC03OhXQl6h0!xDuq!;$lk%_PFoLnG)VrE_05@b62L$0U7E$ zr$s;@_8-g~Py&3eaSp(Cl3uuu4N#9xWmn0#N)cBw&IxVGg;dV2xQ;BCfW_>wbufs7 z3wrC|MPRHkELz7=yOZf;XQ^RAEt%m$9TVdA0y0Xkg*T;H(w8^N{hSakDFdapLA>njoO{^$KExtG)Z_>? z1}ya_Ze1_^aq{bJTB99$MsIc~ri{?C30U-QXOD0bjcF*IA|d4lJ7dTX8!q5^7&2@w ze2=-P#yQ7;_3vInQ36=*&#M57h3dIX$7Vw8V%QXlv;%eNF1%TISlNSly%9dc8H|xV zibCup;<>#k$3i$2>Tx4V4d!?^Z_2Tf<%}1P<^UY=%d-p)B+?R3p909-_I;BywUP0h zxS6CTZ2+6q?drkd3>;YXlp_%-vf&pwVqjxoVN(xq3D7h=_a|(&#D>_dD20X}uKFHD z0kf~A>)9Q1PrjHLiuyS?9X}5@YQv?@WObJP2(V+ADgf$;~jwGSD z9Sw0B5C}4?h*j_1j@ZUmFU@sC?A?y8+PJ6(VhAaueLaw(z)r6`wdL`|7*21`0mm8e zMG7*?sl(Pe0MDqo9z*bvK2+t(U@=ySn8>CoMl+}6R)WqSu5Dj$I&GloXj!~ z-)Zcbt=}*oCwY>@>&`!u`GWgpBxl}zP6OmH&{;r_mk2uJ2G`EOd&cRS)kFWl>KOxY zV_(j8D#Kw884OW8=W=Yu-_av%*QiAz_VlR2h?t8(dIzV*_vOHm?AV2O^ofi#6NZG{ z?vYwAuE4RJsiIEfoFiGM$YU!?sjwKsoDgdZsVm%&(9tMDRZR(TKW9Gl;wGU(LTIcp zp+iFJcZDb8$D>Qn2}w)QWVbD&3uF=TA!P0BUy|_fbnzq0m#_Z?*hC_m>c2zo* z6FW6yjddy~nOr8;O9ee!PO!32eO{Easj5+`F<_~~Xl}vV*4ml~)2lOj<4?&R*6aA4 zdKlIP<-44mOK28+CvI)Ge!7CA2dI}bUgsLZ$%B$w&f*#}xMr*?5$+YVowMFL?9`qU z65pk%uWS?IXqD;0o8C&!<8b5@K#Z=2RAwmZs+5}j<)e4(@;Oy zrceLc85~bU9S*8ZISY30g@x3Gx<>38e@s|$XbE~_tcM!ulb?Hs!&9g(%~gt<8jM}E zuEViGw{KliKgTN9oX}&JK%Gl19;Aq?ii=w1xEIah@Y`JFL>IjYb?yl8G}nWO55vIW zK9mw~K!qa^3&RW4O2w>&^uHy%QuD-|^U#ah-EZ#$NeMng<4$ zA`n@C8AsA)fh;zU8Car2)}T%am&R*(TWmSNv9?v_M9d8;f-@KWi;K^f0n{8gY%o?3 zDSA*xb5%psMm;evW|h!m9X?(s^gS_;ShBEl!cd@I&8|(LJ_D}nn96|7O}L+93OHU& z4sM1zj!zrAg(?Z0)y_HM@y@siaL3JJey$EOK_e#e9TYB>p@CcGgj5-qDRW(q?$})d z^I#OkLxU_jufvLUzM=QRh=rSd(ZkRQsJ_E2ZXg9$Vc2fU8O`O`-jpz2UXb1<#MSVn zM-&xI#!cy~T(o4#)O8m!(iku)?q#GwOGf`fMv(W=)%DO8)SCj9^{0eH({Wc}+97)E zS-{cEv8_BAROz7UN?ZBkl}@-1r?xs)8BCh%f%<{NS%R%pqQb|7W57zchs$oM#pdKI z&tswgpuceRYnn`oMFPaeBm^pD2avFjs%a#Lg&aF4^klN(k%H@l?YK5pDX5udQJ{`$ zm#{krkzhVsXnQX!uzpp2v)mO!ZBc}k1dw|t-2C)dM7x#tOOCu^cW?Q4Gr_QlwAsqL)0ZF zle^XM%(=sNJPf_{n%BemnIv9s%(}y?IZ1@wi5Ab3UyIcAIrfp~OP-g6%?c*nOVjui z5IivOyW{i z!f=?xm3xD5?L@f&JV!EI!xB8>GOM@(r>+o8+%^lYDDY*Gut3}TS?V@ZVaHOK>JpFm z7`Q;R;vCmuIqw$5p#eeKtn1(1Y=%qI>sBg6LAIPd3x8gqfEKS5ROKo5Qgg&p{~qhm^l%|LKa?|Lz59(Oz~J0(2s zb@FzLc&zhVKlNgqt@hvmT9xCp5~^@fXpBroA7b@8@&5de9D`&^Aw+K5BQ?flmJg0 zg=4B3Q;;F-v#oOqs%==+EudR(!@l*@!3){&HPpUR?Qws0H`NeaDRaUQaoEFMC8Vz1 zPc=k4=8D7JDgKzD$AFB}s5#6~8COB2IzV-s6QSGEqCzTey~(ricxr@NeJecH*@ExF zTf$1Sf66;iu;sj@FpJ_m9=JAh1JK|}H~_ZI5!=3US%?ye6`P#4WxyRmKF=NNxHJK0 z*ry!x@rH%pigH-8V60=pa;W^c6$LxOaKYFpipRazxZ7dO(OYaBcVQHg5bv+K}r!du> z(Z%8o0lltegInm@LL`}9Wb`%^QApGes;g6W)FKWS;r7|lz-(r&M+N68oh5dj;f=bM zqcd|8ZjG~(AXKMNPJYf2>+jfyy(t;Y;DD>Rr-T8Uc4OTp#N&rP%vFiYQsJfCmLXvS zJ!~OPC}wgA$K`(%5U_cCZ>gR)&2Ys2Ry#5oNWV-Kgrk^PO`gQXvosa2ns(ym*!f)1 zfv|~aN@!`m!xoNBIdR?wn=+?zI>?TPF_E1kUkljjl(O5H02IB1QP-SC|1(&jQowZqDco%6= zT=-0}6=K&2q-v1rsir8%meobP^J(xdy6B;BO|6M6ug0n?lLCCIs=}@-6h0x9-i%9u2G&Tu6cE+W- zY}RA9C=87;AQbNoSK$a1iTNJ0_XaBB8qHW132XHgGv6j*s@0^vIz>fcxd7xV;OhJ1 z*{|I#>84hjB;pmBCc@(vjYz)}7Mdn;vFeOd4m)wl-Z);cn8dB1-|1*1tVr$G7b~U0 zy}2B*C(BqktFyRRUZkp+YZq-rO7pr0)NUWqaeZEy%Zb}=aLd51oN(-7@77dKc<{2j zjF)H2NhVglP{g^Mc#Vzbh*_#w{e{79q)J$H>HDqRj}a`1;dH4IwyMXpl8USWTpTrL zo*hm<1U>V7AsMU8i^bi?X)ZEu9KoKAU6H}*Qw&p|Ai6@DE5s>B++&oiwYwbs@vL?U zb3$A-i+Aj{2_yE(&3*samyAlHK~hYI8*^M;*2id zRJ=`$)t{KK6KCsvE{8?vBwn}W3kR*CgE(ilPJ@ew1Cs^x*mExXT(IaAFs*a4W_;qt z;ymSNv1=>EI%OzVDdQR^G=)+Wi(99Y&x^9)XgIXv41OH53T{^#3%{^QW~yI{XWUH= zx>O!lGy1ZU3j4{bnX4E_-!qo@cT#2CmEaZ$GcJPcYX6e3qvLq#VAVRLQaui|&jBUJ zV>xWB08;um^nC>2RO{ZAN^>5r)8dGo3PHmjPRF4f@OXMW2HlsVbSsL%!22_ogX{_{ z|A#q%6gM7nngcRPaG3U^4CUCdJ=SG+Mu_ZGvl6~iiA%NHg!O25yGV%bTgC4ZdUH5q z=fO@yA|ChuXq^#GOnX=$Q)h%8!42V41yTPsc!FvUAU+-NTFh9E^Gu@ImP7R4m~aeO zD#lcBThC&3YI&6zM;C<$_6H6>d?zai{*CsX)gAV?V*X}WT=dxSQFpnpZOYH$I)Aqe zIF5+g3CY6y>FO0!bfq&$k9~rR91{*rc2(2#1h%^@D#2#FD*+Ge7=Zh0!zuvApzE`e zt@c? z0FRe-ri4VBN&Ahh-=9nj&UfKW-4xs~>gNF5ikiDdZt5oC(!scuu|_i3;H4fD;{7!o zyVxXzm*HZ952E1kZ^gZJQvi;pNy>nM6$s_yQa}R-UTyLvVun+yYI*KBi(*XV)q8mT94_h#gPl=cX(vU z;XZBG63Vfpj2qVy0u0>r`(sHJ+rf-P?3YQ7z520bE&&DaJd|17U`@95THMq>uVZL@ z8#gD69yY1lRuFE%ry|V>Go`~r+q(+N*fyCKWxzG5uCKYjZ314yWsDOW1)O5o1!)k5OZ2COlpNY#0_}Z%1*8M4%mKK^@GwM18E{dAjdm-1H3gRh&k6Cy zk@4FK#*=P2bQw2iaKJk%ZV^zuE4?0-Ny9*k&BDvQc^l!|+%P5_1D2CUHIA$|ZB43F z+%I4xdegBEM;yBVS-cMUVgZjSsEZk#I-gRw!%ZIM6h0SNR&zkWn}nq<0YW34 zpyRcYV`mPLWttP}{(*!-mIdnWjM($(65T=O#{wL;lX*XjLlo{flsKiJ4qJ|$j$;Zu z5LCP7X?h}=5Ec!ACF1NC^$mVvav#thD8u^`ZjplQW!JY*d#@1>#~$UtCGF2VSwMB{ zoRx@Ii$XZpyULtQrg&54$F&*j%&R$`N|?$@7DF;F&Si}_abxlvkRs0AFKQ|}YJ20*9Cx{0CgA&G zcgvElk}i|zWIj;?v`a2Y^JAxhl^9c<%t>D!;^TjR!6~r%;$QW}U#xlYPx=>zOvNpg zlFXxb(14UL=QHR>EPgWnO!f*-f8l(o>E284zw?hX_;C|_2>7xIGh}Cecy-(+O5EY{ z{p0i<>shHB`P^2dGge60-L`6

xG3+#EOH^}U`I5M9AF86xjCB45TNJ>=%de_SGC z3u6LPzC`AVzy-kAl_{BE`No=aE<^;J;Xiglp9#;Fa0$;wSwWZhM8S?LP^S2mXc$&E zhVR(Cb8&%P=TnH#*~9g9^<7$ozrXC8P8!9K(U@^hnKVi`7WC6u*VOMhZpLOkAFg(` zbx)1GU&p?>oi)N?Y6=JP$cm&i@v=9k@fLTuaH2$NEw)hkT4{_*GbTU0~on@S| z!-b$Q9yh|{5G0j)N|x(!=wl~*!41?cxcRf_za+=}J>v3+JbovODd0`$5#GTu=Hm|D z;bGt!CwQ7cJ+5J#z6U(KId=nxdpuzMkMJO(c+zrRX_)MjS(N5@g*41!k;&{pvZwFI z4P0|RH3VvwsV!Zl(MFT*rdF`W#^6-Pkjx*4OKsfrC>iWyvf+DgP4hs-dRPrf@xZWJ zJY#bvc+rt8#UBb$F< zSSh+n)>8uTD>Eg-Ry$&AJG4Zd%*CmGaOoDF9o{7?aEcMB9m|Y0Dh#kOyGjf-wD_D| zTLnAL;%HstxNrtH|4w}aw=!UR=?Hwh8bZ~*DF>cC#uCYt?3k~k*(HN2uYT$Y*!Xpv zXdJ&MblkaGaGe&OTI^wnY_=`p2Lj$?ni7%u6|p<=Dp78VB)Zx{<#VcJ z|E2M>#p)PtJ038gbx0Dok?c5lfC?CsUETn<8`;Pn$pNi9Cc|@@*crM(Bg%(|!^TFT z?}tZ8YZYAiejJ*fnuR(B9yYR*K-szD&X-L&a46Qb;Lc>h3dz)B7%{l`ep5b&w{XZ5 z8?$rAdSXi^bRWx!zGki5ST8>O8dJIPL=H8(9|d)|qz`xSt;_9jDLpO`7?TBGanr@F zd=XnAQk|0FEjt`-+o0j-4O}K;wH(hW+F1(XD~FLhNb#@sy)- zP-yr;635U46h9LM z>4t813Kwr2rKwYp4Nkhihpt|6gk#dlW3u7V_e#BN1HO>NJ z)}sU(wfE*nvm?nw>6&MC-++7gJrX7AMNty@^M12QJQH?-C*LASJdYc4THaY9aJ|Fd zXy*l7{tOzCp&;g|FkmT&_lvAH2^n;#*bfQCx!|&CFm_GA27hg_#~Q`I*x#FsPc(X( zDV(QJkXF7HE|Rz6aCItN4#{aPe289=lpqaQt8xA{- zbC5}>vbDH$;Hu)BymsQg%Ewa9;9A!@7i+kPbpb;YR=K zlKt~Wh}R7{%RQ;%N~mCs5GEB5;}d!nIATKn4zdxR!mEb_eml+OtLdG}Dc1Ts--2=# zBk-n%A?eDFU&DnVj{CW`Nj7elBB!cm%EN>XbmXnT#b#3uSonq;LgIO6zZ1@fVL3Au zuBpHEfe(bH>gCW0$U=3GHI+xahx2t@L~c1SDs4ZMgnypxS3IveZxgo?QBl$13}MI0 zGNl`#91|-YVF(H4Q>u@*p4ZSknBDGe;1w`pceT$tfZb)i#jjm1XD=c))jT(;K!5V;lk)M4uUVTT!6<18} zzgAp1H6gK0#lWO-jcY0ijPH*yBhq)C-)VMTjB-TfX)zlz zJqWT^KFAg%c=H}wP82_#T&xgO$6Hka=>13b!RJ z>ZP2n%o9a`z;1(Z`UlGNeU03Y(X=NV1))J>e+g3eLdg zaGmw3QU+1Lm@Mz?n`E9^>p8sm8ccJEiV*cIer_+w$I^+-awQrLH~`ZF^ zc9tt{c=|e9rep_JGGDpk4`nrj6%=iW8YVpMGjRzXQkrZt z-j)bNTy8c&FI&C0RpFu9uztEQA4L6E-+V*(@(#T&O4&qO0!qNJ5`l8@c3OyHi|6=t zT8OYt8imG`ScnNy?1Bz}BVkB;D6!PlYxacD2NOmPh4oe`a-3UDv=s8gP1wX5lpzp0D z`%>^wngtFNswl*Z=`7YlDu@V5D`(jMx)4k&<8yI9`k+m3#r;1Mc!FJMl-4%^fc>0S z+~ZmHJ6Ub1YN5Gi1qR=*r8W(AwPqD*qA``9tiQIbOXFTyJ#p0ZI<4Z1cqMQq=ku1% zV|r-eQ1-*dL-E(8?7uw}>umT}8Eydg`qT8mTnCXHWWDtCZ~yB|Fnaa33)*Vt>@M2O z!N2siGmY1#*7a86$ACdl^S^tom)N(2_y z+m<3HtSOzPkVBP54lxwG&Jo|$P+)`$mZPDN@lq??z3;%1BZ|t`&XbPgaVMX8uM-R* zHp>nH2tY>I*$J5P5rufC&_JQd+a{~ae_azE3Q!%1%4jG!>iKY>kS)CEP0vnI37_6= zx5`Oe4m8bnt6CBIbf}{@8MJ_ov*CxAZR)aKlAGBpVYi2dONhxlQ-_8N zURRoxOeN6q1nUg}?}pqhS_nDD430pPk1uAVyZI+UY}p&-62wnv5n zIIqmjp`~a6NSc|juaIp<+XbRFhHiykeES7`Y_W2Flmi1iKlyPd%xQV-tE@y=dFPlj zH3M0bX2O`MwJP|ip};4+z5e{9gI{#u|0l&(?*rDX+_`7T`c^3y#rB<*z(-5IU6EXheCd%xH|3qVmwQQ1 z2Rw6k5Ij9~O|YF7)NAixqx71b`?2N8{%ND6fp6PLomyjQ99x%4B;UA3EAqWB$q!k{ z((n*Bvc>wmDoFHBCTqSE*yK`M>tIc~N%q|0qPx_}cgGvy^%gf0H`$MT3)~&o`>&J_ zhCE0OU?X1|LV#18wk#or;oOT|ngY^^+E4Fok%qX+9qPBnaEVgnn0{!D%oOlYi!Gct zR5>_(FPylfxNw{b$6Ucx+vEh6s881>R{+Rzw1kX$FnzTko)k>AexpkOR0!uT!PYda zU9#BEyKJWF8Z27$v0%BDsYHxU%k|i!dr&|HLx{$fmTA1#D9Ritie~hJ2DZzu)kY|) zy9pK%uzepHKLYZ(ikq?FN`@Pp@hewZbc7u3nw4`*iY?V^Kj7&Ih^T2!CK6B%Ux!ta zi*K*7+lCd$d|5A600n)2ZQ)^g{!NukWL?AaTj~)kTiZhf7h>31jiKIUNwj zMO=ND$*T*W;VR;z+v;_6#{HHd+!IcB+oBp0r{tTtny98+e2(nXY}8fU4rei4j=YQ4 z5U(HAkaZjv(B?nL=9>?8#~?@Ubl$NEFonWWRocoDUX)97jNQlGUX>4ZWFKNcb^&6S&SJo8pAef3)BR}3yiC1IM;?!YFlsz)3?sevbQ0*Rs9YdIV03r8ESW=UX z61|2$Xon$@^)9E~YKLSP5MA0i*WpgMCma`OMt*N96#Tm z3pz|;qs`A{ie7$f`K5X63?@G>-#N0qZUkJHCW4I`3Qo<0w;c)7N=0_uSBv$qeLttD z*<2hMS2r(!?@$Bw!b>dN<9{!ld$C^fsqjkJfZA-6xH6V4dM20weg@Hl!672dyipEP zI{_+wX#Hx0upjlG>_>N`Cv_b4D;KYIsicZDoF-zjt73{j$ewAHxy!c2Vet=E&}VT| zD-_+2uxop9B6x^q)lGRGd@=%Sds+jy;<61=fJ>z|8Thuu* zrtdA1ol8#arxpR`lU48CSoF9+;SIR+B8M|xYm;GyQ(FWR>>D?-PlcdI8Osh)!AmHI z!igZl!SY+-Qnr(Ye=3|Sr@T(ZAy-rDdd%B90Ye6`I}JM-6J8g-vr`=Q^|e(D9&Xi(3}AmUWQ~C5(rFj6u5)1F`jHaYTqorMnAq#pi|oH#JC1y7 zICO41WD4vYc}J98kwBj;Ee4>0cLm!6o@aKj0QR;FJxe5>D3M>8_IUeqgK*$F5IQ7z z(;>WD;yv}faNfCPmpT?sB=s0<%L1XNa=f;TJyq)Yjmmq&E%;P9;dn!`6I4ViS7F&E z$eBr&Nwj^DTCITNeq~1vBDvQiI~m8#b-fI|G`WI_)@8;LV^P&hLx_+{%fj6WEAKyt zg-zINSMMtu?B>iJis_&Z3|X~Z$~$Swt=R-37vPh&cWLk#Cwc$1OPPj5IyiJG7na^P z+rXe*vxL1Do_;{`A*`*u}yqSVGx$9U%g4gtAFyUXctd*>RpMhXle1d1n(uY2^KRCXAaXf-J}(A%cnL ze?XZlRU~JxBjb9_)xff_BS)m;T>EwCJh8BF#Ww5$^GR?{n;;rkZg#Y#ak8`4xK*`a z*HuMvI}qY!kgoGYVM=b{Z)*RO@#KGpmlsHe^h+j|g$tyJSlW(~@P#e!OD}GAmMGQ^;OUd9Sg^&&< zS5`Ee$<`z}Tj58zHEHVngm7=kFw{jiQ<3zAp^o=>IW<+}V52U>5n)th>Q~dt^AeQD zoo)$vmI0vR(iRVPUs=+poXlQh@Z4%l;5qwb31tBmBE;SZvnw*mZTE?4idDq-{*x)% zfO0=|V0}pl%ayMF+OPc}ZIP`ePQsy6Vm{t&i9jr7Fn8Ay;{pPX`0vp!n1IQB4tB*% z8HUM!HXVe|z08CDwSxkNDnLBj4ho`gmXw2Sr|6rl)k8bwm%_#TrRU_M9JH78nIH=# z7iSY1pp&5<64>ofi1!4M0FgKPJwdkJ+E?a`Q|P6Wr)1m-_k^R!jFDj9hxeF|7jyOj zVu?Q4hvafR7kM=ct;t_+pi4mL=8w^X5#>XBfHzq!h#o`G05naFf_e4yz#DYg9KPLtPnwh!LjTJfdaWHSD zM1@-;Jd;fj9UD(glfbJhC=7-NqG`ytLsVF%w*nJ zWf!(nITVfgUamdU@RAL7kU}HsXSxm&k;oB2mXL7BINV)ojI8zOis{)YGqw^0OaxHg?O0{>8Z$=d>sLctXKI9%beE{%9$35R)C+#(w39az1IR5`{|B7aC)0tvRlq`*u+Flv6`&2P^vgZB zj)42jyjS|pJP#Oh078g8--j8Dtjo0IUtLQ@1aW-biW1$IeZHBp2y`c^QLq%;zF~&G zC-7oZFxegnW^kTH%g*U}iMFFNIoA$bqCx>L-r~=8Bf(~09%jP0ae7i;9Lo2K$4#45 z(Z%+dCWk*O2NFWOJ!H>!kP zrFS|Ahc{L^=*TIx7+!gk$Fd?(EbGqVlKeuTu+9y;#>7@^HnRt3H3{s^|k+ zKXD}+V)sksqi|V@-G|@kPs!F?JN75BEYN;Tlbakp+bZj}CWF)74$_X-b0U@B2qSkt zy>>GpXRx}hM3XR;!$pDzk@-Dco&5g!%8q>vAk9zq=X;gKN8pFg@8M5wG_q*2=6h`z zG|caBUcp$C`8nC~98AL^c}1&smQIcRct=%DF-Ve{YNs#xB|#;h78}4#>HME1|5k&eOMPPHqmr zpzqBojtwWDnu7sE^UfAt33$Lt`c^mr1uB!<7LE~$?h<>Gd4i1DUz-fxD}jslp(V7jZ?IrFBM;WJZc==~Hv^XNRJlThg6J>?iI^+9akD`csc3O(h$xpf5O9{aG zeHJI*xr9V(+KQ8X*kz49zLwV(B{Dh@S_eWo+ZIBpd@>RjK_$PVo$iF8@KhIXEkIh~ z@nV|b)$ASxZZ9#5YM0Z_$}F*f?kVR~Q=&+chC!xDz}_{VD0(5->Y+QNgGCLyc1HkW zxpq6AGFpzXpN9EgBfZrg8|l1-Un~~Ln1DA7BjchsaJEnDAdbQb(;M9P!bNlCed1Jj zkQqdbEz>fY0kS)|YnKFEBG(x*1e+A78jwjK68m9J5phA*#b;G=z?*~}$HihxBo<#^ zX7LnkDn-MU`O0&5JFtSQa`e~fEgvhcD2A2jh%0`WTB3s`5LRxiL|GWf7)ywGX5ZKL zs>3CYy0)LEeETRd+JvLvOT@4(Kbw3+HCe_q`N{se0N}*zBc{pY`y;^uk$BSl{+`3Y zreTf4`5Ta<+Z2|KyB2nVcfR$6jYUkhtf0XrE^BvQ?T--2R_=PD)Lvb+Dt_};`B4}Q<`d{={Frt$tg zfbFn+|8@S(O*wG%qrQK>^I_S`Ez?8)*P9QAUE><(pV z$2Oj#ScpVR2SWV8tKTQW=(&K&4S>d`oms}Byt@47c4oBG(oXHHctD0`xvD`Afmg0L z`SToY2toSlaVMPpfN6tn;Cs&!$5eJMlK|XpX|0&5A_9uY)fXP8uCSLlgyLS}o-lNU zRcKT_#kN&m?z#d%W+-n|{?7i?f$f7H#-Cz35!fCoIsL1d;ZD_Fh~rD;pnk_eR@van zBpc%F6^3Cij>}$G;%9L_PH`r49J`P~_s9?#B9C_fs0q^l=lxGJiYn^3PI_ltWC%EK zJ=!+<(0k|g`;)6MDi^I1Z0GNlOKK}G_V>zT>}~DHDUZwhR$!}qh*j=hLbinXT!5}~ z>q`tofHv93MotB_wjOPXisIn?R^lY=MOGgNo1L)BG6~D<*;om@p@*f7Bf(>PS-edM zbk~U@GOt=UJjjqFr<(n_A93ZEBdndOm@vp0y4i%n?X(ZJs!C6Ubgc^8E9~D)i9KL` zVS8S*uc#pxE*(hRm(-(O)EHlA9H$E_5jXkdvteaQoV~<7;V6n^p2#u4Sy42-Nf9}` zBOxc>*{86P9ee%$2pfVFqeze9vuIF5eD=mwWHrTkZXv;**vs<{8)GKRT40=OOZ2*$ znq|b}y+{z(os&l*5IQEj2&JFO*$z)LcEoNP8CP)>wDwu~Q+NPNf&iwl=DW(NmI6lF$ zO@A*OoYq?0p>TQ%Q&-bu(OHV(PhOe~`ZEB_mXJ8t-rWf?MFaDDmuaHzVEq$pi89zp zFZ?bg0`!V|v?&p6ncD3{5E-r^yKce`QNkJSYtIHCu=8tT>^&>GcO{Igsl0g{Gt;#> zKrobW=-JjYE9ypwab(pK#x_x+!{9xA7q5f>$Ury0Rx23$+2+V{#{!=6mNcgWK8dJD2dem8GC%zYUFh zt%p%h$o4e1O$}?>NXOXx{I37nn4LOLKN%MRz^}Su_R8`92LXY%g0{hY+M0q(JeJT zU0!lh492fr&MhXv1`k~>yIHRxSgsn^$zdCBH#-o->yMfEaMjcqqkreBJJA$-JaGkt zOFm3+FAZIN)#;#B7m~cwD@<=ESpY3>s)Sc+W-^WYZlAId9*`7lzwb*St zxunIe*|6K3z8qcIRX4h5ms#Msg@~G$4kGkp4B6I*^RScS-K|D00$M$Y3f|2fj(avC z%X!ru5{AIvdDs(9J3$tSC2J^sy9)}U!AS=_*&mHjje=y|j_>1Oram+C`@33fS8-7w z$bC-zY(=66IPjv<$hX{Y<3J>h8YS z=rLy!1|ib5Lm6+xNfp&ov?WnL(%~r(gNa*i`x3b2nDUM*)?hxPCiwaJ*YWf16;Srz zC`!nne(u@jget9=P1@c7cNQ1T7JR5=3S?`cSC$fiwp;XCIsT`-w?gKV99wgj?^RCB z_4?`-2x%1JlW$=tC9i(hJ42}e{t!MHS}qO%Jqg{A72K{$xwWH4ubc8iHomh1%SzwB zi&h#ro|*5FgHABtd(8?-xUq$)LB&e!%oZ_R6HggC891d_H=R2Hk0Q`K4H>8uJi}aS z1C5*u@tsUnx?XR!;fIk+G-Oz~iDGc8O%|~nPt!hi0)an7yZSbDz#!IpeKU1H#wHYn zVMxxD)w^(Ei0f>yH+Hg!B8W-Vx%oWEMO4de$RlVB z_3FEyRQPA) zwiLk*o=lEoSHgx+#y!eW=K{c_`AWm2`pLJfLv90IHqDoFf%JbXw|0QKC8J{)k`rvk z3*Z|=#4+I@X&J&TBQ7P6vmQ8XuH1U4>y49{lqwj~;k%sMMEUA_H8H5uYQ3~OMuiV^ zhAd_^V!3b^F-yR8&Icoak1IfOVITR=wSv$rSTqDi<8wr1G*L3bvs=g!c$FQk|l zzE%v>1K<$0VlJ*3I8(6*PJ9zDJt2B*#aVMFi-4@eYX>GrgyA^I#&?2SO|toPRDh8w zU)z^>CGhUhzQj}WQsWTHvLh3-=fY?qT6wkY!%G-*vJbLFMI-ujONI<95z(V=wRK`| zwXk=KP!NuxneQ@pD}du$HNn&VzRni#s;2Jb@ryQN9RQ8M!|2SNhzW4Q{609C&~n-G zZE(20ORuip21gDsj&05`gerp!yf8%6znE=1xD9gk+71l&-5m1K4$;go693u`S^8Xz z9oo?{7NUM1CsAc(zI(SbN+eTFvSsbkOS@h{gH}$HQV=-v;+%P+H^2f z__!zQ#1~z%Pr1c|r5yTff`V;Vn?MxTE_2-`G+Y|XOc?4I8W-Vo2d_E$#ey;Fc|2M| zVp(9(@<|w1se!U$2{X#{NAG;rm^1j3&bA@TN-q77AR-n^)k8v1>l({L<)$(94<1s_ zi>T>%;@J`#*E6}>5_=F~7~{279h!Peh3w%hHxrCX~BG81Y(l=xhj?Ndt>K$%@Re;r-I^ENyi?C1_LXV~2njslQ)C2uF4pY%l$ zPYF5q&OQ{z2moSTi^2YLiop!llbdu?;X}XZHi0Wqjg-CBgzT4{A#YVJcIlWdrP-pX zM9J|EyL8VDlDj|<_4mORf0HNUqDJ8X<<`QNch^i5)LRG0jvTR#FoZbJ0G4=b4)Lss zZe*H+=ELLN*XD>~ET@@tYK}~MqDh$E4-PmlXW1pDJMoro?Ig5&X-jjau2xe3ZD(%> z7rml+eIFJKg0a9ov?Jhs>w_H@fY}@h|D|_D=nC`2&axl}D?Q5G+L^diU(!D$BDcV| z8nXr;_O^p^@n6H!uZtETYWT4iMDmx(lGsF_GpRGP}Sgrf{4WN7S%$JVn9 zePE2#slp@^=l z^Qxf$SX`B#odSfka!imDNjMdi)K>36QFx3%qunV=)Jb6}Dxi96QA3QDCil+;AWMI-f89yw!IZ@h%;?6=~v@2IUesJ|=2#N6sTQh`}M z5Q@T0XzNeHJl_Z!w}UB9!sX@&bVHMXXN#QWim6IxGq2qj9*L-hv92XdHAMpF?1b|= zNRCsne;rG1v2i0kj3qKmb_qMf@Su08UT2uhd#{n-Og|8AD=uR>Z+$v)Jbq`l0d1A) ziQQJcqG1|#8%lxqH+CbT9ZnNL0}&K=ok2wClw+lBCy|k{&9({86}wG%gy7D@o^agq zC85W=LBhwqQ5f0hJv$+p<5EoJc(ELtT7M}JpL8BfGd+iV5H4~;-@?lzoC2h&m#0Y( z0W5#oAweW-?gt(cKC?@scILAKzq;BIGw`S}+64A=6=8cq1%-)uvL`gzNmU_z5{BVu zXap`JK}7BkRt+r+lg^Edt2wKN`HHsX*r>O>58L9m3m#n;r-2K(SuC~ygxY&ngqoUm+62#=cRElI+!f<}lMspD;B8+6 zHBj}59$%S-nM9y9_)ab^$zf(d$^{ACbikRzz^hd~gL=e?>oJG&YGpnRPgRN=uD zG=y@JlcxA9VRID(cH7PlMnh>P0WT&Vm2HAZuih3Or-cLdY)esFUq_T!l}kHeuJT9p>KI=cNk{;n9BQuK4g>qefv7Hxj1HZza&+ zWJtK_c`=iev5U!D!how{-w0D*+<8StAq*?;d_TPM6IVXa!S9UgA+Tkq+tpL~C|nMX zRROTS6wa$heKl_{JOcSuqHURs8ey2OT|%V2{A$^Y@b8y|FTTQGa)0^??{)r5r3}tH zKD$ek7iU_V^sNf(EBvoAX?=xD;@Vnm)VxA+QY`G}sn7fa+E zRn+H92rhqK+~HbimK(t`KB^lVS;OMXE@)!mrghTI9j?6h%*$--3BZB_^Bq|HrBxdt zxYaSNqczm}P86PKR`$HqORsBG^Q89g-9l1cG>#1=KceK=ORkAxo(UGoIb?ot)71pv} za=q1Wjf+k}vn?y&vlI-U_7a%@4V#q7RJ_+ZofCqV>wek?o z`^$_qnoZ!wOJA$@1khIlgS+#Pi9-|nmHCp`Y@HvB2Y(EACvA!4&@;_Kp_BJ0%*GU!nyzDEDYLDmDO z$cV!r<4=m6yXA^Uw_dItBvrA=S4)_Pp4jWP9}+mY{X|Hlom8$hWUfjBG$)Gm5XIce3M`NH!t_QcQUOqO zOGuCnQO$3x;tGh>v*|!^cRRq7-#RdIaZ8vE+kr%XsE9?wRjk09yK8eE>1zo0gris{ zQoM7ghoccMm+Vc7N_}J>#h^8o$ae?Y;op zD9*9-Aa5yn10>F(r4S{LP^~6`6)Gok(=Gx$=KSiP9~%Z(GJdwFmPJg(T(f-Av%PoA z4pHCgt2;j_5;16UvJ^zS#KY830K8elp)ab(S1j5lD@I4J-hbk4hb10^Ds1>P}}Y|h6Yz+Urs|4JQB7nBVg9dvo1I}H5n6Q#5YLonu=;H z{=29dBl{gT#wffr@M5`rTYf}4b&MlQT% zmu5R2DEvskjkUs-7Wl3NzD6^Igzm#h^u}of2b!qrd;hVOhx(eaA_;#o2T4T1$Rp#r zBfFv~#D=RF33leLTn!Pi6YdGe9huN5XQ+zJlU&hJp6t`aO=M$gOEKyaIX5L2_11Ey z!(bzFyZcEvpSCbYzzKa2PMqCJlzI}lZM)0kgxODTBjf5WT`{~1v+yAVN>;-|!>;J` ziav8|XLV|mW(h07o5f=>iXI1P67kGZ=rcVB>5WSt z@ZVahzgt=1>B>Rc@~Ic*tO!#$@ZVe)e=WSDN7Pf{ob*-p?oEoAYBX;9Jz5G`3^~+nDJpwODTX2twz)^kZg5$4m|)orig)Rvr4WVE>xJr0A=9Ee zXW2zGiQ;vr1D6dPE=(O9#Bw+kRb+)FXImb?m2dsfavTg=>6QX_O^i`3MHUq>TXv(9 zy1TQT2vSYLe#=ns=1m+plfVxqMI*t8hG+SE;)}UvFA^Tgx8%qJsgo+@Jk=M-*5n$- z!d2$YJ-6(DTo_`^tR{$L{>M48Kvq#-8u*{2f&@EcyRI2ixjTE+1llddZCv ziepo+BjeKKiZtsBpMSj)=2$*spg`6oj(#^ep3!b&NMyV4;4ayiqpyRXq9#|H{e>?4 z&sQopL+I2`>Wf49US1&lvl6zf@Q}%|A!DOA$ll*O=FhR%?*z~+-h3w(SmZEh6KV^L zVaC5r6h1QqHl^jw=#%Csw z#3S`m`QS>l9x(B)T;=QUkQ0&0an5g)*$V1c_Z|bP>w|SrkNo@F7T>l z62MY#%aA-(NjE6c%8f^Cr3z57dx?be&T{Uy34y?Yz8w-0d!wB%+>9leFdsbRR0pB7 z6*~Y+2rMkfA2;S@00Q^Q`+On9bHI}C+k6RFE%@3G^966M4*rhes?@3|WjC%Wx1gt< zs|}lZJ9+_Y7kqO~RZY|uIYZp4iuzy?x*?3X*0hC(oiOkYx3cmO>b+R4SHi~>b8o|t zmY*dyn=lpbMUYBM@a+33c(Y`eEe|xDk^SKs08;1Yca4AA+!TpPT|IZQE~*)iPoHFC z8TQIxBuHBera60o_@6i2_%4j#;4P=bXbI(T*-DL9Rtmm( z(Kf-07C`WqYmJyehInX$Na!Nxbz5R<+~0nVfhg)zLKRub*$#py))a2$mo~sw!476; zC4kujT-9D8Kw&YtYj=i-du>0IC~_zh8_<-Po$SP>yH+uRC+vZhLJ-Y%X9e@oxTG&a z22<3X&@CdP5K3ecKBDi1{;K1kAd^?oC&~rz<0IoD`at?k(T-KS!?aQ_UKnqhK!Al3c1XU}x#366?68D6oAR;)t%*LBW_ZCD|7f>46ht7uS zQVz2iVyQi@!RTzVqp^54GTFH`5ZO-l9@ExA6wp#5K=ijE1;D_e;k@(B5$zarFU=dG z3~?^n@r7URgu;?MMVx4ca_q7RAm!rw|1*T7|~sZJ56xA?^pMuIDcbYsz@z4j%2 zs%_Gv#}d(?MC5*!82lnMokBnxLiO&p@!=0kx3gp+I`{}uNBY1ha+F{{fPT=Y*p~huGr`$Sd%E!ILJ>e)gGZ?C=eG#0T z5F!%ro&D)cnMjJ2CEpWaDQ{g_gPxX0zS)x+um;h)m)rocDPKzl#2SwmTQZTHF-9CE zKi@&2jZ2L~t;{?@%r3!U*G-|?GN`bIjHg=Xwv?F4a-{Vy3q4b!EW6wn!bJt`qDmxN zRp;Tye^e#M%*i|`aSk-qjjzAWcWWNMLq2<`IJq7pI?;_u0~sEx9E0PYwiKN608 zKdg*vxQ?kFa+CL}dEt}zkktb^x-BtQj#kZd zm(-Irs_D1>k{Q@F%Bk%zY=ktFkzu!j2I<?NN}{jc}f8@tUej`aGM<2Cg#tQe$+SAfg#I-q$f7vL#~G zlIp!gSz`G8Swhiwi5YPxlp6ImdS``4PrXSt2|O9b14ZlDKX{wa!AB{!xmdVi*lDU$UFNuMDt$EmDMOlM08!wtWK5z3iC!MM>d(}*@A{f`EM3K*;4+UFd!(F_7c1LS@3LH5!=iDtJq2+WEpIMQW zEvMjMSs_+>0@!*l5wq*p26HKqms-6r-AhD*>q6TTk~%ZleiVJYVSRY5s>q_psNE3C zS@xy0e-P%mf@SGocx?z9=djjqH4=wSl!b-WN9}L~1dQ5URi^ayjUwV-@SaVYaZx*n zWs{4Y(?HbpXqr+&@n%96gCV>zqvAaV8k)QIRWds7PvbdxFHhQqQJcRN2tER*OIxA} zzg`asBKhzSFjXaR)P25Il?Vc~miQ26naFr%-3SYCPqGOcR*HPbbIT-D#nn0V6P{JA zd7FTai&>9hj*}LT#~;i+mdag&`5@ITA+42W+*(C22|2VR!-^>Pj<8dv#K1L&%;sx} zfhU||CjpTTfU~#~%Hf?$RLe@txRKjQfRQ!3kWG*!l;WE3@E{6UCH~=!2a)!%FUxEs zRH8#QJC?W^ws@lG969TEI(%{PDVhh zUt^;gA1O{!YohMk1r_lS1P<)R?g(zZ6YdGm*8d^>0Q&ForyFImKer-J_VFFZINGy) zC)~2^hW2~U9wWx_eF!|DNYcWN6Fs_(<>f>@ljEJC^8{?D&0h&UTTLn-jJzeDv`-E) zrbMihGf~zrC9=Wb*ggn698CU#f z7edHDY?r;uCzlxqx247@IZ+Jjq35cs9h|3cJfhl#bRy((BwE6!XKODJ+pVuX+qkIm z{?xPh5SLX@Ji?3DP-2-}xQ^Uea#Ljpxj1yo!Zw7(H6CB~4ve@x;a6`GA^_&i4&u0> zu?dglU6@1snS`TAmRVC>K3r69BAJ)pHnKm%QuUrC@Ljs0Uam4bRurPam62}_3Rl%( zE&2Jbo{oNgf0SIdzZMq8MX4=*XseC@j>(X9z4ojhWX{DY^HBt{WdlQ+D6KaIG9^H! zx|S%P8xPu+P{dz#ZX*nf%ojx>70XHn)HxhV%v{|@e`Td&8-+OA3FR=IF{-?=((s+> z)w0q$D45HAD2lm5QSLTQ6 zSxWzyJdf}pOeOFs-==wp9TTi@nCCcmmVTI-A<0_orQ8NVv{()U&JxNAMikh1ZKrH~ z!i61nkYb75daY5kp9VCB>7Xi$-7h^mg)R~2PF>m);ZE!fQv4U6YgOT@nUubFjSw4#gxW?EBK3scFA2*)F3u zrT0#_CmhAPD2ISPd=u-K8Y6t)*`GRU>l*fE7L#X4A}$(U@RK`Pt4V}j|0FA;FFHt3IBlF88fuBs8>i!jo}PM`eaa9r2ARuLy6>-5TR# zeOdDi!sX6EevdYxx(*PZt*Xl2h~eX{0~#;xAzDIWhZ76@#$ylC56d`?p6eVQSVZ4? zu4j?-As>1UXw_WTVHxt*10I@g&B1L?P9+DznDK?U6Q(W`D0tJeAS4#wylE%kHO2S* zwVfa&W?37jb`~PMq+;2{^c@?pOYakY%DKbpl2c;g>h{`Bbq4o`<`CmDh2Lln&fu{f zn}drh5wrJBB(ADVreN5O+#aM9ZtaX<2ym5Z2u1kFc{$$N8KXj=*d_GB_DD+y{QZ*f z19tvzRbRDVkOV^(03zE$ywo&P^!@)J9sjKDFKWsNP^`6QR-0D&=M zxpV)CoqulLf7;D#FLqBjZe}3Y%RKqAnLUPxPxf&}c^e`o-sWV7y`E9tFH29p<&5$+ zL^LJ84H3VU{1_tIk{?6FQS!4@0{)$f{m$Ff(k1h13UmCv^p$#8OgZjat$GQ)6MBb+ z=06BU51IR;hT9E>6BcsBfx?s+MO>66VJ|WH7o?yeB%nrc{Z82P>uk3TgQ6U(dj4e` zfLc$go|}P?9SeAdLvv`o3-{(UkppEV99ySXM5;pqaqJEWK;pROZO;UiG5*>!z!_si zX9;;CbnZI3fyl8095;k&*SsX`gsEU+8EqTntRnN!&KDf4ON-G~RZ0~dEMBXssa8Nv zO^Gc6s&FV#l=#wrD6xyi*VhuqWxZ6Dmjpw|tMt13evJf~Z#{mUfsu1p%Jnb+=W! zQawxdA$#{Ez!v)MMu>9J+ixA&35J%-$hWLyjOGIB=f)A?8=6kEGfQ5re;ECpP#Pb~ zlJJ`TtBjPk*4frYMsk)-HbI&Q9->Xiz~i-G6AE{i<6C6mGBc4b~ZH;kzaA+d==a<{d?QBQt~nFs+?Kj+=ou?Q7^ zf|?Fv!44pbQng&kkqrdxYje7&JU&F*L9&rHBH0*rMY|&K%2#$pCyF(;lVpnEQONW+ z>~hE2Q|TY<%09E~0NAZIOZfP9U)nhz4Yqd99bhT(^U>fb@w7D*!ZG#3JMJ*y(yalH zePqBp`?x0Z>ghC#0fe+Ix+(~`D8(C{cXABX1N8@?p8{Ok>yX1cA)r9mgWhVCRueJA z5<-@IxDsZW>V&2qghM6e?2GvWw+A8IUrGrmm2w^#*OyYg?~j(Nf|(C*5|@4`1V#;q z=9K>O-@7$O6s@2y;}79T=o~MbP3YVZrDUtZvI~o>OI2u(Eg}s!Lc#?r7R;59$hpf{ zjuq;#8!72)m&*|&z0SC%%RxrS`OxK(i zAaVICC%n-FJ<0Bb(lP~@yAdXwx?Hktz@tlL+MdAH6HW;L4>@;eb>TSK?yatwDIP4L z%GhWmxW(FmShwMt&hAB5F?AvJ!c= z2nQZzLZi#rszkx+{yuwT%FEnp)e@G3Wk$Rb;@zV+L^~@L3s&cBLL(+|^fsYKk<80{ zn@4?F{@8S1c)+>JrD{u*Q+w4|hc*F7sbx)ASz;0%k+CrMsf0bu{h0I?GhvFEE~{OB z`2oaC)Kr^&SQ(QFu(h5_-l@;`ab@{@|Lgs|7SJPSX+XA?QJmb|g(0k7aZuY!|K zM0R`Vc>GUpNgoNou&#uo#H0}abt|zCLQ%IIA@*7eE;k9eKcXRDFLZDgE{42l%vXCqm$|MfeW zD2@OAUG&Qs{_VfL|F?gaKmW(;t0werQbvRUF${7n|TbbEV$|I2(A^*okWuoVzm)l(uguwD<>a2sPe-WRx|)T*3&9lxQ{0D6&KjUv zw?X(G3A}M+B97We{=%KD^EN-%l6pD*`Ms*8VKn75An5ew?Bn|wk>~p#>(GrsG3z=d zk^OkdyGI5*dAWqsh&1cO|Hm=Uy1MS;zpSpGGmX{NWJUkCtf_zRvb3T$#+Q$IIM(^+ z($ZY((vlCyrL}%PEUi=Wva~*`8TB5P*5~)XVWSy+7$fw(QSV%LskAMwCGXFR>!*y* zfPKreG~>R3>=SK@SmIL320gO^Ygk*DVu_IBa;~j2D;`^N;&|qh$*^d>3n276^;u)T z{r!)fWpMAg&s=y0qyYmG!XykSQgG2BZtTb}GNa4h_%LpBoH>7z4U69s-po&gYT;Be zrG@zqJnL)j3;6XZ^Narpem%aw^FQF%9$kn;O!oPl$r9nNzE5h<8 z1=hJ(_#28uAgdk?g{Y*UPKKHA*DGNS#QN$DkF~L!zn(Hpj;G(JlXMUzw7Z^vJi%3) zeVGw;_=!HLFAl+0^}Y}-b5z3RcYDY>*ThTmL#9!*2vl$;lUW^u2K(9QJcw!tKSoua+2VvqKYAN{Di!aRR7682GX|09ho#QJb@nb+qe`kCqmjXLDb;2Bl zkAw!?e7EctNBH7;ZD);O23AYRTe&0Z!iBJ5l~oUfgve0*{u^Nj!K>H~B1B^%w9mH= zl9eYjEFC%s)VmP6ohLYiYEq#J*RR~wEVlDc-VF+VyF2pYFP zmc4`UdhW5uu3#Hnd9+Y%#Tri;6lsCS0>(6~&1>3VNm!5Qim?(>tC z8f#o<^OO;xzT%$mRrOSem$Fq=#-ef}30klt2dtBbo$JW4{xcZhhEPPSbacEERzM=N zSz?jbHIVVYtuY{{xsN+6GLa#Y;O*!~9KLe0qyJQ%f?D)^uRJC2Li(xlCEDXw`Ic}L zP*_()-+#1$YcDsNCHoX3S>!5u$dfR{NG(C*Og|OicvR=cO*C-n3tM6Y1#c~JhzU`- z+fw4FDvN5hGPhS%8GK(XVa+eFhzp_USFyVg455r5n9=S^m>Lgh)e>eWYD36Z!it-J zPL{B$!$kcgd`0zoEjS|5^vTgP!$7$N*_8y|d)jS=1-$?*ZDd^Zpw9WS*`5apj{Evs zb3{yKy_?_1k|3xFJC?mfJTZER+mJk$pt_8C*le~$IS|RYH`|0IgyCmZd%|&f5aJ|7 z|KoFZ16U~+_YU|+{?|cBaoRGzBQxQdlp05?w*!R;D$vLch3M;2>nQ8Sg4t=`idG=|e~&Nr4mqB2 znXo~!x%8Ims#oE{04RBWS!SGb<3S+0{yVeBP4-o*&HQQzE`Ol|J3PL>^LN%C8u#=4 zkB7n&ri3e-y(nNOvT8a{!e37UxVUXr@~3S5-TOf(?7B<<;=ROHM2M9`RkDg2R*N=4 zR-{I3-xBcaCjyO407{s*{hezOgR=c2Th3g|IRNlNZwU!42TZ|@FoPJGl4Ye3d#-q! z(3mtV|2;ugnFG{wPly^ur9_(`^I7)VVz@?+ew}X5T;mKTv&-@@=ks3To-jqwGQUKD z$il0FTm<6&JhQii+UB@oeTy6JliyY{3PQw-+)M?Vh^&IL$`!oZCTrmki!=5N`{pHfS_~5CJ;C=R9jWx z0T{c+TUDv`$fb`V%pUx#>&O9qA3k^-gk3g#cO4+sk(?`RKZnK+56b19!0UV_a>HJg zC}1Sss)|^o@p8YHC`UFq%qD+Gcze&7gaPBX@n8~$Kw4!%kTK{Vkhkxr#HkZWc7Riw z#_oG%AFDmN7dZAP24DQqO#FKLl02mVb_?Eo7uBn01s&cK@J;c~GJxwkH=i2^Q*Z>Z zbK`*!?4RDQtaHC5v(s5ZJfZ^g_EMw9f;_cBL?iK8Y(nGCnP4`1Rd`mn@=;Y!?{l!O zRuLx}f@mk~s3t(@o*M@)6M=vXE3zNUX1)-P2@PyjZt>P?T=Q}(2f_#dp~fIQvy!lb zb_?O(F9}!Y>EFuzI8UGJ{Fh1@az=zqT3YtSu!AUE$rwsXHaPv4sNpR?FT;J;*x6@|H?-rlB=%t_H;x+SblV?|%5Py^g2+7%ao z_tEGWTLKt?bec`@F}WNq^C0;NuUF=YEvPz_DKSK%&K?rLF06;{AI|ZY<}UFJH^M#P zDDWdV(-jDmvO{vA+0IzKx+8E6!<2mIdmprPPXJ%R z&ypX_P!y69CntY0LVevCf@1mXoN~7vFdH!mm^{8B+ieN5MVGKP`VprhEJfW-8xMYp ze7sEvcD0vSnSyJ=Q({$r+IDSKBMw1#maxeg-L-Hd6vc_OZl?1xw(Jq_K#OE`_?paUmjV#XH}!#Pd78$y4%sKj{RD z(ZFc$??WA|xU?lhfVo=44%1M`Aqi3~ zK6yK7YXgQ1k~OChTy~1s6TlLd0ub054-{DQg>ax?Kjn_yBmgxJ(9NEJcVN-r1zQ~f z&l?Q`yi>@Ts&ON2DDcSw%E(Z}$b)dU6kM+MRxCvZ5=etnZWNNem1SbjgIrFfuL%!D zsd%b7OCdrvhf`4vhe~g|Jv6=aw%yJ>x@bVoQuHqKtlIHVA6&oA&A(nNafy5PUb#eH zh;NgqQ-;H$FB-}9$u^QEM z^pXgZ;vX`REtT!n@PI6FO+!|E76{mUtQv-8me^eygrYWio>UEC0u$7gD=TW2S`RfA zd^FgLE`)%XYjA~~u<ats(n@9TGgzT(}QOAgWerc2#PnbH`B2Tf# zlT0^x`p;K(S&;-u9ofgCbFNfCQfE0pla{esJ=f+e^M`OG@G2FTgkMhrx3d}uh2AM= zM=#AN40etz?~JocI5Y>{(|2}bQH1C#yIGBZq9x1;6svwEELdDuZwW9q#V`_|>1i?Vc@vzQoU;9YmFqSro8H?HJxi)bj8Fv!tXFLW-1^5Zoy^fs1DAfBzn zJ>e)(Wbsy+_1`3l{HD9qJNr;!nrza2Ek@>PUQ`(3Lo)^&ou_d=xQ);)G68sMZwVtu z47A-#jUtb#hH`xfM}m|X7gk4!l{nz_v&2&8ng{CGol0zt!1viEV19;K?4Ce$nXLNR z#)H(@)%5nGRdrdfiIw_YQ!H zJhDGsCQGw9ptX>Xsa>u;Bp()M;4HlJExUfe?%DNO^go7ZARv*O34;O@Ese7D$qeD- znmSW|;Rd7*(f#^~ox;^9Zi{=ZmGGivsr9c{*DOJR-d*YOSCk|At0jjV&oJaU48RwO z8h;pum`gOH_GI4=$TrLK{W&0;@9zWd`2Oqs{d_<^(7$KQb3h*7|G0Aa$DC!aJE*-Z zT(tm9mJ?Z4??VTFl?S3-+5ymgh_N%h4I)~DV&>_I(vyF z6MoxmB~Df<$FX{7L*<>T?W{~fVsF)UXagNd%+y;OxNjg*tZ4(lsS2UDC5T+XcGFhH z_T4aazVJxQ)RMqx*a;OyvF96)7zBpsqnUn+Jse+ip@iIzNKmhgx%<%l$T+7{zFwueg$z=7@G16*$-s@?g;J9oPw8nW`nkHjl`Rml|ci!GX}$ zDAENG`)iFO!16u`Mu26L@B~=43GV>QCSiu}*%L;To|Wrt6Xu?^2ulr6-LAKZ|9oZF z)2~Kld1W8##&Ns@QR*oN{@>=oq>EqxUUU+UgjRb#B{1*&QsPXIm8)T`b{-;8y4=P? zSmaMCxFOjRrGM1y(V~@mXOU{<4|5|%*3h+2SORO@QRl5 zS>q$PFqR+$HYVhvWn3LcipRkTfLyofC$ zH`SNQ9ZA1-kZ_2DgC*?99gUcBA;hGn+MZ=Q6ivk)5-?a@SebH%2r0E09tc~c$P*VH zcw)#c9S9R<`|VChu$-KeWhE=koXIZ9US(7h@}L|w*e8%vR?L>^u`_z__QTnn6| zxQ!R>^eV!t9MHY?+biQD!sL*UbxYj2Y8gZqckKWfi1IPZhluN7Ym^%yC-0ipe)#ny zh>&{)HTtqlNW^kr9@GGi2`p4K`?xwK%WUqp<~>E zuyg%duY{EVTmH}nMzDrp61alzacRRm(%T6Bl(pPlWtyN>O6P-}e(Xr!(m>Aw04|~FKg%)0o z{m<86fUn@R!CCyBe=arA;1m)5Dagwb%Jaz0VO^t9?u5aWnCecrtkI&8&n=;7S#PxB zbB)|fwk3TgWDwcIA)#~s*=+*vYZ{7ePY{hOY5bTyL4Clzvs}yGR#RzLu4QOt%RLY( ze3Rm%pZlp}uhtICcf44$WuB{vu3~koD&T+}7mAI?fC)~F2f_%nS-%rvESwc-{7pD6 zJBr#*kRWm^T`n$Fpa8TGM#lBBgV+30Y~ga0EQcYht02zE(3OoYfyv%gOS zy<(}=n=Y4sR-Pxmb~!$c+!i0YJR-R3VV2>-ll!uMx1Iy|GHU3Eo$!G*xix2kd2fV! z!cl@&!Z(EL4fZ2JJ;!7A%6I;ez=a6Ddm~{8x&Rh)SZZ(tkDOvlYSWCoPPBua0Y1JdkURA?Z1C9BksxcwFp63-BJn(#bz0H)ehj3_02oac9jj+Y;3z zYU$Fw#F)84v|SEjx__aFuY}7%N>+6UvsZ>ZjJ3!QQ&L|N_bI#~nzpWd5N-u-}X=y-*yc>Q@QtfI;V8(sHqPnW z?Xnk9n)Q15oqsuX0jkrMf{53<+#=&#u%yp~EYgMJ+Md8Ar{Y=fRYg%98{5ZPRhw74 z$>S^GZHYy%KZ`9Ni$`BLaM|gzBRbe7ms{WYVe=*3L?h9|Z;&3C|h}C>LCpHR*7XR@XI#%m5GI$+(CN zumxQ8mR-@?h%0ezXCf>>ik4kao$4u}WM0VaAZb+J(yxO=eCtvWjXR-a1K{S$V+Zl9 z9(t}vJuH6jIoTb0%BP;gFd}Mbq>U@oZ48ob&56VuuI$P?Ax?9-=Ua2;yVcgSQ-FMw zm~FyQvWy4?k7OYXj!eYGAv8}@ixufy1rFiJ+JXrGk4NBC{BY|=q0draYPdrJwI^>A zz+4Qa91>b-8mDGOOaEy2Xq>2HGMF#umQ#+WW6J2535NTFdqB3?ck2h6<@x?tKbY@t z>xJ?C*ZKR|`r$zT9<0auVS!h_arnm^_}81oUX~BMDd+qN^Zr(u0f$=pk9XucRq9Hx zxuwLvUlM+BivEq<(<%B?=f70S!@20v*!>^+@09edichEL{|S?yPSKjNINbb3#>FX` z+Zc%|&ZLwKS#IB~#nNa3T0L6zhqnTS6?1{`eBoLM=yF)@1 z3GS0kKnk%&OoE2Rbl!5^0hcO$<9euAFHFWY9>QNwLJl0gZ_Q;TJRkdcE3m;Gn(fv<__rx>q_rmdvXy?MK zDBgbQWz(r{` zBlJ|>2|3h$qpv#nvH93noolXA4wd6OxJ2OCriWrw`aMDi3HZOu#4;rki|v^Xr{3rJ zo}sgZvK?~_zp`Q?d+a;)+&)KGf}fG|yCPg0IQ$=S{`JZkTm(UAf$kSQLNt>0wn8EY zFnigL91p3%_@AVI%=ibgazyCFm;ro51<~(S^i&N*JEMQ_y%C_PG?Tw&wAF6E*YXOU`dQ8oQAF0x{}$YVRn!Q{ zRb~{CA8s4UpI{slK>CR%@?O&fQ6=u(Q;2NNebN6MySr?Y8wUs$J3jfQd=;&!`VgaGD2m#OZDNHP8I+q0QlfShEquh ze&<5;8aaTsY>JEErnPV@E}>5r)3xY?MB2NNNf&0IlXo(dV$r>BWTMT++|l67e2chA z#Pf^Q)R+k*Et>7lVFsl2@V!Kwah?61>6m1`%O<3d&3nC7&C0XMFPjq+(SDa%FPgIX zt^uucHsw=%PA0zw@AQZ7yl2UfnC}t4s%3iY)AE^n%okzgi3GK$Ka;kONeh6f?2SL( z^AwYNzveMH#%IC{%6M;|8CISgpU=OJpU+o3mioO`I2?(L7@z;R@%G19i^Tj(rl5wJ zEAoE(%3&$&e|^VCy$7xciku!#?iv9nV3E zbDU+*f(17&CvrQtC*K@N4H=J{=FU(fdR6u}hFS;Q)ETm_(Jk|Zj8haLS@zzPQ((vF z%Y64-8`$0tJdcLx&3CNtot>B*UFqMQU5rq?94~I{rUan2nW5N^I3k){$Yj1$o)C7j z+`$ID&5ck1jGM{hdvzTHZ2KMi^koV@-*R87O~ix}zdrL4qsZqRm%KTOc+d5@uj+r|#1(*ffAY#f9*>BpER((j!nCPh zzk_L;mhZpL-`}8m#qaMvmaG`6J4do3@vaqbEs=9_XhlO^ya|6j3CSs9SI~oQ3g_*RjB*ex z8n@$>@!1YLn;VGa#TL%yTr@h!m%>FlqRB;##bF2w{zqk>@O=_?ITU2O zP>9kgli|8j2vb(_Va^9p&_yk9H^L}lzLr?bI8DK|m1q*C2Ktkfc}GTi3z>H%*^5?6 zK&wYs{W^aaWr3)gWyBQ$tYeA^ua_gUl(#5sa>R%P`Fy@V`Mb;`f{ShWK^Fu&Y|oDh zC=xpx0zyG%5UvabfWR&+RnLWi87q&MBk3)EOv0a!NPara48Z7QX{Ma{>%*7Hp)V8H z#2=|%uYzI5F~^@9FrJ@6^9FCT&kaFRbCS%>k5p|N-(_Jb?)5_zWw(s^A(JAbLU2B0 zzuw<*c?en@&~fK5X5`abhdlDVhlP}mJLppK`@1N&pe$9xsn_TQ0I8i>^=$IQA~l3q zZ!2Q?2Vok71dj_#`XCH){=)6^R--IybvHL{z}%oOgd-ulh%x3NAyjm=LqaJW18ssV zos!+5pV<33-rTt6#e`_NhOh-5`TLbn&UUgayCv+{V=q0w5n^wQv56rBZVf;DE1?D| z?$HuTk@~uJ8PcGR@M}N+dJ+;i<^Ai#4IP{8$)(@yuy+x&~pMOR`6$uZpsc20ZXwZ5~49g+hU^CImKST@L0gYO9xge z;n-@v67u4lqBE=n^|-9Xl!$83RCu3D3=uHBbRdjHQ4a2epw)76+zAugJ20b@5U7u3 zO}1A&qP-2daVHvV0$W@0hgR#$=2F$*0jGNB_aR*fyC)v5Em3q@UPlfIOp(JRz~)I` z8uoj5H3c^&v%&&JQk1d|9F2-j*!|mzbyZ>et+lhgtN`w zHjZO)kxfSQ! zVzZaR5>&Ntblu)@Qm?6RK~)clU-e+OSEMF>et)MIy91x&`Tfs_gVmCd%X>~lNYJP_ z&-b^Oy1yjGC=V=TV1pKjwtU2NPX?aI-EB!Y%k(@uYKm|CDJJM6oaq~BWNFjG3QLfs zQtFl`;qMn=_RG_rGOUP&GYT^)o}Mey{mzkpxQ1!b9b2}^mmORFT`&)LNc5(@B$V$M zL_Fdhg-El$A)X-fG}A9^zs}!(*IIS9KCXci!%0%>F^s}|zH|BO^}Q>!QkdW0>FG7! zcmTdJ;JOWvu;t)3ngEuDgO`%}cy#9g5#IBJ{~`Q2Nu7}5#Wa@et5Q;5!nL*D0f2(K zmk>DPP*?@uu~kbDD&42xN(iz_<+J~eCgha7=8TSDmTm4KT@*zPr*x_aBq;Btxuw=*ttnK*c5X`E?U$an%!VJC4OoQF7{i>N?+K^VhegRd0ULemR>)VxyiE%>9e7JZO#tx3bd8>?m)P->uZdLyK1p zKkFeo`WzR+)}`F{MSO$|V-aavx;8xzuU*UE+g^5uxhqtptxgfcLE&9Y;oa*)+?&)j z5i9k;SP3r^@7>liFhu)I(Kp@=uLDB}xL|7#0z=@{je-_`$`F7=B+%0$1N-XLbdy#; z_~CIU{nURuoN7=S(J+1Q$6cX`fwUR^|6d17U8 zL2?xZ^y9qHCxNFkCj#%=dp#7Di8$)-MHLX9s%_^WTN_zH-K~Q{m7!eMb{;03^JVRz zsb@gN>vj+k+l$z7R~dT)3)&83-pcag!pug|D@EUje4CC@cHtf0OgkT^*4V@~grho< z?vt(MkD<#7=iS9 za()a%$LVZb1jjs{XHBx6Ys`s_O~ww6fE(L9Y2i)v3q|>}B5XZ*TK%;qY&}8VI8+5g z&9~-E>k%Twb#J8=D@*p}3b=f%C2;I$0n11XG`0OM^~|+3J%imvsccqQ`3<4Tl&uFD z8?z*Q7A33BR7ZuNlF=}V(I&O-P}olz|Js`#eoRtaJ_xPIt-Q3nZZlWU1d zUvM8pP247zy*X!z84XyIX*J@TyXNii4TMAI{r)Y-3izF~9Tb|rt*kc3#>3RyKXd#r zo1#~Y)}4`S2jAUI6qHj<*;X#%!dKV1`KU8OeQ??QZL$ChTtE*AR~CtN^9l z59j5d))6@D>|)=^htdH`QDKhX-`|&6B5KJl&F}9D^QJH}An~IK;d-hpqc$sC;yc}} zZes~b-jpcjcF+h?n}o6}1ZAlkQE=^*&2juH3(yrs!%7%ousLH_=N~i%mZ7!(*9nH7hQpeD6Hj>l?Ob)JNPeQ~${=a&DF% zHkl8>rR+_9VwB7BTB;?0P?t{xg73ry3U&}=znb(sc9YFZgK_t>36}<=Vv?t18M%E( zE9&f4s|*;ZkDp+mN&pNr)StmXH6ihdUhTPpfhq~0vjw_yX5MgIr%Dl9k z6(V`Bas}Bn-8&Lsbv?!18CxVWdAu^pPPKcV4P$UGm+@Qg$gC4iXNL}mHx2*;Ljaj7 zB|CHgk4s?((p7@=Lyh|g8N!Fcfr2lVkysai)Z7@G0MLDj5}eN1FOawh^qJ9 z4G>zjf-e6F*(%?`wMrZhZQmQAP1y~tNUkDN!@D@wrPx|~A1zD}9t~*qu)P~tc0d)6 zV7`}>&d%(lOO4k8Bi(G?o~1Ui8-WO0qf_TB5SWq-3d85Q+R+dL7e>91Yr#|nkPOiv zDr4co{Hj9qMM`la;i5t!@w+*l+@-?4*F*G+RARTc&MU^X~%uWuxETdSS1kAlL%I2rwNNN}{E4&^!Ul?B?{8rSUTMBm- zm7YuLOtH0E$O%}WgD9E0ySpgFXwk6KJELf#sE6VgM(#B=aM>b!W!oMUa=*iA>^okY0D1EJ!?Xa_?eiXK7e*G3XevFKT`{gNGi z0nkmNRa{09wvVovhLLA=AU)j}BSDH*$lReWlPojwP=``^{jiU+$tcPhuVBk4y(~rf zt0)BY%?lj*(}_es4sRKM>_mMxonJEVo!G@}k}ZJEoP(&GzE6q(7#+b_ZRWV4w>Nh2 zMnL|^p7`yV} zh_SnVKVa;h$P32qM=_(^1IF&>_kZhkn-2rR^?Rk>seXV!{zTbb(*BIH`%^-Lvipw{ z|JN%q&l_GR$P?tYs2dBf^7hXhJ;X4fb>EqEp3p3rygz#^URO*mGLZ@!NRwXn3m`L? z^fIEm&&8y#-G#Cs@xJ>J-^dOe!+;1bGEm!OT;e|NP`zsp{sh%a+E7PY1Hk*Jl-n!mz*_$QnbiHGG-4XPUWwlmfnA<>d$*+h^V-Azev%Ks zAxRZEUJxT@ZR0;p0)_W^61?*tli_0ssf+189C`cN-7*~C&_4?v(JA@InA@pldVDG5 ze<;(tlCly53Nw>OAdypiWH$8Wty9&2>(!2<%pS6BzNHG zNgcBc2iIV2kxcOd|UHd@5%_c5V{(1=NXQLIyzwL4pMnZM8 zX(Qot8&d(et4?tS_nLy-&a3%doWOBnS}X9^LjY*l+Y0C_pr}wypA|5KpB|c3E<$Bx zOIB7{zr?1D2U&3<;-!t@kh(>D(hWyffJ-}80$UHRsCSM?Ai1y(yFVvMj8XY!MvEaB zIGMk0IT%H;Pof7+=9K78m}oL_na_=2r!{{$9jLq@p=I95$4R9Z3ywmbY!WX~#*qiH z+lk&e*v2J|m>q}jTpXg2O*v?U=9#S@1&!U7%tGm+Q;so-U0J#%q~O`YO~S81=tLBR zP^$*YzU(to4xKlFPr_vrD}!2t6`4{ofWQ(q6=Zy;`cc8wLtl;%TIQ*h$x$>-mCcwX zh~Wf=_JP2qQS*j?gJ=a!b{8(V&J}6Cb#uVGjc3?~kw@-rY5kp%DDEGsV?z5P0feNMPiL@cB^g^VqNHSLn`V! zB-;+SoF@D~_8iBZig!=X{5V+gJjLEP3|8586^-y=3ResfxA)%YgsAbX52KuCF0$sF z2;7bzhW`kNycSja6n|*5opC3uN1#K=b)o{bdQfr$kC)`QieYuUlRu42%;T7)*BB-* z*yUpz6ZTx!*>}3-UZwN0ILw^K;ii}&1u(}UArBJU)Q2+j8e?;3^bx#*wlUVJ59Npa zGZ}|I1l7qqOW-M_>{LzdJ^OMThQ3uFI9VP)9O@6ymE&cusRvJPz>K~%KyKNqnXohR zPT7ZWZ2+mgS4(AQtcaQYwL~7qr-eQVT*AkAAbVv zI6BQAgcBQ__*UKKn0OcA%B`IOea0&w!-(Z6n&B!6aS~kfX1s(Vv&HrNb-WP0gLBuz zc)>I*K&hE}c2^UR!_sOYw4vOkv1 zn6u(!UOV|#+cN6QImB`_z33{IqXBjyvux+U zyEot*v2z5^t=`(%zb*t0j6MeBduzr$jXu{I2@_8#Wr97>Pysst+l=-DE;dwW8Ej|VUz{TGF*C$I8hea77x9Xq!7h$ zH3Say>wzO@A`UDAb`}qKQ0mJd|NEsmh87yFDg#2aZhT_HnFI$V#gm-2U#a|cPV}(% zNfDhbJFhFEnO@*Dn|$$AY=E0)!fia%=X)!%X>F71RqqCXb|Pxq$c^Lb$KkB0DQ9J` zYZwXkQ-P8{S4S+l_u=-=81*UEy~HPK{oa>TVSVJ-Enym5i!u^(ju!+!9@cdYThI02 zlYFR^()D@k$c97b$8Kfv(|uSD8y}ZV>W=lv*rsY|qV|28j1NKacU(UZ60S1kKp@aj za+VOKOf}ec0=TcZOBkthDpjW9wYnwBzCJ{oW93byI~3IjlHCtdx6+6UWo3|I-+2*v zX*tCLtGydWX|%5HeHY=unda?^WyIl|N8Yy*Wk?5IxVJh0QVA-^flwL`scccA#X?%f z3-@X+%&9D^o9v9Z6!bF?9k#JQn z=>@~Giz^cnNVafmQrRe0;$#?cqk`T9+j1P7@pSgq0FkGe{V;D7Ij$ZF9V9wC1*ax5 zgh%0g2-hv)sEK4n7pwhnqBgH{B9Q@t*9aHyi0f;jgry^oWtCk>A5HQ>pMGxsZv=#N3R0$`ophd~%&s|PANGN+KPgsq2 z!o{RYO_jYZJQJTYm#y%rZ~;lFabbL&yT67e5_f|D)5Wj8J92>6ovNVm)%Ko48*MidWVaMVCg zpRxd9%yGTtNT?3ZKz0t`lQclfCcZw`BG%S7`91#LK88aDK=+Z^;Egy~bD0+ad#Qmq zEUM?flp0(vDZiS)lv+|ME*?s4DY>3ZBLu=)xLPa++h<|adWByPv)33J0s|$YmPgk zn87@r+C*sz5MOkt?l~&x@|L6QvZkIP_S8uff*iIVL;)6=TROuC04%pEAB=uAGos^( zsHp>^*DZ|G$dHA^`$JK!$5$Y=i;U|0&M@2=yNpJ^%jgz0A`1yl?mMGQO5i~Al@X6+ zo=NO=5F>0_g?-MalOoLs8ut!V!#_m86{i#vsw($}Zp!y|! z9WPuPb{E=Vyi8@fS8yjBwXAd?56Hi1Sy`jYlKd$HpnSX(W;VEgj5chdt(?Y@{ABCi zrIh=|HtbT4$B9BGZ* zA-iNtlE#!ew!LZQ77w&AR}V!cuvV%&5Sr`6k2ghSS2RTWtyVz2>egpxEE+rh&e&uQ z;FLEQA1hS$wS{pUST*n(KZ?r3x(9f!026LE%do9t#tW5EA+4|;pbtCBKnF2gn3{rw5zLO7iB1!9#&9WVJ=t}cl*1D*n z(?7qXBZ!SR2}4JSI5?qBL3yS$XQMPm5k<1)sJlvhVYAIq^nV4VVbAeVE>pyjbOfa9W)sgjFmac=;#YtaQEA>z=sex}Z8K3Hu%ni5y zUY){uc*>{hlqO>3z1#F`U6$`fs;Rj|TA%8*kt^tXiRL zG83jb$JGHbAkWne3gfWQmmM_Mx!w|RyJ$=AO|>Weq2X1RqC%JkVoRNuA_qnsuky7^ zP*MFf|22Hb5A$ET`Q<0O*Fl;X+ypyjBhfD6VQ%6{p7VWfMqo_aLSHt*GM-L^M+2Lp zcEV9L6WXt)Yhdhculgi^rdahwuf1>^$KT;L!%*#o@%BzPCz$9qUR2MV$3E0|)qBtP z#UBHIiz&G5#OIX;!A}$cewUPIOA)xZZ=!m=UL*XTkyS<|4_}=TR<`Tlh6W?8XYY%% z%%`}M5uCVMqh(A%FEXz6l%C`==Y82z=J2!yN-+1sLI zZ{5<;`!AQTg|YIS^iXS=e7*H#2vIMnqveQ1AxX&!hx+KaP3*$O@J0{4G~nwTE!w`C zT4K*}CmhuWSI4b2WtZaXA|Rnw##sgkOgQW8Y&czX zdh&y9`i)5w5`1H@4TY2|OTQ&>4dvf8Tnj=^rJmePK!Uxw%52^eb7#C=@m`u)sY3>ayIHWIG;&fJq)QhVV=W9X(^!|`}lk2&`d0DPRR2JX95 zEfc9Eouq0UJ~rE%p7VW|!mW`tY|Fc*AB=zjw3_m*IvSTKCOZ%U+po0t&O9SSk9J!W zXmq9QMU8F?#Kb7CsrF+71@K9T4PYB6Bw%2&Z6NkjdMb8npbW%eskSA>(aGi4?x;fni?u=anQ(vc0RBLuoQj2zNAGMy^wQIb{j~ts@{cQsv z--lT%Mqp70C}(?D{s0YIY(0kt-^JYtN7YWuS6Obp z?RE+4up}R-6+|Mf;d7D${GN4iSttuYDu$3<-~{48;QR{xw+ZDj)9~>y93m^%43>lEK&4j8QMng8 z5(vU@BDCUMTgw<)SN6TUJNec-GU7b(5)a)&zyZnTPyuzZIEX{&-hxNN&cWswO>^Lo zZ3?aqhTsFB8s$LfuEl&PJmT}*(t*HRQ@?QRe(GZ;c*k=d|CalZ;xLOCN1`1LuLEtO zH+%AvQUSY-*CQuh2PS8e&)0!a8*AsJY6ayxdb*Ugkc`vR!v>6uZu%d#_xIMt1Y+&m zvICIQ#j<7RA2uT1$hOJy5g$^fXTHfnWP%bF`XB_1g5E8GnAnLu&|Z|7K&6{1PZ<>+ zpWVTpV+QgyEsgQ(AvB<169fW(osY^$$YXyVP&^6WGYJyt{)A{W;u4S%N7-A;^cHYP zH;h51y{mn1(oxUA9-P-dhSBo~uI`;UN(>rIKHEV-ctzQ<97G|6Vsdv;2<_tgL6ibI zY`cq6r4DJW!yt$*OeYx|&!S5)oya@)nw$IDi2;Zt+$L-%;^CL`(pL|G6INi0O$8%{ zcDfHn9tX+{be0h$in>&AYniE*cE+7B*8NrG$ovK(}>5-?$@9L|BJ(-caB zFvg$|zmpH;@IK!Wo(>Mct>Q#R#lQSX_n{nWAd141^JA3oY?DAZkAR-|dCN?fy>V{I z@Oh8wC0Ngek!ax>T4HAu%?vuvm65kzZlO;?-Okp;8j$z=w}=gpXAT?R>Wch8Gic?aw+F8lknGL zO%ZdPu!tWVXv8SSDS>Pr!lNlfCv^=mLz7=0yNcX1{?|zal8~sFx#InACD4zSKe{t3EM!h>G z7vq&c_Zh_h&O#gO3t-17nKE!G^`-Z`txf z`1KHEx1Kb7x?`)D^lgopm2W}a%(@a)c_v&-rNxsKRCKN=`QlB@L`vQ>HA|?9!+jvM zOgx}s%aB$|Y!eMfCkoe^)x*I)0@tRULoDZ)ExvPrOQcUn42ND6u%>U#8qoZ_Gy2x7 zM9b2cE4Ep60ZAG^{hj&WFTxzA&*<_FDf$>uCdD(5m+%z$g()2483dKHDn+8$lFzfM zlE;1CLw!j|U|KF~;^dN9-`+?>kh`pM8mslq_Ws_ox+t0Nq5-R}a@%)xL15!y2tfv3 z%)toe87mqm`9rdoQLpq&qpN5D$esb#HK8ok+z~tppMHbX*!Lt`IL|0TVP?WV=MgkIt+)M*KjIz?m9@gyZ5H4V63l%X^~ zax9iG47gRFR_=a7`)N;2WS_TF=#%`Bnrk`kutqjG9QZLVWXWRBFxiF%tQ?kAlOLfQ z6VPm*8P0j^Wi-`=zG%oO2Jg-|6`xc#au{Sp0fPdk0!skbrZh{)316LR2{K|qI_E~nIZEwb*6r4$JUQVNz}F@Pogf7t4oxcd9d9MOEp^QP zy&gN^s1~ll%|F$`V@R#xh_RH2oVSE8|4oe4IO@Y3c6)Sl)P)|&T-iYGXu6KAQJ$4h zxN7j000n>?LkO5)F&hy)u?SqrR~eFL170j+s+{PTDKIQ*@lSX*rJ;&L6#U- zzC<^FSw#g0hdt7?l2X$Mo2FE$!z*gqUJnjla^LGQXO!cW{X!oK0)&v@-B)qB0sx=& zAa0-q2}ZJu3v_>$!C}NzU#OKJ;#L`{QyWDQ%mvFBTM{z9P>$(EQ?oJ!`Kfc z6}g5)zNdzw8kt(-)Mlw<@>MaJ7*O$>h^8K%JK|M z?jD_64V(f@(`QWM*A`EvQJdkDr8&Tyof=E@V{yTVD!b~2g0Y0iIouNNduptnUnBgZ zsI+XhI338AQPNduf~$-;;p=Jl&Pe2ykoOWNBmLm=p>n`T!~+2cJu#f3DN3?o2U2g9 zBl=OS&Nf%TEvdF<2{Bh5$s7n351C|TNL->XvBrOAyhJBs>N(~;j<;$GQ6V8-2SGrS zhCSZ85{Si-`<1adb=kLdXvvk54waL=gzTnl19QD7o9=4^V{Q1b#{qxsfPmacW>Q7% z{?NY8Dh|1`e?^Vorj-~uJmhm$+*tHr=+-iGm$ZvJ)XYbY#S%{X=N^6hL;pOXcRr%- zqL+eLl)l;T*neez@T*~yU97CA>Bi=X4oz3@eA3pr7Jp|{ zu%U%MH(jB&;KRl^w=a=bLu&;e(*~e|-ono@j$g{H0~EH$I99z@WT%n4l`7*kn9H>~ zvK5mtePC?dh=n_21F_6kTOEL>L`!_HPE(%`6VzMnE(fCUA@i=dJpUl`E@Qb|GAAO7 zvm=cSrtHYBLrwHAP)yk{qIg?zd)L=h|CY4S=LVCLgcBhWptRXh2F6_kX!%x@s97=D zqh%~W1}#y~cShO2HSi6LD0)ShZC?bA!9)Gt7eRGkJoN?di4`=?h)bX)*=63u1;H8T z86$3V@=LZCmOz{wSMz~Dl+D9&<`18b79n|VcZEz{O}*+TR-u`za|>-!&zYJ zfBd)4@A%=bV6d{ztf96?IP=f=;hJiVA1)t`@x#~ehxp+q@)AG%qnJ_dA%6JhchCvd z-S0pDVU^yg?t$00=;2G+pQDHWl+Z*E|Hp~{>y;ST*V6n)0B$uCdeR7fBLH^;9i-HU z55G#;I(|m_*6+2n-aaFJ^Ig`-%3$8pH{So{ZogO2SxaQ_cRr)#ZQ$><@qY^Z{hkI- zS|A6dLB5z1nNz%(`k89rba7GqWCH9(9fl4eF8!oFpDWjT_Hz>}D`yWqyB_ z*VLtHWOG=#S19+@H3M**+>>rSW08JhAGHfwq@!e6zX{s>JGRos)(Rm}%FZ@y4cYE2 za@L zYtF1?GKda2gHM$YRgHwOzE_oJhIOgxuta3CxZZATVg-ga$#b8_X~5xo6)Tl!rh74B zhiVAQOEF@Gg5qnMORP|RqV=u0T9p|R4YGf0#6Z9 z@|UX0?vr2uW@K;?S}|Wze)(eHLBYxC$Yi9h5oczT zx)TAx@2MNdyHbTMb?bqPhdp%{>%RsgN$1(GHrCC(ylOvEOkwDOUc}!7H`-wBYoG!JfWFSuJ`s z>2;Z|rmH|O=xCQMJ$PG*Gr0Xb(YesBTRs7`hQHjN9ymnK@b~ly1FxE14;o+MkRFE~ zcSsMa6;9LUJ02Nzejj@_N2C#vKekrt-G`y-ViBKoi}GqX6eQDK2^j#E@JpLyw>T8snMN8pYQ=smH|s{QSG&rsa-#h(!n%rYyQo9&ZtA zj0r!==kqg@@c~*=Kc87%lliO&V?FG76BQEdi!C+bkScd>TXMXr3I{oRdQkqPmrQ!x z6J-Z$((^V-SjP81mpR*bzZ zU7GI0f;bWiUJMf4~yuSd@B z@3cB|*GTGRuPk$7ZVRsUL$WO&aF^qp_r@4~ zjWt%&mxQR+)!q=2bXG104z9i8Nrv3jckNp1O5ll5En^i)-7+k8;l+ zu=w>52&fnH*WML_KxWBH0c^cXK)1*f|6N>Swq?ThUg{YMK*}2f!x-4BFWtV?b0|g2 zxDkGzG#)-`vbX>7P^A5qC5vaMBl+{afZ8KFJ@cI~l)HW86`(l<7qLTlWQoe$tqnGo0VSkqis=z{=!Og<50+$Yc!Oi__I*r@$k zl+1oV7)71DHXb{pEPReYcQ+YPr={6O@u0@(fVnXO&+Ut?0|6kgD-1!jtmBx{5M=DP zvYN&b{UaN!!(Yzj|69@z@P{=_`q+3?RI+UkW(hHlsOVq%x{j|>1y(qEgJ-$IV8S$m)3w>IEUdz*#S{f*jye%siV6y~(GQFSp(2XZ zjgj|8>!h0c$#YW}Y>ovmtKhCV%G_#xnO{>sW%As6W-{_v&M$RbEI@+%B{Kqfo*H;N z^KAs;1mW?ivLle)Z7I=;D>@AQaTS-17sz8qTy9uWYwbXIl-oy9%@U4U4}B&h@r%|g z;91gb@8siq1`0jT@6T=)(6`!V_B`Hx1C+gXt%>8bolR^S^=$ZHOL^{QW#hzf{E$Mv z*QO6Y_V1V(t*;D2t%01t4!ywwHBikd}Lc|0)plrdY{?P`G*lj^g;k1&h#!|Oohl% z-+EV8Jp!GcdY6klPGhaOv5Fxb+ilDdg(Z|=2P+R0W!hh-6Uw-!U|UuL4dD%bW30rX zt#-l%)e?X%~GOREVVssggb3*p+9`GmV$2m$ymdX&+`rHs5Qex(C=wKt` z9~b(_F)$7|^L+?^+z3xmL6^9EtCcL90K`B$znp>Eu_|3c2%36pM^OZL&w49k#!n|@ zOWC3mT|JB!fWqmE<55@C@Kd{uB<%7!Rh#_jel?f~&Hc zaN8<2@vVu$lhsrBce>famF)zzesX@-N*ca#!#V`_4L`Got#L|Hx@7~uU=j4iR-XaC zlZWcXcbI+b^Zk$KJ5V>a7((|%w&`c-c#_v^jHVpV8}hs`4mBbTTyI7Z> zR%^n)x^COw<1hhi?WNW@a=BJ)2N3@t6CrdT$w9A*WBodm349=XG=CV&ShILf8gT=# z$zXTx_59;PpL$l5B53FPz{Kg<;p%6K0>!$u>v#w}PsX*Yq_`NUm!6DiluZwh)@xFV5oNNU zMp&284m|HpARes$=u4y7HyI4!q}hYxo;O_2S>NMZ7B){TBS}shT@GO+T?d-Ej&yT$ zbFoso8wEa6N_!VxNlZZsa05g2%MuJ4%g$DK9v$M1O|*OI3wda28QvQGcFRucQhCz; znOz);d?QayO@z5<#nTAVAZ&~bft!4zUB6U3etp3*67&l+rccIMqNnHFQnNp9gps43 z-{)IXal}hBVav#x&1Ub?D26_BBxGGYh=TFXeaWp=BJL68(42&b20!xFDk;iZungxR zBfIgI4l*X-Ax}Lg^0bmZ24snTn)1xsd}=Q4QL^i%o&UJd=fdckXK^;d8^TyIqbRtP zu3cKeXo%S|0xKBOKF%j2dw7=)1HH9Ip@tyqh!{Dco1y|sFF3)MqIe!9ve-dFbS-Z8 zZ|!L4Ge-_Pq8$Vfu-pzsaX#qwqT*cbpcdO@fYlSD!sZC%A)CjK9OrIZHZTqpU(1-m z?XK4@1@=D^+{iGNls#|SZj8;jnyMYIJe|i0?OOTP@N=*l8(-)8gUZ%`A4EkRjK3a2 zt~h{xT}6og$i>rP89}XzCa~j8U)$z#9dBbbG>kjpxJ>f!$i*KI<|TRVoXpSUW255~ zNBUocWtFzO9D=b?dO3PN>6VR7A@pc4IkSu9PUu_CA~x^64>_}aZ8grEpS=*hP}Zt@ zu-4~FD&vk-Wn&}kaALtevw5EA)rG#Sn6VJF?Zyb)glpY0ihh=TMK+9oTJ&KabNZ-yyMH@L;#solcSz85n`(>LYomG#FJxecyjj=7O5q3tk zjG4=V78g;gtA1T{65)$2V*y6o<&ANQT4Wz5s%SEf)k-#1xi{OKQcODb9AU>bsQS`# zvCpE|!)s9Wol%y>R7<*!m8R~CcDBes_y1_R5X#anf_%#_`O9l^YPO`+LI$i~m%x_rwqb-Xz3c*EBS3x1MlKU`>ZT|Vk5WG9 z_4s-X^j)#bniz0dIE*Je6*@QHySVx=nQ(mtG3Q1kyG{2ByYM+yX76))gTED|ozf+OK_e}GpH7~DqOq)-2XX_r4P^mXB= zb;|gW}m`ePGrZ`}hgxe~I}zt$tZpG1o`hMdA!*kqcJh9Aj*2Pi{-07&2f< zTEQma9|#5tYCQ^8Q1UEOPyT=02!liBMD4{^MeKu6-Dl%VU(dCGXO=ZvMs;*`={A~W zrO04F7)EJq8BIQ489|zr<-#&b>+zQKHmn0+4vr1>R(ZC%6;Q*yI|%ba44JRPT5{$M ze?1Iq{3kO9v_@QZ-#%Qm1W=GNtL*3^&Pf8%mGBS$~T+p)+~F$DvQ2-f$fz-4rk z*9~gzrU^NsD@1AhCKG!n5Ec$Xc@1sP7}=!~tE}U&@+sA_+mY?OPSTw*#^lv*zf=c6 z6D|=3;~+Gg>bxZs8TZ{qTa^5i(Gjji5%#UQZgUj%;c7WFsSH*P2K}{38`l=;*rZ(c z5#`8;D=LOV4%4K6T<9~$#+y?+i6AgPg&=e7Wk`ytU7l4j$#|PaxFf7PJFRzO5&4fn zis}iriyXi^1iIkT!aO_4AJcm*Imr`+{>m5%AUw9NK@=e9ZU;h%*&XIjU|DhZ`4f2Z z(7f#rV(vbjeH!9oJiMM&8c_;YfhU4AdIbE_6jiz2jLFFO$Avy~?6{BEu3rfMpzDFe z2&%nI*Ebg7YA2WMhdQWT*Jkag*^n$_PmvR`TeB+4X*slB#C1rhmcR!5qa(&b^?o>Y zk#R0$hH)btSGZI~SJiZQGdyCj;Jy&_o@^w4Iec|}IE9z=_&zo{9M2?^zGAJXY|=Nd z`w!`xD5QK!-&v~4{Qm3w{de&kSk*(Vbj^rCk>vJbz51^*V778SPNsxWHXI>>QBo{N zBt}YTmO~0a!CH=t$8TE-U-leH(OxI%?^plKzJ4kW&L+slTHGncVJvW;oLaH2eTPJE zu&TPiH#M%E+p=vA%$^h;u&IBV_ScqH_AE{QXQuT;u!-rx!IHE*FXy&=CYH%0DZ-PP zO)&AePUL?;AgV@R2){0bcS(QVv;6ZeCBAt%s?W%%^Z#OEbyy{FivN3w|7ohCPr`pn zlQ0Qf2#CHvc@J?RvoE#GG)aim-iIYee%SGaMO=|3f zQ<>l2>3QE@K@8v06YMyEh`y!opnYkWgwOPNVb?{f2MK5YF4DHnf%a498~=Yj1Io2l z6@dS}P zJ2oozaI82FnHi-T}} zZRCR$)%>B_8Y3m!0S`1pdc}FzIC%X<2N{k|;M!ov2Egw4{TU8!pIh&CU`EWoo|nS^ zoTTFE5i=7F;rF9=wy9`yFxvN1F@c--*ColF@Phu_peFB_%rk0zb2j8lO2}rAip0aE4 zXFysK%5t0DpwTNJEhLbfcq3t?3R(7H{X?qIA3RQi$rAsh13el>w2Ch*+`T%P6T4*lRJLXnpq8strIe@ePe%5aX$48%7`gL0Hee@$RIbN+>{c#%vjh)hX7aF#BhPO)boo!l0r7+oI*l<)US zy;I#`syDRdCGF2>%RePFXv_aN@qfJ%16fW`f)(tT&tiKt4B5}mjc02ie?TD1=gtB4 zlF9qCmzfk@F}Zkd@^UuWc!J1j)S)dspng6Vlb-vtvLG_Q`*FmXA5hUePw4NkKO`7Y z9Smgi+JoO9o7;?Ze1&YDpn8(J$@;${*}i9N>mu+1=;|xLY{QyJpp+P<&cpWPKA7JA zHoBlk@G|Po#hyk(fu76a=|=)PWM`A*05jlLT4=KSaLuv-_1N02H11Agw?ZaSr=gZx zX)FtC-Np(ndycvR>FNSp;Yg;K1`uIQ7wLja4j{puji56%Ow{0=?G>uI7&h7RaL{|H zluS3~oDni$x_fsOvO1z7S+=SMP`s7Kb!M0-*>jnDxgA_eBa{H|>3246b$jmhQd$gz z1#6Z~?3qM_ldZI04_O1yg&$#jHC=V2-C%s%61{vcc>6xi&6gQ%V0`VxFsW^m^xO&g4`k9o>9}`JyxU>q16?w4cCnz6n;D8F=(F>djkhiJgRPV~^)b=0_;&eZ zK*;2=Qpg^$*9kGKQoQ{abSgVBS5w`y6T!iFEqiw0IHZvJHM`0wPuxd?d?&hS{SjkU zRfK!vB!6JczN~S~5-SkUF$-$(*2+N-ZK`P&Dc?Eyb7PaP228~?iw68k_PfGRCNs)J zXhJk(Q#?WLUHj;j$MZ=|W-;sg0)0G9X^;u?MQTw_IRJ3Jv1#06(V?Ggd1x8E?_uWr z+&LPV7Y9kxIe^N>7ELvI2OuVmzsdV3p`C5}eo{^r2l!D=vVaeBqx7IYlq(Lf6u^=p zidQ}ZKyY^3BpENjQ+txl`m#>R{QbbuMPK1J);>{v`AjI)8u!%Tb7$_^w&XY<0)k=F zHzEesy)yfEkxsfZUR14@!BBa1y_brM7`F|3sbU}|{8RVzsp1cxP5R6MPPes{{it-2 ztyPgR&8`2QfAwG=*!0+=Y3b6bhdcK{3*_+MuQ4!2)4(cb3b{4{)$9qL!4}3~6GOoX zPhB?>I#krWzor)3N>MC66908h{HjIx-9=T3&YIuf={rXyia7iJ{^!FX>JNbWI!S-O z2%}z8d_0*&A3d3x6i?ODln-OLr~DI`NsBi5a*7zoG<>MfcPs*og7N+D_jf(^nCW@m z^E^TjvM2B;E)6@M`u)BD-p-8d-4q>jXCy|JS1h`f@x6XQ2Zy>SBTpdGS!F~;E)kQZ zjD$}=MSTSV%$_Ygl@b+#^&&N1p9Rp9oiU?z!!Gq?WP?P&EW^l6r&GMUGb0h=aT?e= z%*z(1D$MXk0Bkj~4DUq|dc+ea;ZhVqz`)zLjMB2{g`4=mKELytTz2)rlmk*|NNz7^bT*F3m>!lW_8`!@jI#)xB} zYNbm%lm?Jewk=B>!ycO@5KAN+2`S-&cOVc%Id|J+n15CMvc5D~XQ%S2#olCM74(`9 zP1adE%3kxSd0cPU_J&c~k@pb$2q=qM78|l;%P7k3wKQBxlu6mCo?+Aw7R_H70T3bF zmtiakPj7b>Rr+4q(wmkKGI+9LuzqRzP=Vfi4ZN-8MRnJgiLz&2!Me}&A7swb@EWyy z=1s%++EFIzz4+!2ZP6n^|F#+b_N#kq@u&bT8)L+kTOpCojR;avTXoh@#vX**{(Z(r z;Go5DJJd7ECJMx5E7fyMeHjx|S?`QH;i$!nzMkzr9!aO)jl~w6ly~xJocfx4x8IXS zcbMOY(JC7sjIF0weKynN-L*uS+(8$Vow0zgnD4D3Hb+(v z@10R}x-2jLT4Lvw)Y@u~q9)LaVp@Tw-1^x`NE3>20(QldaGB8#z>z~=ab)hnlGw!s z3RVRE6t_J`K~dj)CB@C!8F5=t%<6Lky>MNe9lwD~?b2;0bJ#3prCZB|#0&3W$6L@R z0IQ5cH75<4XbH7pMtDmQB}4#VD;zIM!GO6`m*BstAI1akF<&6@=UU;Et$yEz5m>bl zu4A4@d`&kXJ0p=UbhW;eIF#4UxD$?AP1G!5tA0F|7p>;R`hMa1Y4WFx#>G+(OE$6u zCHG8%7Hd=(R>1C^cegnpN+Rf*{7$#nDh@BhR87gAj=(ACqv$xb79jEp9= zOYj^@wR9lh0-D_dg7C z$D$Qw*4z9(RJrGcb~6kbN-MN(Ie1m%j)aNIOy%1Op|)a3Aoj-Z7Q`Y&b_fAKN}Tr=U7{?;(*K$YP8}sHD;pa;JO=cA;ZBl zMHj0ffWIc=qF91g?{nLA06NIV-gX3#@Lrexoe;qaEyc86;N6HqWq)ZsF+~Gt%7@nD zarSHZ+U3+3&fc~jk2V$2%!mVQeCqzYI4mmZ4cobiQvfwAhh{50;y8&@d34~GKKuJM zji*X@rakeUV|d?i9THfL0W1fBDD7*tL!yVhPKvBzHK)Td=IcsO?{2b9zTDkpy(@LR zf__t_Ff>xq;u`-z&F}WhUYq3#ft^JeT7HNmo+m;Oa^D^ilrZ zu`ibbp6?pO%e8DIqCha~j7Se7ppIdo4@R+pl2>@GGh=i_vpI+o6KBia5)y6Jye$gQ zJbeOxDT**m>BJ~qm*Y>2(tctDSW!gu*20MZ1ZE4isODSDrl`?nGLCtUd5{r=Y|A)g z6hkZ#<=ij=jaA(bjGPkao7$a`7xm!i-WW9$eX?T(C$xzB^NlgF?)hdJ6VYE?zcS(* z&07J(nDOsv_c~TiNYXMoe?5ep0nhz9m58AUWW`}F@$yDi|7r#SZp}^cBzAGTPtmuJ z2FmU^NX)M|OwAp-1Q-EDsyx?I=2D8i*;$BT$h);Gqf9Evx|s&T!@a(6;C^l(?$&e_ zwGA|sw2^8jOp9+(z(I`VvOZFLd@>e`ua0}9C;3xG1rZV4;gJpa`lG2^=1#B}Hrcol z_T>*{We1jm?9`CaZfonV-6YM1sRMk7h+c`?;i941RXKQDay(K{7HyKlU zJ+?E9G8s59yO#I@CHA%sL>uS|S+)}hyK|DU1|UG(wJ3ru=7m1>q{~qs&|fOoL>D?< z-tVGtG~;tQ5K1CufguE(h`6WPW#F-;Q!L7}@bVd7gkTwYD2Ye%MR@kGqLjO3Os(YX z$_L|v9j|(>6oRH%#@z9fiPsW=?@hrDOG0nNjK7i((gzLN; z$%jU2Q8pZTW)np(^ie4s1M5N_-|Ooc@%a82gEErps~9mc-Q^TVOg_#AKvj9CoBEmO zbp(+$l6t-TUaGK(>ZOfKU8^{B7I>v-<2}d7h9~cpPt)kDvMS|+D2dRk`7TOMZ*<@?bJYGRC1kYrIL@GW4xPGvKCi`ci9+ayGH2Fc4G{<65`vsw=Bq_ zy#+g}#P%yvR`|%VKM(5ANf{*cY#J!=Y7S@ZZw;i-0C;cP2I6UT=hVQ6iw{Ihwm$<< z?rQd58bFpBQOk#V$_Ax|c%43C>J@I!ttgG#oOJ>%)>c=jsl6ygDO);W8>|}FH!qzy ztt}5&RJ_R;WiJucV%!;t4H9CwG7j~$Gwy_=-coa*EoW!oWyzL73K2y+^p;)}xRLz% z-dHPAwBMVI1!vO}*`aW#GYDfZ8N{RnC~F8LYk6GM3b0JQTSm{~7vsh_wFkan)rhX9 z5^f<2;n5Dn1gV4p;EuKU%XxbFZCpD;b9cea(m{u=5_Y|-!hk-l-Hg4A@eGOBplj8miVkE4trbl z+KJ%y(H1fQ=DrrpKbH+-z!xCLej5z$ctW;BoUbYmruX6Z>qd9mL;=Bo_kZ~8DQ@GF zkh|p0DWBsQus`Xq1^@fa@XXaynkUc0Tvbc6b;aC#D*4w#NW2HkrXytYoj4P=9Nb(h zFyhXUg~cxoJvcNPXSCA|I5aK?EJu&(u536!@9GMH_YnSEwyvk{np#*!NIFRo8Rb2A zOW<@(v0=9vzzb@T!hM|qk@vyArNayWRE|J;X4p;Le%5+Zh3||z;i%h**(m$OZ#Joj z@B+JeC4Xwp7C)j_3fM2<;w_qaxyv)!k#3p!+$U+klqc({w0YB>77=mfn)@ej^&*dDPH+oh(iIviK2 zrya&^>}yo+j9BLcH2h!;Q|nf2!Wc~7vB&28=4<<#BAhr)bb-8{*k!fPm9jf?GjBlcLE*^l!?h68$J6dOfNj01`OdZcIAy2z@bU_RJJ z6Z9^!1}lXJsx}D+d-^r($^*(#uzl$d1-RAzii#`rZvCGL`SEn#T z076FL(jUUQ#nxT^r~%uEIf#Q;8u+goMumv?>4VX?#mwu3eGOKW2xg|wmH?!Y2Flq9 zGIX6DukSKsPKH|EWN35pNBq|gwXV`}8l&SPJTFT7+K!7c<9*#3=l)ELnTi|S&KPiW z?@L#1B?k2xa@ABGE2l7)4MFCu(=%5?kfo{}Moq|}7Crafh07U`vS@1c$Avy~l!EmB z@ge*&1h7?#e?pc-K#bdBX%+AsSXkc)=SEOp2)eLLE@Fwrd5*_D8Lro+Iy?r>OOw4A z-Mj?e>9S7A>eBwjb8Z~+mr1Kv!QOq$&R>S`u4HLqSGT8N-&Akdw8kDpU=RJ>`KZq&)1tKr=A7p=_pG)tX6YG&U#bX8OKyH z^l8>6PWTVCorIsKN|bpma<{`&5lzs5VZBY2xrTjIr&_{saT86>x^-LJ-W4&)f1Fj) zrIJe^r0EI^7L7FEShl4lpY&c#?+n7q&fs10>RYRnjx$bqCX17|e&UTj`+JPZeG0j4-jH$Cu?EN$ub0&C{9j-xNgtH{<1W{T= zgM%AFj=0EMLd9LF9GasRSwqUbIht(r-9n!P?oCewJWVtp;YN^bL?a(Kq8iC30s%&J zBd|%$jip$o*I}W1wcxGF08_V>dLRr9zZbO=PO6^}t1hLV-S6uQ4l;+4e7HWaMU?Iw z*=W$7xxL{8OSYba?fG7K*KEIYBuA8xy%_W&2wS&uJT3^}=om(vdYPUdjBVNdlnR$95pt?$7AjFMk^v1g?!AO*XV*A%fqvtwVL{4rUpF&T31o45KWT z@l-kHQFGxyNQC5V2ZAi1Eq6=kV%_s{-3ZqUsF&BNEM}7Y%(gkA$G<7YTRTQJzAhdZ zIeBpY+Ky4rK&owyq1a50JK?DDz!1ne8>b1bBc5xW>Fo6+f4=jwzB6i`@58?o>!~f- z7ym@FeiRXIq&pwh6^=7^#s-i~vW&!D;6CP)ah{Fve#zGZBez*4`fG_vLto97n$8Nc zS;w3O)j({_`@NKWXaE~N30GZFotQq^X1){VCH=c&-k0=&2Zcl!9SJcCxt#AjT#Zli zzi;{e`<;H?(c3gnmWQFC2gEhGwM!oJz4JQX(t=ZO1r)IQNS)c~ww-j-+?!II-n z?_lT=h zd0y>BT;9Y=JuueSi_uFvaNn&jMhzoJm%JyrG2+CQ7PT-5-1{z!R-Nue?S!K`CD>~z zhikV1C$6q$G<6%PpYh)$ZC!F1h9Jn{4A?C_PSF7~XHPGx)vf7&Jq9Hq0rS7qg*^8A zRSpp|L0Nhdepqp`mucO|;v+Cam^}~Sj~iibh?{zQSq?6%TsV}`d(r8zXC?iS`jOB@ z{+3KrU|U8Ww0>=itN}a1sV!v7g=?{uQ+ANGdfPFaf9S(o!_h=Nv?B-TO!3fqoiHfg zRyinm8wS@_1~27_&|o=2dYeZ($Jh_-92}N%2x16{D}iY15tuxxkoGdNY;tfxF&tuV z_;BJVM07oIXml;#O=ok0Skx%{(%Ec=GnbvVvxA1>a<=R6y~GBdg#EB_D1{tk_m$qS zGY<(KeKOJYV@k3eamz*9h(B>iG~+b11-{?p@hbn2;$1$m0EpuuOV+<)E(9iP0Hoo6 zTF_)%-~Fo-li1%GIFU^tR(_`czqj-=_(W@0@*fwnwmH#i+LQ0J5zoVyiLdksOhFv<$yloTU3@ZzpE~j_MvdGu z{wtQ8-p}Dag_aR1v;Rpg-ROFME=oU3ex=mIQMq;mu8ZE2mf+`W z-Lq`T*{RM~FjrMI74d9{Tb+s6R;nb6pr zPrTE`kbo?4c`4wp zhrq=uM(WlUV!DQ=*6A+NB>V%BWEWLaJc<+};SfLO+`EXVdoSgB^i1NK13+Z3@1KLW z|6%fX!TQ8Hb0Dx)d1}5X(HOrIsaTR$2H9fGfx-)zXv^Br7w{0?Rr3Jn0R%?C11CC2L-(a^ z<KUMC@ zCZ7NXP~UEF5B2%}3L@#}_h0Yt#ewf5--Fl@<%|wbx&{P#AL*V(NxvIm{|E;3W?H#U!FHm?x&ihvu-Q^2n^QHPpV@}Vl8S43>* zccE|u7DKdtNF_>TFzVd{VHJ1A2tr?T*Xj_>fnWuDe*qH3$($u*MV_?nExqevsMiJ; zvnM!SAebD*0<9WVb8k`tl5x`XwMlgkPY_nyq`E8j*_PUYc`ddcytpeGmGoQ>HU7D2 zF^n8YiVbjO4BR3n+WT_^Mv?T)Cm5yGwy_O46-U^uFk~{`3W4m=l3~?8_B0~Owwt~ z_jj)m2PudJqr*sS##^V!UMh=;G}0{-i)Z&5+4jl$(F0C|k>R|G&LQ=pe?cUQ)W*%A(nwc?xZ zuXT{sDusR!RqXD>$-%8eF(EzXpNedKH2P6ya(m5sJrkKN2p=TxqezQu@Xwv2O8aa0&z%Ejb$p%;ft%!*5A^^UhuzlB5wON(2Vppfd(YOya&UE5 zjs)g>^Og`m%x{M>L}BJq4_Oi8-uVLuapv=(=7bc=@ko&UYdI0f2#`oqM(wHd3eIqJ z>2Tum7>=GDF;@%+w^2m$v;-bLr)ax2at)hGHD$!t3$123VwSRttswZKnKwa$@b_yD zPkr#rdBVZPcTorWL0hDfq$|?rY4Ts^L=XECDwmUeF`Ivi^VEw?zHGJ2%i(pQzN%M# zSFaB&mO{j&ea_knzONBj+`;z#-T>}qW4mS76-~}<*~NzqM^PR(8$Lyt)lk1@Idpr8 z!}Bf(icSo=WC^0aWCCPcl>Ey_#>PcPqP~WHU}WT|?!3r|=RMH?j3}4^CD_&imXEN7 zS5dlNnY#mH7d_F#O%$+kJ0ak9QE**qvD)V1&Z(*<-iiX6s}Or2XpimXW&X~nE=}QOfCGtwISKFMTJRI?Tt|j76fjhpbe2aMf*kQ>HRwH25O%#p# z7^7w6(TFd59fl~JraC=Q+87Dy(SvEVQDWptITsm&@z)?UKvp66?ovLl>?9-zyN@H` zv$8{y4M~o6=<~id5zOXRok)PLX2;t&3buqtZGIHhEMeMTi%P7o@VhpA_+o!Gp2y0< z97gh|I(WF@%`~#n2OOg|VBzzmbZ7I9E2r2uK0mcvmH#aC$tX(($K{s-UvWeh`qXZ; z>WL7O!w9oQ#oQaP;EMw2)5Af$lj?GQY;}l4um>e}&aduXCAi+WNvPC^mDxbnP ztFm^+op4lHjVL|oa5nd$vWU-~mgGZac|FlsvLTpd8Y)XVF)||_w�MI2iNY>k$hB zdMgKbxD4XmXF+)}#JWPv8Ad!h^~ClpAt75_>(Mf*V+mWN3uD9y(JdwZdI+jEwM2k; zFGm$GWvS0ZpM-P8^M#|)8mv_GW9_?CvmF{JP#C0p%YyK#o31@a5FGG$z0{L%oFzn? zqo_L0NNvjneU%umEvN4r!$O}r1}%9aa4##ikS!|XhMv-`If!OaWAm0#p*3B;%GhMj zla}!|2^mKgJEQ}63A}(a1Tkhfs6P<2R)+@c>RfLd@Uf{tCM1C7w-FXWE=)cjMp$%2 zdkS{KM91@kyie)Tq*1nOC=bgx4#Un(IuFCPc9@4@J7FlWy{MgV)NH(>E&1oeFwtzh z(pr)aBaO+A-LZe?mg1CN#+k11+jp_2J0Sj(FjZP)&o>1XeNdKNmM=UHh?dcJ9BRsy z5d}ol^-odM2$MxH?{zBrbP)eRTNE!130!t73a{cqN9;Kge}*_-CdY!CBr|sAD8=`| zSTX_p)Oib4n`ZU{wLd@42W9R<3FC{*TTng-Z9K{x&xZgKL-W9jlkA!oY z;T$h**o{=O6>b?<j_oQQagammhWyq;BzVPI7*sk+HfV@}GP-Ey2eJ2tHZ2seyG zfWZgwCW=h<>Rya#FjIvo!XV6=9IwAG)fw5P>*{_Td zG+)jD_jZV079Z)zukAp=(yj5(4nryIQW=73i~ltGf?MS7{cqzP{xuP~@%R6URZPqO z_-|kTH(ESW?f<|NMtldZ$VRJ?^K5ONuZkOeQVpceO7`Wkmr9 zcF228_?#E&`=I|(;{SRj&M9l`(s$ygA=qp z9DiFipLTkuex_P(fzWRwlgPZ8Yi4dh*$<}oXY9*v^J;*Nh?f4yYbY-eudvQGrp{{K z(nSE#GO{bFdFKMNq2gF_W5c6$lGn945+~Js2WfGs(O5v83}i75_Tr^A#juaJ;>8Z@ zDVuilvgm-wGQK~{aiOO=;?|VQeKoFJtmOWrTNmRZyZ1u{qav>A#362)fPACVRfjdU3Ab(1eEV^ZUA0 zcywNSA*z1Ud7Q4wF8`hSX{ap2Q99&b#9}G+dO?>#%8@znPBxqAQKe6|Ls>?R{q}r* zwk4p*$Y6a$PMGUs^Cj&n+vU15=>yQ@E=;ltd)=6A5`dV=(`fs>fj`~$QuJbvP3)zZ z#5IkNhAk0$%9Se{yUl8v9T)w?^NTO-?zwXU@X0j1SBs>0Bark)ilT2X0?Fo|aF&Lx zu|tpblWiX9h#8R3W_xP_b0Y(D{~kQsw`7`2pCl15?<`%ADM8d{!zLSgjRk#S!%0|M zbCZJWL3UUg(d1i}FbYHEeaiZ6oIs5%lMSC&Z_mheOPexYjTR4{Yq}Wp* zf>B@>$_xfGHjQ*d9t6zs5R2lF>%ZU0A1Q5$CqEm9LK2r4Wp!&$;+M-5o=!31dHQ^w zde9NAaZ6ogl7m*XrB;V$cG_EVkB5xxPEC3ueT!*5zW=$Lukdo;&6elPZ*fH8TrE$z zl&unbN`_MEpKPyJ)sQm=Pi}@S>+yFwmR319i7ig@{En@6_zt2Y{=R!L4Ii8oF4D

oG927X%2*-1I-$AqE{a2~p6hZ~|;sOCE_r`SI2-$nZnm)#T(tPod7=kDPBJ*K>2%cA-K&A&9S%#z&f zm?m;$;aU@YgnoC`(H1S3K3WH zp1%GTtu8D7;k&HW_3*vX9Lbsd8;e`?^ZOr{@9%s%O-?-Z=lieydu6Kx!Wu9$&$IYM z6w#4K_Ii{5qTUch%^n(Pdokg`M!FdU*3dp~hVzW^I%TEY^Z ztOD_cVl!>A5WY8t=zXpy%b_(Cx9_8|hpaelrFmkZ_kx>}S;G7`?I zNW=r~4s59)CUX&C8iES(cEu8y)iqi|)>w0gn&IQjNwDD{Hal38hJ(m@%~jiPhA^_~ zoZ`6yNAE;Mwq;ZvHU($15|L6paWsOcoj7C~YShY|fGd>j+`Z{jF^0KY%U}C65C%Q= ze&|!43FNS5{?}aH604pOM|7SOtKV`&*(DNh&2Wgx*uqM9I1<5?Yq1=~c_QZ8Di~eAlZ48ujkYPv>g5 zqGUI+yWAZSW}+xj!=WgN`k4t%?tbY906Noi`Y!WiQe<*tWGp5&?=!jk@VP9NAZxAMNP^4W|h zS-X<%P5qzB3Y=Wa8vci}{=Gc%t#EF5&r-yeBLq;-_kA7QKcB4aV&K)a_?L=yd+Xb8 zmBH2}!>jc9RlLugG)HGlL|#^k_L~ziwV<&7Y(>=PV(pOqZL;lD^L| zvC@YW*9QR+9KT+x$zn4ul!=#i*f5!enLjc>ZJ0^1`<@A zs2B(gFOp5U#&&-vh`^=VC$?W}__c&yCw`!8>=c1`-QYDTVOsEJdk~+v-?=8u5%A3^ z`M|-ORMwPE`09j+TVCr4++Z%gJoSm8>RcTe`_fSxAQ87XS{$CDa>Jy=smVG5YbXw$ zVdF}&a}eLxxT>c_L)FAfPI2I{Th}_iIzhH8Xqrz==8^=bghP{geE=$_76*4@Ua%^0 zJR!eK*5d4hU$Wuh?aYYsK#YYl%4vQV#ps&xe#+VEniGL-LeL!YJsUzhD~?QP>*O_| zn&$d-Ah-}y(I1p-kBQ5tIe6vMz{dySdy+z~JT>1xLPhx1)z0=2>JpUl-SyG{={jV( z6CSSxZpqEGSI{jr)HN8@3#F;yF5Pn9K$=dO=_=AQC}6}E?&+`m00lT=Vf6Ix&QV%8 z;)U|n(@O&#;3`i@=7g)%^2yT6^W&axmfYU*w3wEgFhnYqPb@3xiH?U4nC5UTrs#nK zm-sirS0`i+-_1w)(Y8ru>!|@gf-Mtutr4j7v=DwjgfSt;AhM`e4v7A``d=oQZ zo}4uNlac4A1SEZBd`$*Ojt9QkDWL#}@ht@OxiJJ+G{qa$rwzYT);(P!I|m&4ZzU#N z(5fB{OKbpVufqYLlWe>WqLk3yS!zNhj%Yo|$jzH<5pOa&T~imSKki`gWBu0weB&E)kErlynDUZSq|8y~G>_BQL1|=~0ZC z<6@Vtq7>O}*M&X_qPX3OQ03d>vCm!=ZUD8o{SzM@>}C6$cp#-1dEmj*j9tdBPM3un z*sM=q5#9#LSAY15ruy3}-wFL>`b{&6S>%(uzMpsr8bBv%!K4lF4M%&f<*S4d$|K(t zGlF8bR5;Wl;XcBW{x#JmPOP75lsQH4jYExDKKd5QzL(q;Y+x|s(sa>JbxgPh@nN!J zBLW3iy6JpAwx6rnuRC8H%21CrG-8i!7WN*uQ5*59y9lLY64k+v3QZB zN|Y}PxVEb(ctV75D6#U{`h$#^DA`|S}Pi3Dj&Nt+|&ohyDF%2WNULGIQveoMS^YQ-n%vk`%OXvfC4-yJ|_o!6^D zAihg@rq+*x0I*g&uh)Yh=9tCNR*CD`GqtFw=RueY!p^u8jw=Nb4)c#Lw=Fvl#nIcZ z>|+J#D+4&0;q0i_)(KvP=`?)y9SkLDLRz zY>U?Pz%3Zn5Fv z1@`sS!65Hjw-!E>Hjhzk(c*^S!GgJpODxB-(5JmE-)~4o69^*Yj)chl%Yi`1zt_xq z?~>9m*6rsmDPj?%7Rg?9iI^tMA5yiPeN!;yI(@)$eVaaG4dIZ`+gq2(=oG@GDn?mD zI5poHr`l#aRTugsfK`7Yi2iD@;i@XaUFN!f8&gJ)err+8w-N)s^HC2Hc-AeVk|KZ{ zF3MNS5Q$t;^Hqi{d%YG5a%ar=q}qK>F6O@hc~;LCvG0gW|9QR;+0fuXqUVci5elpR zuw@HTT9~6tbLjB~Mz?p+a7WREuC(8(vYK$z=AssDIXM%Lchh7i)HhEAdiLj`2UdHQ zPtMKsuA{353Pj~+1{$Acdbf@p9u$PeCSKdwJ+|e;_qIi1fAY;9m#&eCl9VKtLjsMQ3Q)zG%4w{KP{>;l-Bxbbd7rP^ zTG$sQJppYsq9&uBfz5iz@z9^O@Hk5Mr^DH#9wxci;68-IEO#ytN<()N5y5=;8}pf0 z9xfawwDavyENgK%tp?&)iPNs`jnNz-xyMG{ITEkQvez8Mwx}s+jx4&iY0aI3L($+4 zQW4{Wt8+~#z)tHAkcc?h9=>JyFhdLnf*0)}7Q-@cln(5A4p$u6);(e$G8|N-Y>i`G zotekmMV+a$`$3T{h!L%S_4#XO&Q51v&Yaz^#9$Cp@kM>wC+VT`M0Iju{uSHV3JA?I zqj$DKNKfKgo!LbHnHHyJW2;r{3Y%_hKsneJAFT6rLc88`P8DdKY@>|XOAM2dDYDQ9 zqa;{P&2OR*#7Q>5lkuw)eCC0i5JXbhwVl6AGA3DR2p{vx*uPAN;ti=Lqos<5u%-4Y z6;2$ z7+QuZJ9=BLt#-_Iw@lK`DBmnt8s^R@($}JEbcEyb>0YBu8SwcvqkQflODo1dF7#Pq ziOH=4@U;_UwwM&IZk#Q1;I0t4UYX~dlFd$boDeIKIlMM+ADzf-W?MsMgITp?gYdT` zt6=ZtlVt3V{rd6#aU*n*qs+i;uCtlL6Hz5WGo~nOTaH)8X3=}CU_?zkucK9ofEl^8 zE6AZ>sc5xM9n(WJh!dhqT;i!H& zfbqXT+64`|XYY03Ys>4pQ4Gfz&X1LbKtT(6tTfR*OU4`9P&63~t%W?&#n>{3^dd_M zcgA}))ys&be3Q-e=s*BFfq@Z~^_jNZup*p7^gUbl$Ex4a2u@lCBP zznfHlTA) zq?o!~l78+;pFm4Ic4QQLtfI4asbVBp)GpN=USjN~g+3QZoX6Gln~b?~;aK$uhzZmX z3RF9Z9s@MT$9BrD5_g|dJF8gS*#NviL%$v-3|mp__i60Lu9QYt#>>}k@e$BH<9XCvgPx@xo6zU$ES(FeV%!MFl@(77+0_4d80LS8*etCh`$(4$ zD^{b@Gnjyf^g@8DQ#x*zi#WY!3&Ac~`|vZ6ebKFB#S$t!5AOM9zp`g?tK@yvl{vzF z17a z5@Sj@O6}9$#fCZ=KF)jK;G=t5xlMf0^Rl-vdUl8&w@-BWOV7!+JiAjnGeMSqXeY64 zgVVEv3wz%JYhM_P?5S{cQvFau3+4E^9}4$()gAhQK&@h9>51TZAY2Zgk0*+*aA(HHt4v<_^ z)LS)dR`6N7R8xg5Eaa*Eq~-ZU0D?6xa!<9$$+%0mnm9JcDOfYMVM0T1jKiw8vOjq` zyg~Qgt@^-J@FU}T?1k%w)1zR6ta`Cthe1})QsdgMU8-nkPWUX9e7mIu^VL!j93!rC zV-ya4N=cs!msylzG^fIGe{8$Ty>Q^9+y3oi;h9UOcq%;0cOK1Q761~O%Y7CU(EZY3 z78F2#f}RC~5mbS-#<3!?j2m+xh*EINyqbiP93W#mf$Qgzt8SHR=iOPi>TA%8Z12s- zE(HnBm>3RSD!QB#)~qX+?F*MUx1PZR*0ke|ZJ=2oP0gQY5vA1{PP0g=id;r%sX%{A z^)3}};M1xn#**lGx$HknCChtE&jjA3B2Ys(7GB$J_UFQhuNTgR%Oc11j-DO`lm(~f zF0c;V}a^$ITyx(m$SEF0OlydSn~F&cEWKj zjAB!TaENC8w)=<)MQ%p+ajwUs&K5&{6k``dCKtM)Ot!w}nGy&RyBEisE1pU%#YuOK zrh`{B4Wper60__EQUet{g_xVXy}R zaNH*j=_upsL52imle^0xA5=9<0in%WJ-Xb{b8dDOzQ`E)g+6r}AYy5+anWg!Zz`%b z@RXKgM@_Jfpix;>HFeUts;Pl?#+`6fO)2n9`x?{vHHG@?0Lh+Yi+jB@={}%6-<&`# zj1!;3tHS4+T?~EJV3%k5D}Uk`o5TE`{*ol&&5qUeL6RhfUrhBNNzPeT*`i9$?P0Y0 zSTZg|JiFB;7lK%^&)wR{T0@`iZM2{PffTbdSe%jCA<53ct7TjLfgl}YqwYsbq*hVQ zn!x0?sp`r_iRX66v2z5h<`YNoPar|f5fkpNnvfc_YdjD<;iQ%>`Tb1TIh*^8m8=Fd ztw)MOj>Bxnds?=e2Zs-xsk6sP;R$fI*Hgsi@DX(O#o;3in9WXo8YtJ%4 z(Ajux4W7whd!^^fDq~bS2VXT+RgJ8dvex=rH3VDr3;F2+NoO~Q(NGDN8ZIXkL|o@} zShgc$|2jhRmM(TJyAwQtRi47%WNqio;PS4*iorh^IWgpEo?;|0k>zCY+{$OXYEjL^ z&x2s)pW7LUB+s&@+nITJeritUF5bIMsyGa_WZ$liMtuZM+>W1nq`$6PNHxwH>RR#Rb)^1))?W1~5fU z&;K}akLe^4cu?29gyGbRx zgj&mqvFWXH?YY8B8e4U9imW9qfbFN|G=HO-56$8FI&s6hOXaLgzU;1@EHkuLM{GF7 zNQm1IuZ)1{mvCSNZlwiJKlcNNegJ=}iQ-ylA40BVPH6KB%ewT)Af-=`Rq7J-y%cj`WvIvR+x7phva5&WxiiV%L}!3$e4kK8*q< zUu;INd=crgCzpJZ90~pR$~Qj@f-bapXnV8H z_o-xj5f~*^a>FQO-~8UYkJc8;v8e7vYl{i!-ul;-X~KZSB9mH|=-T6*qzNAM)Nc@t(&q53{Ug0i`lh7O@ zHHQ-e+2Y=_%ixHC6v8Hau2X+DIGCszZS&RNTb&mDnYjnkjJT*YrFHHuqT58jj?7or zpT<4Wp+C$#UKYMY+ovAhqN2#WmI_>u7FYYy99c$k_4-5OEp2Ww4nIki8s6&ZQBESv=rNWBn;XtUVOOTXE>e$l0Sxuf@A&v~AUZvL0K|#J76f)HIIJ zA4xOHWRESx@L9kXzoIVvGYPT1>eU1>By0=b+6iBS(84`IFkWWg)u8f3IrnyCTy!wu zRPriLZ-=ZJJXW}NK+0g06wN41gtEJKKymno>LDVYk!!}tULrB9e7u&(ofOxOijg4D zfZg8OAOk2C$U2Dqv$9V(pN~C@_T$a{p=V_iYKXF>rQ)SxuJ`7I5SsgzVw5#N_9c6B zCSxQfW6*Pwh{~pfa_t9ze?!cg5j-y+cg9&G_kTX@UvJyyx_1Ag2olvc2D2VR6sKY2jVE z8QI6)x;Y|Nl`dQ5ilf6w@2WEr6ge%`CN9LMimU~nq4(kfSYokl6t{FxH00Ny>XENX zG)^|Rif~poKs_o(V3(`OS4N(DIdc$2bUuzR2LhXXbmvY;^5OC(R>>e1ypZfw22A{d z`q~*YI1#a284C{{(?R%3#I|)Dhu-7nEyP=i+-r;F7*&ZJTGqDOZ6!MED-Y3Ci9_%4 zY%(@ROB}Gf-pgQG#F245>C3fw%&E#*gNW~PYmRR*g!CYa4Cb-9b^=kDN<36W2wrRM zRYe>oO45vxPr$n?qo@hC%vzKZz%|exoeX(;&Sowpf@)B$o`}4Hbv9imVvFfs*RN>W zbCEE>nlNu@ImCq(T?LlF{iv`Z)bemZFm zMDJ{_>t!?iddPgpaTutWxkpTgdo7)*vo8@zGk5Op>Yvi?wUW-h+-s!~m9A4qdJF4j zXZlx9XU88&m)%UfY-0#Z$Lq;D@^!UCkk91mTJkHsU?igndiw8I`s*plY;LSCjDE<( zSjxuA&g-kAS6$5r}ac@4W_SAs73Kkym@(#H*;>mAjgB<7Eum+jd1LFP}|K zNYV3JOUcmC#Y?;jUuDT?3-nf^BlrrgMSF>k@FdpQTdTmdcEH0ZMh8q}ayLnnS24%dLd%#31WU32|{Rmd3tgGuh`S@r{fLhZ$YJJg2gTpYV4zmTWSmq``! zjZKybcHTyb0l1Li+5BE0PByh|1}-H!+unz0MkhPhc3yvDBofI&AB>*kpx`&gFc`UG zVY6iJrv#fge4B3y=el01FG&{p1C<7Yu<%RQ#Er_v(Tq%cD&%-TQ$3* za#&mr-9ZRyqIX{yWtKFALB;3{XX?vXZ~f@74!mDP?RrE6m3?<>JaHZ_=M*g@YT&4M zWL)$J{(a^!Ekxn)DCjyZfR6y$TQi0r%EHGxV`5vQ2N{r1S?j$f5F?%!#HuQh0C2u* zRe*N*bZDF`TzI437`fR2TtP8DSBLjOo(gPd+zCg`CR;_WIzMQ(4s_YG_xaGB906(; zHF4Q7rvY@?^R1SvC=qO`UUJu&TneVJ2v}uTXQs|9&y8Pa>pBxyP^QS9E?Ya?;8l9% z5MQSk*}XK2ZA%xM)8mb}i(83%9n2=KT@30H{6e1wHNdWMXIw0C?$vBiJuieRJGJt$ z>j|SEwzWI-y={l_C2OgqPrxL+Hm4n_mqT*`$6xi{oQP{f)%`7v4B1ybpr~7}Lp326 z5TUJ>p;#OX?NtU@1|6~Z6eB@L0H40Ky5a#{^jzgADm9IEuF;8|4ku`zS zY+T7zRqavjGGv>K4p(8EeHEkF#vD#vWfTiU_Gbsi2#Q0wGs>=qx1O3YB|=oBtEgEl zwbu-OXUtqW`hE7z8Y@7JReuwbO~%!H{w9Q(%h%I4;UCvE-5d|7L|aVH&EeEn3@H6E zc`l1g-G^)%v%_ZHCc#wTZzX0;IB9lKpGB@~7mI4Mie|&rL9~ zWkr9>0-;KjeKyDdlh75gO#BuMObA)*vx{G7dfu~g-q15H3RpHWGCEP$O5jmmsH%3v zhEJTY?GQa`aAjAF1Zb4XuZ+2Qp=aGr@Pb0NpWDfsZ{WTU?Ue7_x^;Yq%9WO)i{GVM zOkYtF<}{vo$NKwta=Gd4I-cXyf)nf-u8fW+3PsoGhyf7Hy++Xgve7fhiE(QD!Zzwy+n-#Z%4a^(zqa)L!wjW(BsU4ar!ZS_=AVB8 zODog!mHu0$|7r1dwofFnzt^Z&u0&Ekbsck5zCow+in%1#`;s*iPuNX~uNmV5Wrw42 z`1dmZ>&cuspJLDaWbR$cD>lisCl}uRa4zsleic&1q{~YouVmX|r6pAgi-ar(su-?x zWx0$hKcVt~*WB}q1x!Tik^UM-?gver={)cE#4+uRWB5Ab_d3pF1l7tPcq#?tg6UsV z&znWvl~bI9HzPrEae3Zzk7Q9PK__!~!z;a>Jhn)n?fO}zx_7VF>mX;On_Q@zI77>xKAvqVo}dAMSGLiha|Jazc;>JJ0zZasPr}} z#i(@ASYpEIa*`|>Vge|!GMM=$u42hFo-1t(7!``G076=>t*Pv@#31f$y?ZO(3QdZi zyO%RGu+3HX;wFZBc!KD=UcAQjYfX$MdAvuq#s-UQwY?V*a3?Pe=0)nAjRRsJ2X-mA(Ggo%6(rBu)ecN75J21I)~TW6$k#E zO{|D){I$09cpE9`H&UQH>^aI1f*hRZ!7;_Pf7gF16#V_O{w{?ggL z#$F`iD)G-kw7grqfp+a!1NV`ye9Pg^V|l4w2AnYy*7V%V9$@6>PyXtv%eF6~Mp=)`^$Hja8sdDtvd!MZ z;Rs@Ulv?A^;}n(^TN(7&=nizzs#JJ@C%yPMq&JzQjE*2gSEb1;f0iJjf`AKpoc8QpdSxe4&O*Bi5 ze+_+|nl-UY;_)zWI)8X@Ayj@=bD2mgKVJHxvUUEFIZ0FLvNLp9e+90^6k5KT<)|hR z*siDdv)!&7x&}yMYsyCCkxSFN>L^2X@ScU&OzZyni2xqh99*+s1K7Pz-e!taML`e?KbFWqG9 zC{yKlb)&Phsj|Nlrh2n2IdO7EXP=&cjxc@+zsW&u5|Zu-=(ARW7iH{hMZQZ8-*>jh z%{JZGYHg7xC)+FCHdS1VDv$IA+;=IEgAJl$YewRx%cpi{tn%0R$@ujmV4})02|?6g zEk}*eGRkHHXUU%NHHBOY76okXK`PmVdO}p)q>8|Q+e1GjdFPkWP(1~qNPIZgI#cIPT}QGscO8PNggmhRU#LWVu5et{raMud3M(X zC{~Pw48afXA_`Yno`?2I6d4lEH*|>+nBQt|UG7rInUjCTm=b2CQKEcxdL~qG_&6ls zr9_z$haqveSS6cv<>d({cI}7uly#v`Lc#0BsTssmg|<2UByKy2cTc-FanYRkOet}_ z3G#25TAR4OAzT{$LHPDu!Ce)DcR7g(uSds&zMi8Voz6HE@6NarhE=@k|O7z_%eN>Uh_*RU> z^OQZ?wQ<~)__EL^V#N;a#+T>a@xQnHt1038_Fd&>zw)>L^a4Ke-Ja5nSA#Oe8FyXAW< zY{&PKE2kBxu@5Cz5R_RYAZ=_z;g^rp@n4gps2wU_sJ@x3!n zb>%%cG6zPFPh2k78ry+EaGEjm&O+5V7&-68MQt(;f+vWUE)nQULKdtg5{_qTj1)@K z*xqVT0~gr^w6B z`;>E1peTU)anA8h5UlOPq{Hv3K@{y$dE#AzC~}d?^Q0&qnj<=cU^P4A=s+S0_?!-m z-V^Hn+7DAd?2J3%s5o$@7k&J}xGCIZnLM&D3c+#rzNh06LTsCn-ap#F$*rB$dXtlx zTOQZ&@T%ftpWfc}t&H6VrES7Y3wWsn!7q$g8@j}Z&u-jHoN7SUbLSeM1k;r+&a`$gPDBeaK5p=dV)5lw`(F`ZRk?YL5k!xZvEi{m>_I zi5tpefrGfBg*h~l?XlX|%bBU=y{^Y|%sIrjpjPcR|XyIU<=DabA z`3AnZyE1atPnZP7=tD@v)VMJMHInRsG17=~XG{Tyovax}Kf1Jhe>)>@?qb$cueox@ z`g!V!RmQ`@)6@euqj^m!sagymqraqiS|y6zeXhK18B;k`?&ZUI)|ahvz_tq>)t4>- zrzRAyy9AGKfc`#9#XT7DVh>VzHlk6_DY0dE^LCw6Qz!3?JK?0Pxj2&@$Mu$>E9*?) z0yDA?ZCo&1!#`cSgN@h2Ly(h*&)D-Vdpc&XOv6#~VjkjoKLs?Lwv5fBeDbni_*`)& zaL85$j!xTM&9KU9`@D-u_CjKr<$|CN@CW294qcXqx1|WYfyy65>?V)g434&~H^w)-KRkOQB ze~S7G(>^x9R0cf>If$&7WpCr){M7$W#v$lQ$eskC&f5SIP*pUBo@?IWe;coJZ7c$c zaVH#C2Qg563WxiSx7C4F5j(19ANL)p5Qu0NgRWn6mVC(U<`o~;!ctAx7^lu+`Sc5Y zxwtSuZeMGZ1=fXyzFb_0eQwQ_QHDoW(SxWVqdxDV5{SywgQ!GH23_G+qgX*EE@Lo{ z5b>`q?#ft*2b&Hx76;r#x{6wXcFj>U;=0fl+jEVMB~fOFVsr%30cU<^?BB2z%L!-~ zevmS*r|hm7ST&7|i-G`5CoGd}Iq@jf=DoR#7M5SRepQUY5(#sk3saeb`imR0Vr1b1 zoc>m$?9AIED8*RCj*>6o#wco~L_%Tgj2-~UVz0B0BZn5oVlWPCGY89kxhl~U=`G$x z8ERu^+zCg0;sSeu)7`{+G{7+(Z)lti7J2P;TyAQ z%Wn;K*;+w2X{;e0QGwt{UOVfxuhiD)9tPi~>n+)eq4V1OBg<#(~+EHaa zaNs9W!ij^MRdeES_&%OEB5rqHPYx$j4$sa>`2EbjIg|Trt(+vJ$9e0Np-*%-C%P=w z09k7vxaQS%%YAi$J7j*^;f4#`NPp$;E^s6LyK~s~9E<(O#)6-|%#^6=U=va6i6inU z9XLd&3hc=w{C)_t^LnFY?d9wv_A*tD-o0`5L6#jxI=wA_^hooK3p)Ga8y9|M^qh-> z+S3cqt;^(kBo9Dxq&AU8#v`kAi}#Vkh}`OtLP84%`(9BT zjBpVeP?cySsL*y2zB)lRd2MB9AK))fs9Q$v%Y7Xg`=n`b*O^maEv^SXh`Dj@7u=*< z;3LH-CckMr(+gt)#|D~0TW(tC+@lt^eLwwH<# z|6+FDx3R*XgA+eBB#xCEeEm;D;#$bKkPk!B9!KHdqoiuDHthPH1hE6AR4v}z+7DeS z4ys=z{wx)s5~h@sRG|bU4z6~4bMBpcA#?)Y4>XosyKs9=5J{wT8Pn6LQ zTn2%sq!_(zxCqx7=80zE7y2LoN#1rQThFpji$KwR>RJDwz_~-uPQ|#3yA!7EwdhO& z^yPebtF~P4M%lB4_C&B}AL;F_tx@U4wJ38sJ;-oM>3cdbJKWyx>G)-frMXnxzfG~d zg6zF!1De~_)LDSo;DxOi0m5p>Y%8N%6cbks=D-NVTlIHF+4>WnRWpKpBio0KacXYH zlK_zKjc>u3qKS6aI5aok5)m>vBH;<|RdbhS)4Gvy(cB_K=EbfOSIh%I7H{nog(r$J z7{w^cOB8xaS4R9p@hRKeiCd&t#qmIp?XWuxqT|TjE%UBb^}?uWN<((ZpgkzYL`z+EwOJ;8$3hd(or)Kmb+QQ7g0PX zzvduHz_<)imnffZX}1*TKC!&$UK_vIK%ScnO(513(95bSBEB?DerdH>5FF;!^X0(GUN=G91c`Py7S|A@-27zr;{)YPB0gxD;>TFr>F zLR&vB{p5KMXN+!x_kkcMy5|RU$W4@?KX%5Qa8fK{s2Mz;54MY}MSdt2AY^ME*{AJd zEEjWodI4wqluk0CTKDvtcrrWFzq)1;Q^Ll7avVq8_uC>Z#&v$0^W>XrboKaPyIJyI zuXNd8%SK_7MAEZ-nK@`lw)9`Y;h~&|A~rJ7r!H`^n3%A4SPPTHBu+}6w)M0p4#HJp z+z-OEKjb^f$>_$(&QP>r)vd!>q&cGv!oCWP_K8*^zw{8+t-=gdwq)kUoScE zR&k3Q-&YC0JCBUL8M6(O7B$yjSU#a4_hvYJirjHM37z}pXZs8MQ5 z?fbdZ7(}14JJ*c3uoU~om~$MA@d*&686!4Vsn;68jW(iY6k{eVGLh*oZ7jU1jaD;? z<W&RI4|@eE-yLp-1#Uj*O_?k9K?;F?wW&C7|6TZ>kI5WIH*8TZy4 z0{e<#r9TKx9a@mLOLL}L-5Gbnuam|FG~LFaHSV%6;Au7&!o>4_DE4gHIULOtW6D0x}s(N4Qm#Q(xg|4pi}=(~_p zIW;GUTK7LRCwU%+Ej3QHl(CF;iLxINo^&mdHzx#tRwdSU-s5%8iH;ARwdO<5@%Dt9 zqC?L;kT!&2-)eQJY3@1pb~tv>O`*?ra1TD!Qm`J65>=un#B>Y?f=I{XNJU6u-#*h<+G^X3qnKr3b;Y zAbP9Y$z=s*#%=n7Zpb?y&Dc%-OFO4Z*co@iQCEZnBvtyTD>53kjQrzM5M|Kfxy^!b z+qeJt<3b;wf~?&>(B8d7nUo$0Mv1n0L`;JyrfjfpJL4ZW!r;g}^7OhYZ+JK#-h`L0 zXzHQ>>|dOW{d4uiC=RVvY_~5L)MRWZP<~`w6`kmi&}WJ5D&MI2qzR*{${SpJl~L^C zoZX>uMHU*6^e)o^Ol=O64vYkvDDhUK=;IdbT{X@Vd6t+$dx@;Bz8P(o66M?HN>np; z&+!}M(8oLDPB3u8S)4$H?=dJvK|J7e(E8o+N+9Ra$XW+ZA# z+vZ+sETv6e)nWKn@TndMkvCFCS5*sM-XCk#nQ1sjpbw< z%4cWX2~#&NdP&rV>^o0KjMHzr;}WV;?N?b!b7m)UdgCV&wjGRZ4X_08&;P z%f}>JUtct*Y%XQL?pIFvD$?X6au=Zh6sM;S6R%z6q!m(l)tsrwm&&=-P1>Ag<;YHT z01$D&EowsH0N$GeA@YhiXhMw~_Etq`=gb5Y-Ngp;fcuJTgTXCuWf_IR=_?U(-L=8< zi{CkS!dQ40We^B)DJIiY2H=)+b#6*{U2pgXRLObC$lgC)i=U?@j=X6ug_r$uZE>gr zj*>m&@+r#n0{KE0z7s?Ty-WAMZHuG9(>4pV_sG~HKWCtmLCdr~|L*j5vEFq&Ae|3WFD{WocH@;QeD~MUSH=Jccw!NQwHS;c;P;R>SbKUwz<+gmyV{WY9pfCKLeB%%P~lF7hO9DMWI%lkhzhq8NS= zG?g(;aq7r^sw5%GhHtVCm6RmC#EsGQ{zob17RT{J|9c?OxhvHDkB1_TLTYe5C|u$1 z6fP@j2nA!pS8Ymx3@u=BL>r@1pQ_ zMwxq#Td*6W=S^mq5>5gx_hQHo!q8MiUJ)(FD+rm@XVC7)ba|rJ^^A+UzzwXdg%5(v z;z%IYy&c3d1hVa|9eL4*GG}v>YGV}LNtmT;2c%SiQq!WyH1@=P(F7ux)NmkV(8;4F zh-?uLGEFEASwu@MYfKVNzBSiReapM+r8}m5#e?|bZq0Sv?ASwddBbV&CRPL@n$__u301}LgD!&@lf}w zs(bLwE>#hmKE!nf4@GARmfQ#=Bi>q0Cy~0vU1iw3SYD>u^T60*og9tH8tHk zCOFP67IUx45_VvWiDUgcqZmvnFZ8JwV|$ByWz5zlU}9AV5#_eU(EHRuM66+Rc<3O! zdD}X-OT{HGxTmXBVku0$Cq`OZA2vp=`#8zIHHTG>$*qQjhpB{YQ0*ilUZFvs+UcAO zcdDmZ=3K;|r)Y6G=;7q}Ca#xSs?+1`JkUDvI^It7V^}BBXb$h52z~akIWY#058a_T z@?A$B=qagW=bdqex=7`TP)~rMozWZlya_9NW0Yy?WhvB*4k(s-?Fa98omG|@HVB31 zkOzY3gN9n6s_J{|g9*f|_QrlXh(w}ojRy`FtLuRLVw z>HWM?rN7P_NBY+}{dit^;(yJAeqK4!zpoO0kJ;1DD+lR!Cy0q?SW&w)IG4ma7`X|q zHEBkiBV0`HjA^X&HrU5CeixrNN4{x@ogf=0P9zS5%+A1SLL&SfDsby!#zPwHAT;~z zIKx{;ZWy0dw2Vua;rzj!h-!H%MB6IYF3YA??mrY`j)pJtjj{AFv8H)Jm$l%*NW1}n z+HW;h;9AO|=PHMmPK&~!Ebl*VB|0vffIm}=1h=;_--eyVD)M%C7ll)V(6rG`5Ixc` z8uvlO1;OX%1Nd&j2T*v|4&V|_6su-ToSiQv?u;3XuTzau zTnazbSi}f(Np}$8@nk;fX~7jzU8nyx7C;b#0InHVH4d2gb0{Yh+ z31UgrqSb(x-2^C!YS=xCaeP$OkHhY9-+8rS8}_*_>@D93a|K_P3^8N@_gxOPZDbkb zBI9KyNOqabL|=*{|Ij{@AVnH35#=1SbmX1IkHJP?CqR{5AEaztu7t1 zw6n$vRQ{n>-Y;7{(42&DJoVO&@;(#`fEzwUoLkL$s+SUxEL0tp*oKzOR1;ac;b zon!6?bf!`2L7g;^V-w6GMYrl)LqiF=X9!LsO4n%))M2z_Faz$E6-S{9 zOwt?#C686l5-uBjSqYs31WCeL?;J9Dq=i0cf(1txgOLv>3o6~yTna|CUEd}QZr{*N)922`2KJ`CmTtoqF#5cpPc@dq=o4j& z7?+2ki=dwFv1*t2UJ7YUyv1%+u?E`FSGy&VD*^`P;!28uQh@tm709<4U zVKRE0fU{Oe^9=1Q{8WlN<4!ng%iKW9jrMh4ZWAn6=GmhyU)e{xY%BuXmn|JH5>_ou zXqScwR!34QFfg4@G@qvTufV# zj`cveW{4X-5UzpQRb9Djw%4Ic9Jib0&?S!7p9K6>1lf=ftX>m%`P(*ho$&klhUSOe zKi)#X6Z7Cqo-!Q1dxBSGx161uejchbb@qun8AUWH`2+o%*y- z(xn4q@Jl6$8>PVp9;FykgYA`0{0Xk9_ToUO0^jYuINRB5d;73d*ZG(&Vi(7o97Od@ zyn0oP4eZp^E2H!=caNGeydoCdL`{T_H3vpML|^wd&KhM0EBl}WBLGtMhZ-yI_D(fc zV#yt9{OWZ1(76)gibafx%h?hcGafo+WbB53$FG9jVo;U4h6FAX^wM1Ch&Xd-uH)#h z+?(sfgqM$T8Pcy#2sqH%kImtXHOi`ZYEH<7nDtsJ7PwOFyHvIgM#40JHm@10tg%Z6 z?afJ|RXNd6j2T#~aGlx?qsOit_#zPBGn5n#1dzq5{uprTh^pQ zdivLv-hX(J|GJQygx^z`tvK_~KP}$Xro7UBt903}`czsuf3HceTp_jRDu<5a^N~-% z>Abwh&UML}0Umc#-fKqn*y3+z{@0T^0`|nV?aACbqRm>u#r8rhn6)wG&Ii`17-?}Z z{?*fSDwVzJE4}9@MnyAsU^pmi`91&5qaf&!Fec zQdRM*mEZ!03hPl!_?6|;12FiNdzb-P_8c>rOPG$NCDh_$bNb$x#B!+V%2YF(Hn%fb zV+ajw7cQA!by^4o&8_MZbFR20eXGnmaIj~oGo4*>aXKA8)lhXhK|HNj>7KCHqNMk9 z>EvXVdb_x?RN&pcifheeOGL|-6?nfPw)ACzIXr`8 zpFBjCZ^#X%Oi0%pcBX=BzA+aj&f2q`mK`K>E3Wv{kb!t}hU4g^%i@Egk6y?nzn@K%KdEO#BL0xeO z0oQ<=ys~33sbUDyC#v)=9#iH)NSt(st5QI=%YyeGI= zsPB#Hv3Z{@glwFRn#Z28^Hu2wZz>y$j|##A7vS?jLktzR3%aGBR9odWT|59nzpFO+ z{@C)pau17#!(dstdvdpUg6yx%QZ^1wV>d@wNFz0DOcf+tY4(}P8k?GT^u|T3DR9EC zY?U`N@m+9qQv{|d>+}rxy;bS5RuPy&RnMxubj)78=Yk=5Rh+M`!OZXJGAisqRW*TU zG_%k4YQ)OT?!YE334WJunmn?aS1_%2DJmQFsvBE#0nqaz1*jag>LJ2>Tn8LB8JEdc zc{r}wYzsbi8(Gc=>^g!|)p7G#ZfxLg;54AvfW6=6?9_Zg)U~u-vBic_ahvKGOavlO zY>CKb_S%!Oi;uyo@nqxVxU{{n0ad|W)Ha^30c7Q%)5XqmUZ;y)mRQpE>MM~?>#cf1 ziV*y1XOjs{Cnk9zT9Wy< z@_uc3iK$@;zcKQ9qHf6qdbL0P;t?!Yb0aU9zOtSr@#72GCXt!LG^S(@PRx&GN~Q+E`_EQHeU4EN+23Z}PF3H? zhf4lUnO?0Vcv2>=;gVhU5TVPg=*xwej84UgH%ysdZ4o2R`xYys6J89m6#CKv6Yd=C zqPhi4+c}DFu+WObHDB5(TJzb9K@&vgQrk`2tL@hkdYvMpd7{;Ab%wS#G@66RI7cZX zjs>SWB1`|;;IB>)y%Ac3xPZ%1IByf;^9;JhvUz3f>q5jIRjdl)y6m+AMb^Sr?{a4J zZw{J+Ym$88DDq#oa`Xl?nC=#k>_Y>9VRuibZ4CgpvqoQ(IL>5_HmwrJ@u*d-Ce6XY zUO92#%-~NP;J*5D;E?TG11eS=!r*W>Q5>LQw&3YIN0EJW3#GMllorogPcbH*Ei1=s zzAM77{*iC?Ku6fOQt*Fxe;PIglQ?@;2PvBUlfu|I#^ zAHo>3(a&7)%Supsn;F{B*}1fHW>f&LipulLGO7W@WYBo`*(7;8`8gS zb^cqWbjk644NJ>j=IDj=sS49u^#3g6DB*X{Sixmlf*8aK9PpIi15~DlOiADP+|SMX zuUTvbtWwVe#~O|^%VmDH%`I+yicvIJYH;WmM%h!ul&$Wm0Nh=PA15}h%n#bTn%J(z zi>mqq=e55W$kz{?r@g7V17qQcW&ci>FG3c|unyZK6wu0(ZfC7HHJgv^^oBs-OZM^g$)=xTm|cb1pGE<4zbyJnQN+khJf2b&{P|;aqvYVSJO`F57fE2drj?!xxzf zQFa%s@D|3=y18nIC7w3iebPD3jYXyNhC-@InjjV~F_o174cD+OZD$F1mRp-1RV9Kn zzGJvQD^QKE00bL>V6E;@l-TvPL0Im>#TItWZ^HSx^cYFq0q7fnv(AvOtx80O3wk^n zup+)9H^wYhze$IlFJc|?hE0`NiBD7RB@!FNRb67kSz6KeC!yjgV>H1MR&gDx{yhRV zNI2IYqXe{!`1vp41Qxwx0m1S?HP8{cr{-;tRy7*%(6E}(69&iZqwA&;l$}g*RyQYz z^_`SIdwR+(*xE~Vo~BwHF_3zYRjE*Cmhx2!2JJFjpEzD$@d5V z(9?M@%Qvy5=TPu-c#k#E?CuKm%2F-9OtH5`wgSGHiEqWBUZdsJ*GPYLJkM@h*Tf8j z*ZK8MGv2cS$EfF9%zd$c(pqxY2GR!=?}3Y?Eo4s%qPB@@fwyRZzjlHr6mzn=Dv)Li*yYDCjrf#)7hRVmA&=%M13jP?nO9W_S-j*CfifOlI?1j3+|B?8A(b3ivP z0ZXQm@)jUYU3mcJKpDRj?dRN~sBB~fF9&hK%g^~BRY~p%TeK4hYU|hN6N8X|6wpQC zm(ULI6+yI0jb4wmqpyt9br3wD@LGSL1B&nq*F*V=4zD4Hj&XXxUF59G>H&9nVyCKm z5Le)8%c0pa5mHEc-V=p8HcNfypwHJJE{opcMXie867uq)_J(`ha>MIP_>kf79gmP8 z`^EMlQ)jP}Qi&!Rem?drI{R|$S?00li~3Yw+IF>~40`%2KXK67VK!zd($WdPpUE<( zVIO{(4APCSM-8VOubCF@FWIz6M}hk-c3h5!0F3hlGa8+J1E@P>8R@vL@Em$emsPM> zx1@ia(~kgkC;orE>Jup%0(DRSzDoGrd1TKmmY7pzao!FzuD;Ys+;$gQ_A@hU2&e?u(V8_#~u z@cMZ#u>?`jIp0JPNktuJ-D^a*<`U~BO0%VTK)H(omZU5;dVplL^Qpma?p1{rJknof zWF`a?bcr&r2xF%S374vPXjKBL&D~p-#l&o{wRcA5V=h|8uTGb(Kz`;*kUl?e`CCTO zOsI&San5YcYh@PeLF8E0)|zfo&xcPNqaqADc2z4j@nZP^I5Uj~$u2TGuHbw2UWZt3 zm6!OV81})|%&*I1=SRh7We;B)OzFz_@(*oaSILBZnT%hZP-AQ;pt2_wQ{uc4YM&vm zctac+XIJx9Np@yS;e#Lp>496_#l>6v!M0w+U6g?b37E%fJEIJMq}6K&$sDi`p8Lia z#d0hS*Z$l=Jep{b13RN=0{LA0wZ_E3raknmtRj|+LGJL%6(6jC%M=c_1{4WILBkXv@`C6qoxwGqSkVNj-PK_F&%1naKbZDqs|**6qf~yGK_q) zdAOwc23h^oH_~IQ1|XE4{`-+{3O@vFSSqhg&nZW5oSTi4Js%LWKU#NIXWy+mon@qV z>uyUQ)>%*gI;Wqly957g#TeFIPyfD3_&rt`Pgeyc3o)G%^NKj0#hrnu+bSi>lzx8Y znm{F)Zl@G2S@$lif1BvR^pgb_#$t4%R-99K8S zTnD4&@oB>Z-xz^)GVAQvYLtNjYfv)+z}Ao)-x%{?6kTIdJTOMymzjHwF-KDt`edv@ zY*u$=lrByUMQa-`%PEU#u;D)sm<6&PHvv6jw&27Yf|ja?taa(xT`H-X4WniRd$Pgc zpBWwTRYM8tBWZZ@1fWv{pawXV+p7|5N=%V(uPWeNv(N_t6qRr)vGABB)Z6Gwhu31g z)r>e2)Ur>TA&Lj7GOms58WVFixOin8`f_L73BOLNG#bLU9+2GU$*g1-6Ud;*diJiJ zhWhX6FQH_XmC}%F-}6^)r5n!fdw$>|f?>>e0=Lz~G2TmR;JCKIqi!5LH)?^}I-!3J zVg%ue0l|PnaQzw_FkTHC^vSq<4K4Vq%~~9GY=nMR;zR?SUez3txYq0ncaF;T)Ednp zi=hmpRU9}z>M%h{A(fnex}9(CGna{*jE1K zRCUVGXQ7|7Fae2ge?G2pI{Ut$>ntO^FX&tPSde@A*E#*Xpda{OAIwyC{j%<$4fB zmL8(}-lhdMfeD8hOq4EH?=#r*c-A!>&UZ#n@VB-#yD)Oz=c}G?Q`Nnd=rv)k#fuh{ zH84G)rtfPpj~%%V$Rqnm7qh7yi1zeXYdCpz9GZ;cUFNM<*E+JgI*$1^Hk_n*WpiN@ zb4E1BE(K~q#<1bZ?gV^alhg9yW^anBX@10DN4Jg2ULsMS>yePe2H;d@uL@tgcIa?! z6v%otFZ8KMBo|KViXeLk&#Pxmkd19a^4(>SWkJ^1_%1`58lsG|FiH; z6t;ZiT1L#CCm6_9BDbdoxAwinuTGB^*iz^oOqrX`61GsQDI?>evm9Tw)BS)y7-%Nf z=E$f^UscT*iEmc+epnQVbP_k!ijhauEz;km#zNqP8g&O2f+=M^*+qL+^FW!tSJe)8 z+OgTWs;Lil#+`6f6vfNh9G1hKSM=ir{Jtp!S=SpL?2~<@R}pV1rl<44gMhN9*8rH1 zRq4b<;k~;i@LnA>oI&`jf9ggIqq-FJbs@~K(o0wYo+wtX-3j>8GZF_GM|Im=Tk;xJ z;8^gVFNV1n(s_EBhi-Lk55Nmr-8%LF*6baHS|O~oFPe~ARG6TP5`AaEolqUcfplu{ z#a^aJx*XY&krxZjkBp1S$B}gMN!Kb}YnnB?%PA{%0dzwVKoIunz)=c%Zt0|1eQlmq zb&}K~*~dXtzX6Bx5%7?s4`P_#Mr{YrKc5YHoqgX*=qw|>ZzZ<$u@&g)U+47mR^q__ z`kKa8qNjgfCHx+rvRJ+NhVQZzglDNGJhW7Io;?82gwodPSH4-Y?BI!^rlcftWNHU( zo$$L8DyQ=6$N-6$rn0cEwDYu`k}|2nK0keI3`c#ng`HBZu{^tMu=soB zvZwB0dYP+~kQbo{+%c+k=UxZ*#a=TO!A73`%7pW3{t*7S5pe6G| zbFyr$cmkmqiJ+5wxiLCUNUG6{VlgMz0+U=ATbzb$)I=hVV$P=$fhV>${LaYhh46f5Z;V_NnYw(K8>0u6tuEa?359nu z2ZH6n$zjL|{+sfB$fqvd8F#|3gPNVtpg-(Yg6I+Um;N0?82ut!V%6*PJo!u(8 zQ&q%J@bp)VTze6#VwZZT0oe-3BC2|iD2k4wcSUGu>Lu%{c*Q1j@+u>C91mc3M&ObH z?^G``IwC=5YcwM+5JI`t2yi#(?wZk+*3q}IbnV#G)sJJzHk&wwhp}`Xv?`+~5=Z;E zCtRhLZ6#m6W(q!mBlNXs%c?HZL&SdyUdI!$W$-$WzE{u5+GfK+SXYV?W6j z+OR+$`iZ0H6$yA&UU>MtPhC6J6%Xg!0v`U=34loY@=>aWOp0&2mTJOAZ&E43g4q2d zrgTT_{+MYXc7qb{|MB0>yb!x>dqSP_8@%-Z5PwJPZn;*(Zb>*IcCYCN#O@P$LF|63 zX4HE??0!%GlamW?-2jCC+^AQsJJ|V#*}dfb8MFJRj0&^+Z)g73lNp~R@MTuWY?7;J zpcXgnEx8bM(M_Q4S0RKu%}XJ#WRAK#h<#frwqevnRg8Qpxi-hjEx*j@+V<*Oet?kE zg6sVnNAEL-Vk4b3#IwqD#wD)W4*b22;3wd3miAZVkOBPtgG*fC3?aJ(!1}d=sj2ys zyrCsO>vaxcf&3opIDVE@Hs(SXv-G}<$J!1rU-_0*9J37ol1-2nlZY~QNxU~PrL5C~ zEcz}h2|bb6gZ(HWAzWn(dyo*zam7fV`vsn?r8jV9fC&1C_JL?mQ$xSv(iyrj7sn6| zkk%iz%t7Vecg!kBPwuh2_vBY7?iK4*C!crAn!6jEuQ=X4`THi@cjvGk-wN{%KityO z6I;0p;xX6UWGj~U)ryVqpmky=3mXS8gAEAH7^`M0uI(5uwZ&Xy8FY$Gc8)wbzp*85 z=tA84*d`!r;?HQtEgue-Zbvh2=Wxk*f_Ux%)ynSIybCnr7x2D_4|`*C&S1F4OFKlo zz~e#17OdNm0F#{Lk)DZwZq=L9}NH$F#-XCgRODrre$R4C3?8W9kiY$mJtfW6C{7W7!1q4M;Afi$LD{%G*Z8 z*4Ls7TUSdv+m=3epHpECm{g^(Ji7E>Pwr|mK^xxCQAvI+MSC*unTyIz%ud-ldYxYI zQ;I5GRBr?NU$Ivoby9fXJY~nvwJv%=hAX^7ytWVD@gB*m^vnz6TvauhfDp0VsuAT= zWi3~1g$OxzW1F>!^_278XMyL%LAJ$4rQ^zucHXLSM3>-cGJWg^z7+L)==)T_G(fE z5iei5oEtTYvC%y z)M}m8@-5ng^D1A;-T=JGxlX$@K!z;Ev~<_5F81QK^pm=nSiV2ls8cb#VFF{j@}j0L zkicBqcZzm#X50mW^fKN@o^S4 z-rTEjj&>kqvH#YIFmxawalvaskE5;$FD>{kb<~7kKa#ZS98l(a@X$l1&R)hrf|d2C z`1<3E@0fM=`QjTYPrDZCQ+@tLWJ4}tM*1s%=5|=NQhPdfb7Wd=<XLelsR6jDh>j|x36#K;8C^%=5Xhb-*A-E94U$RjahS~0-9gZ9MU?XR%wD5 z1q!c0+h6^^E;1 zhg4-9TMO^v@Tdq6FeMH#;IoSdjtU?~(j2KhbI|?m2zLTflj7hBD$$!12Ql1aDk~1L z!o)H+OHsd{?`VFb{mTeZw#x2RE|<^1OLBXa=Z?CM*8j6}k- zhZ-a9{i$9VQv`F-XvUPm_)k|xe0_aqBG5Qu{Yf&#dqs*+B4>|%(Ci~y3j_{zgg;hs%!chEsi92DKb?ce{Xq*_U%bM^C z*E)B^l6~xmU)lQ%>DM+ReW*f?$_@;8g}ZX%z$M&Z94kV>(b7-C??> zDm&%q?VsW9o2#1W_n+3pwYz%uZB0C6>FHxl+|&D-*3-Yv>Blwk#Qz$qu_pHP@2iC0 zV|ck*WCy#+(&A%QIBdz!C#u)P)No{cNPI`HjI(I@9xOp-XW73yA!cHgFGG-=eVqY9 zxAL6;(xw=nPbgj)ofv|J{n^gwc{5Y@8p~kxvW2ZNscx{IqwQ=y_k#z0o5Ka&4=<1{ zHC{FZ`C5(y*}oX?GyxD49&xD#;vO1ezV=KMea9+;Com7Ayw?C&xOLmlvR zg@tA3;LRIY&6)$ZbS6M2d<{a$I0XWwNMz-^9*|uhC@MYUst05arM2+!af(9c*_M(s z60D*o%>fo>n*uwB^r^^IH3#RAE?UhYDm8F_s+!DOkR1O;lVu=%cE=`{9G?|j)e1V0 z3rbERf{R&^H39FT;+Mhv{mhFw$ND^yjwX2OtFl9mJ_p^LgOVyDU42>iDciH}#(>V! z(}yvzr}xFYr+=N(&&I%k|Mj^HW1y#hUnTq=pBT5TygS^;o6Bg@JL;cx04A?yw6)Zy&=Au*TW{NzcdH^OO2jelv<2tbTWm+7`)av z6xhzV6XyE5=r)i>O-es*^xWL&y|Q<8F*bTV{UsD(-dn)cEuGML#f7R0;v`J0ivx!T zh=={a&Vgf7JF8M0oWHGv&H>iBVbFID>5K>py5dNn!=$1)IB{+k-_DUU$MsfoNcW1a zJ4i(^dT+H<(l*I|Bw!PirH3y|R>95slq<@{$&#T>m|2290YB$oquM&Qb3IE#q?y~f)FPju!FUax9&8=_G zIEx-z-%ijq9Ery7Rq6uL*OZUDT?q48)82aU<); zLff;C^vtn?Q|UO;$wFu~!HN~a(ZV2%Z{9fquq+ZuKb1N@EBSPiSm2t`S6L&S+ZY#} zAYUx+ga+eRCpcoo83=4zFGnm5r-xW-VPssCX5v9xG+HXUJTb{G&B@PzZyf|7a#!s& zp~e<_L046IBbd##17vY%aYy$K;CRDvSO2L4`cBbcFL%*|BR zqV`W4e5t^8Nj?f$YcB_F4S-CP7T@nH}$PtD^Q|0=_alrqW zKbyE7Sj~7OWTL4AP2gCaII&8G{ub^g!_b;L<4!oL)7YRD4<4Rxdr!`+fZ!eZhw2cO zA8|PV?@y-a*$J{`#HYWek#V&Yc@K)g&

o6B^ovorIxVWkm+=RdI;r0_L^mkd-ml zkTeI6n*w96ID+gq&7tx@bZM6~Ap-Gg4@Ge}=Qw!T3GGGJTDe*Z`}oM*1$Q|Adq{>tB6U)e_W z^zY6g-#{gN(0G z24d1^f+)F6jJ84e{Sf8=^p4{JCB4-vQ#tw&%|{L#kcs5W5WRnXJaR}`XP=K8gj#4V zG0k4or;^+bw=-d_0fDjUVzB2HaHBeHALFPOuF z!=Uq&q1alLRAQaTlp;FwSt`6^Y>UpkGm6!ka=H$??B3aR7S%zHP%xN1ijb<|o>mix zH5yKY3Oe>-bb#>_^`$3@EQY*uf1l;{^{iCr90k?lbwA+_6~OPkG}q$;=d+fP&c%B{ z6J#?*ux4FVHx6!9brqx-cfvFd7VU%^ZM+BF{MVDc0RR%G@s+)&bAS{M>9uXW-AwOV zMYa~Xx}{p#vxGTii8Jg7ZoBb4D3CQN#{N-&$I)Us;~nBIQ7T);WS*GLuZ;7fu(F!) z8dg`{pdGTmO3jzg^R2G?pQMU>S|RJg8zdVcGCvg_d0XbZ@o<(nlf@Xo4)6pv)*ju!YNLO^8pVAuXJI z^vwzCCR}R%etMRJYeF=VD`UpK_J>yG1Q^>vM$xTU*t#m&oHdk!OI2d{CyU4QUSfMf z%W2tOBKLuARUCVXm)TQ4vdnf zc7v-+_Fji*!phUrQ4BafeZMF+T?X2-Otx1#c#>`5+|r5oYS_!~^4g(mB?=R!MFieUlx}mJ4%|eM(XnQsPeK;UoR6*(hITIaYgq_% z*JWSZtUj?tnBcFBivq=pb17uL*9&b3KhJ=%*(2VyFLxO9H^iw$u zi8mhS8(lcvjRW?_!v8O4Z?ZF2jw|ZU>n%3$4{RV4P~YUM6!zE&cy>n#9GrK*{1i(~ zh?2v3$_1xmk(8*8q9itin#=xDIFaS7ZnBpp@5$n~_f@Z4mOMdtKUZ#f;7aLKxecUw z(PE9ndCkA$(%&MiBe&C}#HYn$;Kc~2jD|m4J-W*&rwkIss#T7xJt^_{D3bk!~l$Xc$O!-H{cx4-A+h%l7Rz&K4v^nzHOA{bxGj_wZNwaY)Z7R zS@>(NxNkP4p18t|q*}db@Mp1!jbmrGV%x+e(Rn62!q55|HM@ju38sE#x0ZIUKd?LH z(#Dkn90gV+6Iwky6TJ&8_LDNvdjoJyZF4$eHhIbAoLZoAqJ|H=RrQN_K!#cLxC9U} zQ&@w8MUr@~4>wK@!9^2tnc=ykdLx{21%o}je)h)L*LE8|wuS^fhRCCb<`79a$>!5}t|)uK4Y0SkVPT)CCoO#u&nd#0`0>DBUGhphj2B)q)K9l9ok9e zTTHQT?W`g;oYgZElRs|aZtWC(KnA7g7253d$n$Wi5nnaZ#i$M%27DUOMb=wYMFWnq zOy+&CW_?t8+zQ`h8SjOpSm@~ed%<*-JGK7($I!Y+H?`qsl8q8ASK*TYnF~}!Ae|#p zoX71Qz;hpFxvD7QXA`zmQDCra5U4vKNCoYJeg~itx#)Erk?g?Xz*X`@Bu|ezEgiTbM|NF{`gQ^QC4W^T`&pFUB1X+JY1=rv z)@50&8csp(EKI3o)|rZ_lCAw!+X11k!a7%>)N5^6-WYl!n-kJZ_>z_=QL2nr{RfPv z{`_+3Gme?h6_%(DEgU(yRsWDRB%8n=Nro=@JAPs|5ajs{{qGz4%wd!275sGpYr6Ly z`F^MT{)wf39&lxbp`3Nf9~}^JPbm{HC@RUlG=(#G66f;H{t57xs9uFTl4GC$F8bd$ z^p`W?NKw?>DpN>q=>s_vk?!Kgz(6PFyf*_qG~e4lEVuHxcZl=X4t-alyrL zah62T%+yQz`Jl_8gq1YW<(Gf*0|O^<4i}vrJ_tJtJ0aok{#IJ}+v{X&ZP4h;tPg)>P6Z@6e4wv> zBsEK@c(M(umTk=rvVtF&pJdPX!bJ^fb@%T?Um{EU_Fa?(qF}1;C{(!PD!#}b)M8j% z7?V@3^a7bkcfyvbFO@4HPhHtW%?d&O*+CURw4#jroT|#Bj(WdU2 z0lmaTNPSh}WIhqxXn_wxTz0DqS3*n{NpGT9!HKTK1CMeB&-MTi=_8-lsq}3C2u0j_ zHW5MNsKz%$621635T7l$fg&ENx@Vl|h}DH_&xjzwzgbTt0yKC$qHlz9P%{`hs&*3b zh4_&g&~p49dM1$Gr|8w<0-0Vz)y##`QX?zozAo@lBFTQF%@go-1N5p0)PP_2tskc;WvnM%?z_7u`^rPJ7%b+dS6zu}&{LTFVds?Fy zBrIyX3sxlti-CQ*mMDE)WIbxa*lgtPtYmBm4}(gc$d=yZ)2QNbCL7FYSS?1UZAaL- zl4<1d-QIil>jGakpM>AD8bS>i6a?)HdT?LXgCV~2P}8z&wdZC<=!gXp+_g(_XiJ2e z+*xs4i%OXX0Es589(GyrM8>=hyUZf#@M1KfEL0+B+@>z~pusr3;)s$cnyJ+#l77sF zy~Kr3BpcCHG$A3S-TNvkWo}G_;|yL8kgJ9aef3sjJ9=R+v=g#Tc%HQ`jVCHe%hIPrUwa{2g{U*$C^w&mkF%GSeJ_`tQ*eJrNXSA-kj zxEshBEtum=#;|Q{yMZhp)Tal1NcrZQ5!H9{|L{aLzCS!+bVCZxS*^n|qc7|oJY>1- zW*9(??_ZbiU6xlV3Hk6=O`gY!+(fUU<06))b;aK}R>TzT0-uC(s%X0YrN$OV%Ic1? zJ`oe#Yd~-$06GCg5hG*J0CXWP8yoS){~-OGDVKZlwoRDSuC6=cBE+#&c5AI;Ye9K^ zn=tP^lX!SedML2UDde`k^-Lybjn!k0OExC(8c;BgNlAA=sj_Vr^SJ|L^Qh$iIdnj+ zRm4J8?clPXTzI#stKf5MNtx`CYVvBppePyN7|jDN7jhtZ=OO#K8C7Dw=s|F=q?~%) z%~Os@1Al0>lVf=;;Wm1lszfpv-G;!z^ih{6lTW6Ez4qe{Up6Zwx!3)u0An*E9-}c= z=81R}Pu*2op#FPzRa~?VQ_~ad!5?OioLQu!%(Wjq;;1h0Nl4iHxVK@jg29fjQ4hc? z*0VqGIQElm`80x-(E7vjk!>)_F|r<(&)jqNcI<%TW`s+#dwYsHZbtpG>OmiuRgwH$ zX(`6Yz7^-CRH$u6cC!AR3}r~e?K@F#a5o#@yOr(13ArGQ>X6OJE+k zWdHNJuSv{GUYaW*&&KuCxF{mi6niC1L3W-6J_zNAM`ouxVTrN5c)K|I2K;>OLke)*(^5+fc5)sYFV*a6k3T|3pARsiJcqIrV~A5X_AZ&7ZEQEP zO12zi#F>nYV+3#9$uPnwhVFi&%gZJcNTN3hyI`}`i6+wcx3X)TJx!Y2=aROBfw%}Z zvS`?)%jQE$O{VPy=hn33-QHPED0P%n>y?mfip3O7NMwt4&btw!qUOob1XPT;od%HT zy}CosRe}_R-{`rT>PtmbJx9jI76D~%2e(u0G5a!}ID{2!YZM{=v@u^NP>Ta&P%3sp zkG+N4Yh2zsM8OZ{KqyM3;QfNI$c~tU>Ogj`mJa(<2TF@FfnrV_7m( zM&U9%IFTI+PnE>shtbJo{|oogDLA@CrwL_GmIIHu_EqHoQxAPrD=+4s`$|MS$Nf`Z zc^+jkyV|*u{am!AbR9u2hjWq}skK3{A&Wh)2C&K&aWpU7$POLj=XUsP4DnD8?HJN~ zZ^sTeiuFt~H+OI|dwWG`3AB-R4|?BI)c1a$MuKdP1^xRt+DMP5`CXVyJw9yPCFHn` zSk7D!q_aB`cp@h-D(`s9qYQCxxk=k9E?jM{?PYzh5A@lj_U^jTSuB*>72HA?>{O8n zUP1(cy$5hX%K>?~1F*q0wd$%woN#ZYsv7rVyxh3P3aOE?J3Z7S8SpNeYtM;x#*s!7Rvcj8C&@b@twzXgZE%DN zAT5*zL@P(&tO11hNJfvIY$&;4|A&Ix|S|JIduUy11< z--WAmeMQ26-k=UYcvDPO{E@gB>m`r(5K)rQ2YQ3Z0_VrQF=4*s_+=CwY12!}JK=Ka zls%CX4urLNdt7TIq^xabe{0_~@z1;I0pN>hBu=*t@!#^b-l_@&)b&XdiXvJ}s#n5T zgpxO#ofVRZxA&91#B8dFiF?K1y5E9xdmn$ngd8D*Yv37cpgG#gq03kmQ%%kxW z#^3-*Q8X8)n3L0mbfJ%mxvvX+%6yEnmPs%u6ky0~LuTQkc|Z{}`$Ww1MxMPkxvB{x zVG6Q&Ck%Ef_)WNnkBd$=2f|B~!>W@ZC-#J%nvU=8zV1XmIARNH!XoO2uT~L$8}J(2 zDe4o~tZV<8I6K!QDX>*f9b6Q1F?yC_#BXpk-urSW_!!{_z0%}-Mi%PpN+U=`&Pr0Q zZxx%UoNpB$I$*w4s1BH0z|QUtIPMKqP*XfToWJc2{jK6bf7ymNh2A&2%Km*`=om73 z?B9EgHub$mtjN@N!bXI&Ry_;P%4vM z|9HSZF5t%D_sPswocZHVyp^*l)1PVkF1np>w-o>Tn@xJKkbKW&MgA_1Mo4P1EE z1+{sDxXJGoF_P9M{y|pr@b8fSc_E8R$b9d}y(1jS27Ux#RPzFHeW#G&OFXZkYe zxU%fBbjWQpZR;T^?8Sv8{e|Tc7M>4qd={M>Q;q)l=krYBk_eSJ8QNh$Prkw#9v%n6ql~uCHXp`2moXwrc8kduF-%iD?n&UqD%mVqdNw~Q4 zV0lB6X|IaIM~1NT=WeiEX<|ULlMyOJrs9pPl;+{SbKy)4(iv5KCrlYH+FLcQMO!oG zy^qQuBf~uAJDu!X|4x`~Lt}G@>GC!WULcWo?Rb65NX-qOGoxaZOF`j5)hK&gmea_` z8bt+d=WJbLJH;j41wlUJc75(?7kKj(b#Z4XuuF#Wk)ePywsi-FqL6!=RLPMXhu(ZF zStLs4=yNEUbPrk0Q4vBO8$#6u1jsBlU8X#dcRb-RH!@<@RL&gdp7VyUHSg`iySSN@ zerO+F(;16-)xIGdcZRm_^H4FxAYsSq>I8gn&B%X|Op))FNu=|4;+@g!#I!4Y07~Q3gH#^2Q%N3ETdyS5( z{b&@24i6bQ{O$OpHVnBad>PK>X1?PG9CwrM0a5lUnM!v2Tyayi3coPsGJrcinFI+O zz2l44NOa5{A3Ge3CU<<1T9v4B3cp<|9q{G6(1OnI_`U^(4+(QgSm4#=-Yl)n<^HF8n+?+?DrEK367 z@%?+JAX+n_mEIrx4u7u5+YHec&6k_H@ zt4qY=h)Axt5;0~YPs&~*H&q<@Rf(ihMkl)UipW0^mr+j^%n2eIjw&&UdWG6};c?9Q zolpVh?zGH7oWz7>Y_x@U4yMlxf31h7l3WQGJJI`8F2rB1%lF=KhLnutdhYz#976KLdB70vR(8eL(2^r3icPi2v&3HwS9ZOBj;+X4QI9Uou}Z*B;;0V;5{q5L8>84`Y8GNZcTf))eR)Y zL&3cKomD>wMREpmrYb^QQ8?dR+KA0?Yzup9qymTnL*BVOPR{eH2y{8%xA;w?oF5%P z`Ye}pTXQZ5Xibqh?Np;Fcpb|JfT#)Fk=X;ld;AMk9gbLsCLM}m=QJ)%!1oCfUEC~P z+56au79GtxY3buS2FDSD%C9Z$gglbAYq#^@$YP4l1wA+f?BT)w;}jG{gY==NIoKk9 z6XQ&Am0n(PSFUjOBD%RE%mjn>`!byUP(oKx zxYn2K+sKG_n{3_4+z>UMXKts+nmcInx7sWo6SCSSU6%`{?Rv;0p} zBsS!HeB%-Gj^On8+7g*p*8z5hU}ngqt;*YePTXam>`>=VUveQ4KO2SCIY5f1itaAzTv>|H+Oq{p5LOIlUpkXn0* zJK!kJQZ{n#JY4R+i8IU+ESv0u-YZWy6`*7F?@!$=2YpTJ~GC= zlf}{0f~}L&WH@RP7X4CNB@s%fimWD4*h;>U$p(;HK=DC0q@5f#2q3~ZZEDDSsW-Av$$CD6ib#=07aL_d6?@DRiNdz4oisl z5{LA#&3MWhO(;SWA1GMeYo={t{EWVlz;pq;gw z?~4L(!6@@A9{`RQ72RBC$R?NK+dvu@r+Aji%mQB&O-IP(TR-uPXfRpd=PK`rm@|Ex zta4Nq0bS!^whDxIJv}Ws17mq}>r(6zE0Jqe2gQV+Z#@7?D!61RKw`B0VGwwA!Kzd- zM-0#yuCv?`x3DbmDbPKBBAp!qgmfxA0ZWQ zBkTwm6|1g_AZRj{x+JWr8TsyehD`vo$9X?6)~P>b8x zoyXrmp?M!E0AzCY2LKoMvI**{NchLA(XD4viY@T5Dv>Cq1fEkR@?sm@eUN0)1X5${ zB@#DA#Je5@IYN^TGpVYGCsN3Re5s0@Of7JPDzRm6_WKfO;t1=Jtg8y7f)D$U%&J;$ zp{%DK<`5+=t7;od;zPHKTg6Jd(K`VOZqr#MUBW z=%oN8_l;Q(Negz0FWIH4A;%Qq4mb)pQoHB;OYaHLNrFtzJ?MSW%33q*q3&ZlJ&7$%IJGm91*RUFBO?|tx4ChZ5)Yx_K#0p&+{b#6ZyZ;W)r6IK z72>T%XMDR@SA-7Ba}3q5gqZKHLpB6wh^`5x9l&XsI7AvyZR=ab18%!GZ>*zl>~b%d zc(B#2CUQ|M=H6-}0MUW%6BE%^v z2dkMMw|pY)!fRpwa)vg&kFky@Mg18j?D{bFIn~(0$+bDS;n4g&O^y0oyWall>DR1W`bH) zRih|GEl%{u8s*H)h0dYI9Bo?^dyS+TbXN5$5&ZBhRRrjgF9lXnWaHZm-2pAD1CwAc z$`YU}cump5d7|qofQ$&?T@=6*(lUBohYws}94JCdqLnm7Akeoxv?G!A&Q?v4NxEV6 z)U=tpmhcCCMTEc0YG3LyA<{Fwcw}l=mMgO3SGiuQglE#$QF+kSKX3J z{Wv3aaz7!VWwAcQx+ztJ?aPGDY4n-u31}{)Wch>ZZVcehL+T!@mN4-8h`PTL_XcK` z|LP(6&SsI_!luU>p_3KNr+u1a4@1OI!mO0S^WJ$RF$LYgnO`Hdl& z)0xMQ3q!A9W=@e+?wEYo$XqA0hV-p2D!TT}Cx^9LeYl#**~Mr=nepUpzY_Y??wVEC zi9Q!vPWn%MC;D8G1sV^XN3`nZ;iR~dEv`i$+`0O?z(-E4*0qOMpebXcFju% zG{0?h=$$-aSP2a z+jdLwUA>!y*6p!+i4dYYsD4GXUO+*<0^S_$R9dPdM6Mi z2*Bf#bmod|a?aR+xNwy+X(EE=D%ECVaa%PCK1w7AL@G&_nC19b-o$^xzhe-JZc@ED zuu{B{^_e!pIoWEnG52bnYQUg~%q+d!T#I*}-)BgvqbhHSe!Pml4j7x}R-yt-U47XQ zMU^umhkmK6`@fjgfqmdX9~%}?Cs8U#F`|^#*m=vTWtN&`&-@W#WU-ul{BH4z-QhKQ0Z($#v^5^GXyCf%Q!B6}XW zgzcd%gpy4>fQ-UB0PAMsqGp%LMWF8+yI#J_mZu3x)Pp_L-alU#_(Z*BP#j&fHi~O- zcV}=5KDfKPySo$If)DQQ8f0*n;0_58g1cLA|9HX|>?)6={5T92)@UEcE? z4#>Pv>uoQh9&MaG8I9>4c`Dr30~H(|WLwKm@S`B!JC!gzJj_C|0G?tT15&u-Cwbx{ zLV>tuO|4EXiYK&eQ@H7HSx6n&wyiev6vUb8NpG^&LYii>=g~z4TtaB$fm)14kyEfiD40 zRNmQ1L-r;aViAl+jj`O`Ky@(g)<5{UWJ#G&rPKz!Ns>fs1Win9E%~sI{o7f<2ZdMe zD(-st1HTQa9b&}qCdTa z^4=kaZ7lcUeug2Cms(LPd_E_9dn2u()CTV{wG+0BvIKG^S*(*7BBFf*UWDtB4u9H! zM$w+3ZKZrG)`xW($I%!cxxXd2D>LQa{T(s?UkWci6w^V_#gob8GycxT5bm}&*h;s? zy#N8Gf}vUwv<&gmI<$#+j`i}m@Ur(`E6wCyNuQY(^9np`}k4++L4V-GTK50tC-(+14cb-oh*LgCtExEn*GO?3`PEDNC?l09* zPqi^r%RC{b_dDU*oxf@h;!4EnSczit=1hrtQ78R`=^h2tGw&(cAUCl^4#JMq``Ys; z{5|cu-ot>yrTYu5wmo6&Yj^2iQA7!*iWn^b-WbzcVlv@IOvU#Uzk@}SB-_aj(7m5P zkA=>8cn9S~`ZG55h?1yD^1i2c?60k`46>?NnA2oI(W|^zOV#UuGe0@BH`GqNkeLa4NS>>yrQRLm2b) zuQ^GuD!w%9U(~ZbphgDP$zqQ+D+7U-ZQPGcjr+s~dT)x5LUM?ql{di+a?$mW$-}T0Uvi{ z+?B(e)=l~MYyZ)Im^an$0cU_>x%S##Z37s`tj$GL_vX}KbP4}=*MXxRY$?`gP?V^w zZ&2>bODHEL4HQXP4TsSk8QZtFCswAAuRoHFZiFHZeuT)N1X++FfXCmA&JTV9pmYWQ zY+XJEZC+)-54gYv8TMyi^eFdk()R_K9T&i_Q6`SV;HQ!#BCIf&*+T?+KxDZ{sQ$)I z5?KPiFcxOH^vyVq(x5oqPZN~y-dE-2+4NpdG>7AUy@Ib0(}c#;59cSmDQIgkCz5WHPUnGK}n}Uo_$EO3nhBXHNT{ z+)V&am;i3L#U@I~2#(3@;4jdYP=ObPKl!!G@C^ADbs_d+_=4}7eNh7@#}dNkVgU$ z3KBcW%Ef}33L&Th*HGb$Q{d+U{!HNr9T@eUaU@D04SkJ^c)+>*2j*Lq&zi=CV?KwE z+iz=^yrjtPSMht>KA1uroEWONQ#o*SL-jl!qSl8Ts^79eVSTh7qD_ofpGGK!*|CKi zNS;y&hd?(fdgq@~I*>p~l&#{Z)?r=LYeqjHo$1%M#8^$ZFnVCu!1NKVx`gM`?8gI= zn7kBhWY4&W0^M5io)%xfiQtu3-_pxr*k;8nVg=sQjHq*oGkA4|`~3)lgKklGrgW^B zw`qaDaQUob{fDDY&P9+`1j9oHg_-wnjF(ObY5;ygm=%C?nA*-uEEXS&6DYy1Smtj9 zAl%vS!~#;E{Y=(dlpDg%rN<#RUBj+%64_ayBved-c4C&74kFu)7xjvI;%{nn9;9?B zCx zl=?haLLk}4CUWdJSaE8@{d3QxV<$Q>s(vPfFqrRt~{;eee{khN>6Hr8GuU zrKOX8O?7?K=| zo}v^@t6MB#?)(Np{+g_Q-UKOp_qfC=YBilY4txc(CwDS)YZg+`VuBAnR`4}Df`lF? zM=D*uB^}1Ir>8f%x4I&@sp~J;FEYZsb1)VzGl|bbeG4Ei)!<{>|3z-uU2QPe65~PP zR}I(9=D~ghx78*|jV~!`ve^v9FG<6<-T9?cgiT`F(-*3I%4yu)LCbx}6xI;$dz)l1 ziqQ>{c3K2eOgB%b1XKY3ZavM&U{k>F@Ms-2?NJF&t*q_Hr&qp52bKy-n3*Y2_b)|jcWQ_b8ou8)u`y%Z!xa12hTh!>q^@S1O zNRU=knvCPKHRWL)PV~3;;CjX-Z3;WH^-lf~>lcW&X5tj+#Eo$=7(4l$N1u5ZE*Bs* zZg>*gsX+XhT`u=tYvvrcB*Oqvf$vDsEB8ER-x3#~w}%9^fTT;I{0GUPIKv<7i?d|%@>;Ch38#Lub?b>Y=ho)suw$}F8<#! zw{ixN{#0KVxTFTeZLIzhj_v^M{s;u{QA~R$cecw^{oh_u)W*}AzQ+fJ9h zo?1CAp`TzQ=N+^5de=?)_#(w-EGP5g0r|OAd`5BEsoD4Tq!Rz%cA?Nsyz|z*`~;V% z>tNPP<~N|E^)CIl75*iw28G(doJQF0aUK#QCx;UlH zfEl@NNsO>Lynn#@UOsaU>>D|2jah@Ng@ydmlr~>bj2p3jchjvvVI8dJnYrjREXRkF zDC_50)DM%uwnDu7MsmKHm9w_c2I$6Zl{w@Wjl?gyGVk6mM*~8*)W8U|TE;pc%aF^#5B5~Uap?E{pr+JC= zW&FZ-v1F|1qoeb&(A#>`V&69 zV|vA~hfyB`D~?Orp#!}MSOqN_e0)xe#5?Lq88_{n{L2fnsHpRpj&rYEpEVuBZyZP+UK4zNW4*(^9 zWKFo20cdjj#1~c_KDG*yJ-OeFYTo;0%kQrXju{QGba(@y@5++CFZb?rvbb9tR2F zaHP|g_)^X>d0)iwGwOLEghIWMjieG`fA~IAm>HEV0u3-V3*Vqub<`pZGxH}Wa{%cKpC%UcJ z=}2>O%r`T0<7{1^*$>$6;KH|r*wNKJZkYi~8Ed^`JR8aJSSRl134ZOS(&O-@!OZy= zzOhkGG%zdwIR++lBN=*z?j|oX$ml5gIEv$=u#YyuKVdL4YuMv8SN(=;rOzBNbrboV z%x;urg3(XELS+s$lo^$T0ZjQAT)gyT!gfhH;|cVybmDg?R|coD-wUl0!8heC@r z1&A^vGz`1v^3$brxK$FF0p82sWlvlH;;(t@FXkH^tvtGLU)5?fPl=TDSbL-x(hZXX ze5}}Sq7xqCEX5Z?DXUaroM$;*P@HKkWd?CDJ1ee&Erg;W!aL3$% zM`*Ply^?amm<)adN1#`X?@Jud#ujdAyV!eddJa$B8u=auRc+5YVqRL9i$~jmf@?Jb zk4TWee+rmutTV1p4ik}oV--K#`_>44pnCP~oR0Q(be+~sZP-{M$+ROvb{nXhj~Pw< z!>X!|l_X^hd6Ojq^Wkv0%v`Sj-Ibh0J~DHsLt{AAqsd?SdI{g{lDezEVvgKgL;+re zQc7eAqPYGBzXXzlCEQeN3RXA1a+1-V5(`h)^h@1x3>zkwK!kk_roIm^8*6CC7~$_= zg`Fl3o6pJa6u2mK8~Xno=noJG6DHprDK%8*1P=T18d?{VoqkYK=*2`L#tw}w(6kYf zyzAP2ijFriEi>)oiSy)dtFPVQ#!4f2A;yl|Y6l<_>B)h4PFz-RNp#3O2d)_RO3F~n zJAGNF)GlQN#-y37kHRj84bauO@9e-oHUCC-#{Kt`bnG~Mn-#8b)tir;1e9D`KPk$T z+*l@%e`h8^xkI6~TU8|QTx1>26*j}oi@G-`dP}p(hcS{z*y5`y<0RCD#p2`MzDiSa z1P5NzHBOGq@OPID4|q*-I0x|c%a0)rnk3l)mMeM}gkIb{7FBKsHGBhk>T;-i*|3I5 z8#5T;fXr*W&KjQ3V6#;WJObt>0d_LY?Y%-zi|cotc8`=?4CRQDnBt!5-Z8p3@JCH= z{}lgWkN2U^c!=Sh?3&6QZ41%m~lBmO$M};VjCf1-kI2vBY9!G44R`$PR9&dfz@5>$I90Nf6Qp_3K;l6Y#zdYeCzbA1KAZP+>GVAW32c$!x$IuOSXeR6yk1v0RB+VeU4ERu z6yP&Q=l?oq2rN-W!)1lRmpQ3R@iiOQE^-yeE^A?_JntD*1l337qRN153o05mrQZ)D za=#im_fkEu?1xG;+|NpaY|mvL`~#7a()9kRjoqx6LE9!XMC?XTiLQVW^33E^w}Bhj zyqXS=yxT22Nrx|ALA(T+ynVu^Rq@B+=8yc_51zTB2n>?+Q5Qdl9s4vHp)6yd2GP!XPA+#Xn$9=vFqLbn0SMT` zUiVD%VZJ5jB;rBWav^m@~cyYEA>S$IJD93plQFIFSP`mx15AKlQ6;edihN+ z>UW;dwlB2I8KG^s@zF{PQ(9NXM6*i4HMs|&Uy&aamSbZ~h0$g6fLYLiy3Ub+Qb+yX zJ}3AkC9ogy-+7wjnYsG_!6>gRT?^0OS2ZS5HB_1V6Z^zHUo-FfHVT$iuKwE0kCR$Gi_1b{@)kT#RxB$_8l3OttR z_3i`{m5cc`L2($Ns@v{ndpj*bm(sGuK5a%ff3nGT#d@xw+Nc_Lf54BMNfh!Ma#&$- z2Qcy7RH{xD?Gd|1|8^W|7ESe892wj#TsBvgsoT2K`JL!n`+Jjb`M>EOLpp;yB zqD-=h0|Cwtydk7cP05%7G6Ig~BBn>~eET1;2qdA%m*b69P0{t?Fw^) zkT9@XDDdu2K(#^T*HTB9985g$t*A8ku3tg-H)hn#w1QkMxl-$*2cQITz|v8uL<+cs zQNUU;uOkKYg$N458jTId6sN-f%e9l|Tcd&VyEFWQ^v#qEc><=VuglT_wx=NhHqEZP zE6`@p^ypP|K;XB6U28&yD6O5VI+2M965khth@`dwZIZyIviqrJJ!p9yl}PH)sU z+F(l>a{`BTK99^6KQ#K_>_am8y5OJ1S%n|xo%=BN)3Uzm!Anp?u^ZiV8`@ZNK_TxB z1wzf02a>iKyVMPI!(lbNAOg4K8@Jx(zJKKnq9v?$pV$i|@^tXbWtCeOe{Yaqz-;Ns zG-pI}Ep@GM?sfmJ;lOgpL+&+9Xofq7Z~E)*lTqH+FXJPK3`I<$B?z@u#U0i)^h_6D zC3%q4WdtxhR9S|IqguPv+0X@+7-CEit~52^uUB8rgznL=<8#NR#POodG#Kz8tF_L! zIuKR-t@!!@-i2FrlHnEQbVk^Ye3UhCiHR{ULkTE}fk`)6Ku(b%T3gcFOpHQRCCzVB zdKM$-gOKF#?igWMS2rzR4j7KHRxum22=A)Tg6Ef2vSULVJcRrL@`Wj;p?gHEkNy6) z1`pi1MGuefmG7RqAst?H_B<;YG5;r@{w%}*fEu8|9s_+mpahW~MY{z3*ZgKH?Iklc zJ$n@jtNm{s&O5HL&uy>b1>yY-k>GA*3TbQ){ox&R|0RNPfi98wc;GyxB|?a@$}4~W z;|OA9Rm?NHBw0Nf=UZWAw8=i!)h2SG^jvS;y+m6$Ypj0%`4YGQ&51@%iyjNFG!w3g zS^|(J8L2Ers-hp+hZQK?a6Qw4#NBHOgbL0uZ@B((fOoLAv1I15fq~=E+GE)r;M|2# z6yCYuibe3lN-&eJgRQ^w_|ijJ!BwSHS#Y&A+Y_lDsNMDcgFgAE-fydni9X4Xm+wMc zDIZ?g_5ps)WHfO9GCPQ%q4NyZx{m~oC<|2`6)$l{p*0+C-UjwAsB-2DDA$bU(}9D9 zL!`6dsy=lVcUy3h>~r?b;CBaJr}2(kG3WAK(?VrZV*c2yU?CH1<6=RTy{*O?!eD14 zbjEU(p~Yc*agVU<^3JcbKuI;OH2Qh-#S#^&)D5|6d^SW*L&L#q4t`Ois#n8zYPG0a>Th-uO!{9F)uG3 z9kNM9%ONNOIG)sSf;r>MH8dsN`J>_A2G}+9F}Y+8e4JU)VcTKl z=<$S;@UZB2V~amkQf6K}H&!4(iGv4on|rqk*K?b75kI9}H|etlr~hi_Q6^?KlAa)B z&AZzjH9*ETmM4fG`rPh%8lvU>96&t{(zK_gDqd_#Thp9h6i<=8QUrq_AD{Gk-8~;* z2_l5&m5n?}(4Q9`%$Z}gSgXgz;cmh?KTUi8KeZ5Dwa%3BvDGPo>Op755w7^#p5#tO zC3FfCl>IE>HdFw>bn7(eH^yryPZ?%D-Un5hWvK72`>YYyU=Eq+@pfMmE)oZ%@jKgGd!j!<2Uz-0#qgHx<=;IH;;6fm&`CC+v0_{)u+MIW#Wo2rV2%TG9wB9gekE8&8 zB?y9TOS0$K!j%a0mUZRc5rL5i?ET`_(GVfXV(qrn-y~O#@Frgm^bN@7XC+`Pj4QSI z?XY)T=Bb1?!hnZeSLZnohjc-Z(I+8IC3ejZvO~Ia@%h2lC`NZH-$~>T2KDR3q%$>Y zDIU1?8vvzr!|MIgVW?}q*}*fvbPF8YyURMMq`?lQ!h;MG`Gns?`G+4g3*PXKl!wI& z*7XH=Qpe|Y_=QyE)e~jS!^lLbjYywsDr9_fM`9$e2m!i@hAT=KLo{s zmJ7T;5yWwSs*o4o$Q4Vk)z77UH-WF9(yuY=Vh)@^j(~>rUR%IzIPK8#v(!;k*A{9x zYJHF&;0ugKrKaax>X%Q|q{$!1dMI%#gjrx82aMG2&~m|-cC=@CF&=pnx5U{?H1$Dx zF1lAedk#*wpBX5PtzG=ETN;GsS~NW~z1t!b+0Cq99+W5;XR#)wRACmD3*|gSj2DUh zJC?S%NT#c_KA5~-l^;3PvZR~F=*u_+B*wC3KTEf7sfrktTwWXE0_EsIyiSGw<}H#n9z z22!;B{{_;du-JH6{A3ER&^Q}wz+3Ya0*8^ojP5`KleS=<`azXxcU-(gz>Wy(1$&V{ zdsjV7;DY*L9y~YWS^|_i#3VN{pQR2S_q^Dtfkf2;FJ+mVX2v<)XCtYakMQ|ZYw2k> zCUINz2z4c;bHxfyXx~T~(S}Jdb}$gaGmPE zGS8Lz_Sl%S@p?zwh}YkW2f@ur3+lZ0#Z&w;PwJy>U*MjVki^{hb$HE)DrxC&>yrJM z^)I{c1$A_5W7?o=y^@e-| z{4%<|W`(#2XjoT<>5$?Hqow=a|D-R0y%27pIM;esz;{ej@r@ZHJg5D-(z2q>AXvK; z6OzS~D{G{_Yz3o|k5#wd7AHrCBzeOAP<)%f}LH`@5ZWv@GhU zTU_LSkIb5nLBtqDc*Qgd`3a-Xy8XtUb<@BO;A4pCCFUA#BF2%#+a^?rxu!WSmMPYV zWs&1VnVGwQh)Q#?_|G!^>JJf647yI*sAJEW8Olrx`}a~WvdMiu-1f-mjZ*6^Qv&`$@9CMM!yeeo2ZOwetO<{TT}=Is!$} zA!zfG^29s&rF3Y?>hBfeuqFVRW0j6rL@H6m5c3Y>)BpZyq2R{eXOCp3v}ko{Lvswz ziOW8yw@U%;Y5QGQht3U-59He*(p%TbHDwm)IW)H$nP1LXKLr3tg0n>O3Tsg)Fj8ZS zhd?jp7R+Z`$#+*NC*tvkP+eL1rZ@kfD@)q)bbO@%MqU}4rhf5yf5#&jwnNUniBbDnGLP6Qv$8z=L3bbx|G1p?0pWp!$9{H| z4HY@R;6~(xZFGN(jwT6f<@TksGFX&SH_AIk0@bUj;Uyr{w-d{3F5P_%!8#t?(P9A; zGiR=-UiS-rCFu-`41!?8q)8A0-Qj)i#9<7q$%(dtHRSDlN6l0fp4`naPXx7vzkWkm z7l0_`em>4-hgNbK(IqmJD=#ebL9>!`TYLAX%UrI9m2JiQGWT!=X znJy`{K@TfftSSKjDo)?$*2^`m18z_6U=QOlg!g!MFkCDtCie{Py!yV+iD#^Z5Y@Si-byf+!l_9)Y6xSJ^Dbj{ijq{;DbSCRPO|e9(D71BF6R2iDZ9=D3-AB3u zQq8E!E$I1}lfOG|h#55<@e`m-@RM;x#nlwgIqbT5! zSObU?uyvbZO0*Hmy-9^tG&L$=TxzpJGbt8z6fx{+JZj3x>AJo~V$3iz`7KFeV6mhX z(~n7!rY~~RS$I=1fs4sOAXnL@Q+9Id`UW(FDmSVG; zkF2Ghl0rg0DQ}^xeP~U4EDL4&#dpYIArO4$neGs1^NJ8l>2OayytEL?gH7!UTXmvR ze=bsS2<$Xbg!+E%(-(db-EY`i<+p;Qt{P3EBSJfKBIcIwuVXA_L&tkVD{CHK&SPX5 zU!O>G5IMeXbnJ?Ww-B2TM_s4*4m3@^A1TiH3VAub!kk9vv`x8`r$-$tG2`R@_ip=l za0l81#S{(~+KAWal6^ew7Y>IZuOL9~XgD1fk5!2kI_)Ft{x|qNx{63Qd{7*Pe~dOM z;)BJYsn=qr>qG~}KhWV{lntpr_q@hlgB?6uTPE6Cqvfg?$&r|d&=%so;LvRk7SRm4 zZwXAMHIQqd7Kc$k&0^qlYF_opryrFjM#d@Tk{gHz0U5}s)|pIr-e^rt^tbqL3I+|@ z+F^8ahy^=$npr(>+m*i(56$RNj#wurjRHy(RWeFs#1Yc9^;-{`*CUgv(%{Hx$DW;g zWUm~9_WteOSrJMaz3Lfo8#_MVfjeCHk6!@ftKN_=6;?%FeVRz3Yw~%q#IRe)1B(y*Fj_?At%ji_i~7AeJx&754NGC(z7}0&0 z`<{s7#O!Pc(190j!5)>rX;1$Fw&=-#aQQPS#QUdyPhhVy^EJKqWe6PY;(UMsVibp? zXMzm0*}C-xxJG_@2>6>gC9b@48Q-MWTm7f(c?9rG#Gq4LLvC}%8mKKgvo&yzJf7fn zaivR{ul6)IfOX{yF-~UpL)g=Wx=aHl+nle|{Q8^z^N`pCbK06xCWmLoDmXcSBT+>i zYl3gAHL^oKp1CKNwKJi#hdULz5VM%d&U6PwnN3QW(OIH6#n#0dgg^*0+tnhc;f!Ul z1bwvfDT`V#D$Ypa*k;g=ysKw$Jrn-V+SAMqo6lDwDABsYO}nBT{pY>#!Ir0QcMtcC z)rutPR?lu7Ueo8e3!fuaLp=-+JE>8kiRo#r*ig{(;aBazd5@DtiE*~K4mi6O?q`)} zV$Uf@@iB}M1>DN={!9DvRDN)sA6J^np!^+kl2s-=S(ZpEjr!`5%Nnru;Y z*hl+Uk-R*xk%V6a)yBzir{18`k}sk6r*-t}L}VkKTlv(zqu~!qH27Pyj5MLyl4zzy zTPkA8r8e{_*?NL*gHZivSK%iDYU!o;j68Mjw=$HNh8S*DJgq*b;KnjHPi<6zWQoO` ztEJgRmW;7!L#0{YS-WYPDuN^3s{NL&AUy;JhW}fiinI3QUc&L_Xx%Oq!8A5xmMa_4BDTTiBMzJK`w#oBnBpkvkQzdA zelsV;eed8=XEt2J$Vy7mCTK~FQhlU;8;SDDM}k=DsU}4(yfkpUD>zk18`*$3nXcpSNC4&K|>> z5t2Mdaudt0BR8F1Z_2o?{s^8w1(1bArHpl_{$lk0HT@>Eh)bfT0mm1ll?*xnLSBm@ zQLt@XfAnrEODCe&ZPg>|{E9Z36hF$L8cX{30!V*9Hc^HO;h+bSx9tgAL467bqvgg; ztLY@x`|A`{vGbu$icnYLUFrl>EWpw%*dIL*QgmVLq*SQ`M4ht+C-D&Gr*E6H@Y=Xh z6h_OuVJjfqmZ@RtL%8&fuE_Jvtq@r_R!d(8^j z2*5H9e_lHpv!V9}p)7*>XG1CYB&9{93DkIWXMlG^bAf=*DmOy=YXd?zTu*Mf#BZ`r z&8LK{JpsoadK;n}DNjku#kkX+^2Y;Gi{`%N5I2B98?vB>UL5}sFVQ~o#~DpLw);hTp9pE8oza|yrGGyn6f zd}EN%QI9(sT~GhR^M|i#teQ4L&U&MYdj_?)8e9eX@m;id=l+*wxZdvY5G-d=w|OBB#H7?E!k-abE>AQev;twHkd5F6sU|I~r~Dw-b9c~<$*l;h6VO+~N_? z{+i0mdlVVs3Jv*;^}Dy^;d*N1>TKcx*YKV@cTm*k+hSOo48_{k-+d|8&UzPmKA&&Z zO;Nid*XrO!hrfI_0PMVGv%wFgqw=+%{ljUpynCed#7e<)q})H?B%~|HDt*Zhbffks z1oHiTLzHFh0w}#t<5fU%>|ZN^jBWynIJ#UStq}_MikfM}NR~{4E!7OhmIsxX9sI*k z3eZt7=~^+p3<%bJ_h;PHN_x3Iy_(@H$(xJoKB+VqyQ^a!Vs*%?p63zhjj{%0i*PH| zl3j>|USkx${|f0V2g-iWrsaU zGaMfQZ+CIb|J3F6&EcJ{R{#!imhkT$p5uw@O|g7sKib z4>k0X2SFRB{V;i96HK(pu|3`j19VX9Mc;>pcYhcoabtYB3u}w=*rmIPB)B&yciZISfuli}rLda~A4H)KXW|MfBW< z1*;IRCp0UI-LtA!qkDqJIwuXkv~Cr+Fk&D-rg~f)<@_jsLth;nKyJ!^uLMRADgz>s^z~dHN)f*&P7cA zR~gFY*<|L?GpUW5IdyG7kb6?Lb_EtbT>c?8Md=8tmr&vjc+hD%c8gkxA?*tS93m9; z0=5p3*$8=v`t_^)I4twn{cBY5*+w{5gDV*Z+(7TB^$!BM0ZGKb7V0a zefqkA$NL-aexDH;?II%!lHSDiw)l^BYvhLq26(^II|md4>!a4MSMI-@_r<@?BJ;&8 z{i)+H`pn44o~*Lh9<3L%Fm=%^0FqS968aPbA_U!Df@PP})$Z?lcHI-H?Z!dIW^(xI z)t9;!40Z>8d?kl350|It@iW?t`q(N?0l$7#%%iU5I~vtP#eW6J8_*6I4cg<_g|mbV zGAfJIQ8n>uD-|PEq@(f_?Q9)Cd^tuG9=rInV9`_Z5g=NSBK()=_VkOU#$~|uo7}@2 zG|z_zU4X`BN^rpjKp8t= zZ`FEf^YDbj^vr9}7ybSK`zBbX{4;U3gUNRsu~gEjUH%R2Te{}zza@dA`5v(#)9+fG zt+}C}zGb{)BmdOr6q0lRd-q@qb+?MKPlCT81@S;lLePGkgbDl7;I#K#$~zer=NW4f@l$R$-FwWsMkS5 z&qZLL0D7)0KoLuH+I<*2d@yQ1`taGQuH369uB4E$&)Dg!*-pK&Mz?(6cog>2-WgN_ zv;wRPvIN%$!db9}PxyJNqjlBA)5&T18(W6}3YBFhQQZTuUINZ4Acy=>dC?ryJkPw{k@1Pk8Z#ztx3YSOx!0f< zz`NzTAq=}*4)AeCL4H8YqJjMApj?~iZsAF6hQqjBX(Y9%YmGv>_g+;dpmS$vAB%~i zbte9Nw{tIG;<7;JGZ%YuSw|-1LF+fqs$*gVqrmclC z_3{vm2h!`Vp438K3CUcjodpxY!}jB)lBwF>QddPy#*=KSmD!?~?gNJ3xZ`G``plYx z_nu{OyjylpcT-QUFmt(1Q1iLaap-@P_8KC1FtGlAw5u%Fa_{S*Mt?>_DXR>pwa^>) z9C)#B2ZeBkatxUIR;AHC^{r60tDEN~@ShsiL|yFvW5vO`BjAU5M^eE*yS!L8`dtMC z2`B2~9-f%GzC>N2MjP}VgAZ8mtVnL`G2AXPd@S&+@qr4Dfwk92kCH3=+b@mbL8K7Y zzT!$_JxZ{TE0ELzN-K|nwYPy&SR5ZP1*XSRX&AunI6PCdBKs3t53I<6LJu}ST)1u97G<&V|<6G z$nH=_TbUqkb(F>l*W#Arh(ELB_%GWNd$AYreDrB9YQROIP4lX?rkUqbx!qP{wAJKLP`RZB$jJ%Xti)kU|m*g;_CfZ zY|mk=tJCP^!RLsja`x+9ewjydqxci}|4#}Vu)LbR3-#mm{rz$4qquN$l??cN_}2J; zYJeAWR@1YD0G`$VQw;}=Acc?)UOWd&7C@&VRgoLcSb=13 zsv)DEgMkvckx#zW=LTiIe5_#RShwV*F-w9QCIy201{UdCeS=MCH*?uuK(}dF-X_`W z8WZu`>wAy-*5~Y$0!ZIKRg}#nLK5@l+3Ch1qKbzjU!Ix%1rFrUV3{T2yTlMUAzo=| zt;NaeEk_(4iQl6emW$4dHNbc#@e8~qsGJU+}f`Iz@Lq*KSOv`!S z&hwC7{M5_vFCqZKw8W+^RIphz5yyD%-DMQ|i}X!cT8B&Xtqcsch2C(;az%!S1sXNW z^=Fcck_gIFh?i7{o1^a0Y@8k+Z|X@+v3(#<>sh!Z`VMV@o2sHlaGwiAJ;%>ZlELOc zsJ#BdCF(_Ta0$RTJ^lSCSme{gMjwj7o}lrUUSji(MnX|Esy zv<|rmh~i5My_it2IBZu;KY^pTc=!60N=~Uq*Lo0jGjs_ek}u+$42G;eSNI;tWLO#h z>{H~NDqy0}^Qt4(>MI@Unu<#&l$p zDQWY?Hhs;KnL=Z-Hp#c(m2%MHl~VdRekM76jP^kZe=3zt_0~@7V+k9CoiDRf@K=mi zGE4Y+4j)nzFG~xkJxw5RVDAnkdL6JBouw~yzNp^b{V;4kCcNG`Ox*ld<%*1s%<{?3KO@w74jjuVwFYaF;XvQR0OF$uq8t zQoAbYHge!n>nZ+_RK^C7-0D%1@kceaLnvWIs7z($ass=G{VfIfak`AsqbuX?mLT0* zEZ`m)mmvd zAozUZyTDkr49%&gYvn#SB~$A^3I1r*g!~t+EylWCM4@G>W45JqP(&OF75^5UYO5g` zpvXo`aYa461-n>a9f2yE{LJA+Cjl$*7bETXHn^H1*`4AkqoAjLNU^CARGW6Q#1 zJuLA&Os@8!!soH1n2ypV-+{4}i@L<5v*oUkb_R>^9HYi< zqETPNbS)JC)M+9ym-fO!c#pFqTTz2QbOwZ)ZP9_joL<~ynSl3LWF_ycO2T5s(pke1 z({Voc6L7YL^UXK#?mT{ui7*&nbyb=F5;pZYpDV|sYc;|V^?iM6tJ;NAR zq=Ltv63IlH1^M=(o&Jv{5hIKjxp8pa7MMndB+eMjV8v~-BArHNdJJ1h9wWWnfYUQb za;!*?2hb|_t)>G5+CaRL%r80yX59Jz&9^oZCMYOD zlf8Ex?@VSw_~^3pY9io(Y9B<+`@a>-f}D2skjEX7XGTXW(vM5Z@g9 zr>F48)`>*~^JQZ@u{#^J8_-5X1I;(OJIlFwE>ZI#)C$4*{33VbdRuY6zf>&L66+-t z5tBZRbt~o1MC7~p-T_yR?XawKye?gIOa^L6BotLKP#Lh4JE z1nn98-*TkQXr?nf1_%4@a|_ANY+@8(mQDRof|kW5smn+%J@m-ZORf^Kz!Fx@TG&IJ zR-k1L(}Ia+Q!Zk*^obsqYEgE-;DB`W8VfG90=x%OWP(OqCxf^q%%MnG5DK#Mex0c(Wkmj@;p%C=d1So5nf{(w z!n_W78mhv%CN;Q&*Ya&}5JmMivfNb?uyC0APEUwXbahhB!C;>94no#IO!@UX#neN6D&|fvD zU0c@txj9u$Dxdt$(9~vS z4{)rePS!W2>@$WF`*+)t*d$$#!H^yS;YzYuegfs^(W7#_Ym7ygD2uM$^4wSvp&~n` z*MuB1ood5mrAh~rc3Kg(Bpsz(Ypm5-jFI*k)?RtBO8hh=Udktx*kMSf4Bfa=fTM7& zqVSVUY6@2p1*BfJ#V|w0+z-t0-Nk0BIlljVeV^5iq7&9c^%b!6gwfe@gDsq7iYZ=S z@t%<86>oO4bXCTy{sTs@KXRd)MWPFc>)*miVxS*+XwY|uM*=p`%Bz^Z;yvQeE8gs` z;I9i<(-D$6+x&LLd!+pSiKYME!1v44H+1U*ZurN`)P_hPVA6th#N{%z!9@ciElr0H z!h-vsY@j>gFGlftnflIgsG^3X^JVIL%`eP(6SfxP<&-Fl-!4-dE{Rt~1P$DInfmIQ zR}k0^o(#z8AOGY5P<^r%I4@HhiVM;EWd@8wQ*K82@002cKtwxPxwnHwY>@T@FSMK$C^k1Mq6_4xDouesB zCP!mZlqss(fX2h);c`8iHTtspHT1=Ue6Ku(`uY9a{+%NV`3*0_mqmFtHR7-SpmxCl zBw9;;_FnkU=X)UHOq{Gr|M{AGfig)07Hm6Dsi|764W5U;^M;rI%0^~U@dK~#cloa( zG8_A?ogvX|60esRhDbye7Vsb=@<{SGGHa|XB+`>^oo<_dBtz_@g6HpXRj0~hMW!;_RMIg?@0YJF)!^mK;^2&<7 z^WbGiA%NFqtdY2`B@W*SLw@<2-k8s(rL@w7j7xOa?sA2M)y1iWM~21hUj6*bl}A2V znKtII^vI4DV)mLVR#DC7#?_RP8=(SB9d!|&D4IE4-+=f{>S?&;7_jc+1vF_R7vvmto(~z#g zIrky3H?9`IN49t@AyMmrFFaHX>gy=~>joGSpV(>U&=l)aNIAVvmAnw*N#>j`(Gi)d zBweCtsWL*_e!!Qt+vB08LK5MQBrt3|-H3utjE*}*k#+@z9?xi`70KWJ&?L$Ho1|sGk z0Et!DFFdtS8#F5s<{NBK2%M#J)c}-G;i7>+@ zU3*5h%v^HT?aUSg$YDtPdE7L&%g`r@uUR?=~2Eqpz;|J7*; zIBuvftjv9AC)@$Y4GOdQk`FF{%OOHGD4}=A^#^_E#4@dTr{dBriMDy0CnQpLu@1aB zO++qXEImv9@$4j|24J=K?_-C_T)Qd!B8!xO$U8nRFTQR38|%nKWa~%XttAykvfGT0KIZgvaWFG3y)_Tl5+10&w_K|b6=3V zt{owpIg*#aLh|>gQ`3&MDoeElQos<>du?!st`J&M7au z2uF*<1>FH@xu9D-)m0g<`VSaS{pJ5>?_E~y%#lUie!WG38#ZKt?iasGvD!|^-f0zV z*1MnhG=5!%LYl+5*c;A)ohI!$|4>MwkQ53@aUrIYA?YO9)k!Mvn{P56Jw&mSH6R0) zL}{(&j)3|sr(C-u{CPuIZs_V2{OtnPc<)U4amW1eiKPE@upB{}#@;bIz+qBj7r?I$ zAS?@$Mq!AMB&D;)kaJjIU)(WdU(BS3Qy63`!_rt`1bmas0V7~%<;y$oPJo}DY%L-( zYm$Ksf5_HcO8mRU|GX7{4=MX9f=`)UYUFOu)GGl`ak5&7-am?SUrE63wiYN!qFKL| zByO}Eu?MrrP(b{PUSMEK3n4Fnf%!v|QJY0e9b{mejWd&zGhrT0$t?6-{rBRdo(YiI zMCQ3%3Q5bm>xw8TE5V^Q+58lKgd3pD>C}1Mzy$n81BI)>7H4QQ(4%y`@kNj3kA)T* zh^IAEQOCkzc2=KEMStCA=MIqVnDEs3CXq&YSs>-PR<*C#HiU6Vl~~@Bp41$kM(NoH zZa`#9sEQ&5_$g;=Gv|LGsdW0ArT@H@R;z{U6%4YL)j}59dTYLy9*gBwdaRYv-6$HP zpKE2#^rc07OKtvK$5U{fIt%ctsnhO{NPV%-5J&!|#E!eosb)}QvX-h&4)ZJojX|~x zqS~vWod7virij5_YQn2OuZ$YFc3%oE9q`SUQqSZC@Ki4Ps*fX~?W`R2ws8K`l?R#m zCI;QtAd9-)$L)QUGfy-{?@~^*XcD4ojAS_TTit;H_?@NYFk*<=u1o93oK*9PU}z54 z7lpr3{Aql?C~TdZ%FomT231SK0#ekJH8Sw~nWN3=po=e>LzoE-9_VXxk?*p0N?fpH z0>~k~qhWzWK<$C{y?<|}vV2c+P&sG#qa0CYT2Ay=xjuWs&m1vvbc!JSa=D$j5L99m zw#A9|e6P4JIj6rTnOVW8@4WY=>~#CKr4%z)0#K-Mz}JWl7s_;HrEM0y(P~m;DpkD7 z6(Ew0mpnOpQxWdvcx@^r>^g!*#QU!$0=pn)0$GHhNhwB#+!PHE4ZyVArrQi@DZqnu zAEyYLWX0@5&k=k@VhXw+q@|P<n`;9jBE(KjRV)^+3xzaf z%?VX;JqPkBnvb*B33kdMoMvwpjj*Nq-uqFAOSHDy$*s7QF7tgO*PEDX%PYDe$`lc8 z&D;@3dYBz?2aHRvbpyhPRSuwBbEUJ zt$*hgi&S6R_aYOvuI2yg8v{rEx-}EVKHzcK*3?yL!``VAz5eBn~X=R}5hNpKP#O+3NrnnUpb4Aqb5Fl`Wv%pZ47>)9lVA;44^1s7pM z;>)~}GlUH$1f=!6^n<-989jq=83=g=3}uTjYQwFmcP!{c@9%$s*kc)&{ zjiMOamg{}1Q3f{ExFAkIHSHQk0fy+Z3*#SzY$lvnTy zcO}K8X3~E^Or#b|@!Fys-b-bj?1;I%!VT{&jzbAHL1bQ+4iS<2Di2CCr!b@7A`j&F zAd~peI5#W2v|p(9q8SOBlHxpBvt`v+4UpgP8CFesY&K%l{ZwmFA#Q-oqMdL>!}<{Ig$16eZwcVxQ{ED+4^x0U|BE5W`!q8 zGv+6CLtHj?T+|o%+`nbN@-}Dyc5Ur0ZEuQ;0$5^!u;$xTnF-k0)#@}tL4 z7AaZ4u7l_dDUFNftRwFi*0wi9&#Apsjp*gONLYBNQ4}y)@AfvI zxH@rZo$x~n8Clk^&V4yn&44w%X5=vVGzOS|V8yY!Zx#ZkB*v>-9Hf3-JxMt{S4}jD>2Q`4}RmXX3qYaOP z+%C>9#E}FA3qge#Wz+ZEL~MxT4u{Lam_81J^bv7x4&%V}PHguEd*iaDWK0KdRLe3A z?RIQO%%puY=89OdY#^p;2Fe!P@*(WA%~q`7ZTQ?z&h<8zLqEB^F?Jocab^=Sh)ev1 zTxW?zTlJK4?N!NK=Uh9XRd;An6hKS@u87`MVw@^IV0T&LYjJ$Z2ZA&e!et-$ATr-J zeSN2jLn84bZd7HYsX6dL6bsSU?|i@(6I0WDyq%EfLeSEMsGxFs`7B-tc*EPwsh<^N z_j2eb#@n@C`{~&$iIlXHE1RsZyecbzh;dosTDW*Qy<=glXz&~?e6m*>hG1=)7ru%~eu>;y+j_Ez?ko3jy9wE}m0%|kA5k#y)FFN?n1duWL(Z8(VUh`a1lV#8zM>XeG_qKL?^PEP>yPJSi}{3%ujp}xyMW4 zi4P(Q63NLlRounZm{zKA9GU_wCk$#eV#@^JuCr7GTn*l2H6L7*Ipng{fW*ykIxs+_ zf@_ztpBRutODq1!fK2jwbLdCGkF(tSQAj(K4v4bni7s|Ul*L&5I=y>$c!u!5Gp~7; zu;+R0u>y04c|9DRZd45ky(8{`ljxQsNGUlTUNjE3sD2b5mj~op6WJ`^i9B_^sfJ*# z^Ln}U@__fnJZqkp2eNqFLSQCZ*_6(5L@!~Y4Q~Cx3L!7gjmUyKK z_a`2gcSJ&ww4e<8c#dla9)yGLjw78C*)KkI2eknu{m>mJ%0tdID(=f@5md)%ay>S; zX`g;mt(2*Q^@k=G{~l4NKa=Z4h4-(4-^rD;3E5$EO$4nHWcCx`1y)PwC^aw$SRUY`w@OD zL~Z=}rx0HAKmOZ=fBe(hfBY{OaYRWLTlaw}GgzqcZhU!p^u0=%8S#n_|H5nA;Wjc| z|K7g$A0F_p3%F7EV-mAfXa4w?C?s;0Dbqi4`~EL08itHbtLgVT^_0R`h78P_b)uIM z_222TvgO8Hm!i%6#Z8K@1+fajpG=gm@b?P;=T(?IxHw%sg*!_#o{Vv05sRe71wwOs;i<+ErWsTmo(zCYvWJuyP& zcW!BfAur#)R-g}GNBpsl^B6&;a|KcTIEFkk9nLHQDSH5@i~Qd zirPM7T#YI7_E}2O>iFEBW8^!DZ964GCzeO^HAdvDt2x3*c48?nFkPpeR5) z8c99Lz7INPF>TIQdZRQcwIuhQOAicU^?5; zh80_ty}wnL!;{kMtW_W;eCXxYeczB?OL3V?h(KbsOFu#(5P|MTz$G=?8ycaxQgRAS z6oeQ`*t`Gz0zL#pZ=%3p-(+X&p#`$Yr+Db6YX9|=pZDylNU!lA658>YK!yMDugbLFN@H=9FaYS^`(exwn_UL7_~y-dJ&Va;M+^X- za?4Rn7$!@!s1|WkBInLVpvCJ(HYQcxa|JS)HOhM;k7c9E2NAslHs5;`aL>x|V~<3P z5$FHVqwq?;rfL>x-&}n@)f_o6xD&7NCP`KEDs!DsHLo%+k45vUhRlA9U9-l7~Q2!5CJclF`Ym3JtcF+Dbb-ET(pWSt@~vA# zl0?+X%DaNk_tK;i;96oW{Oer!)r<7ANH`onzyIhMXuG~8XWgpe;KftP z0U*k^&3OX;dI1I}W*2bU@G*8WREf^4sa0AMC^~=e-kJePOV-5~_u>4rpox;@@qxZ5 zl<%3uOJ-v6LZ53#6E1n2^#|JX-P(lPTv7EA;G~%aTq?9=rRRQ6)ks1^f`V#5OYUbh zAj;9go@l@jwthi_EOb#vG&3;W6Rpy}A&Mq35DVcgYS4(}BCgqOEtY{OOR~9O?1-ZG z;Z@>qq>Avxe8sEI6H14Kxd%W=r7rVe06b<&)oB`hjfRj1HyT1Fzm(UA5tCNG_j)3V zHjJvGNlAnZ^LkogijiAF(WwuJEv{m{B6|CJqf>~@j)Ysb z8=@1X)&@N?9{Wz7Ni_iL&uTRww7eb*;Cly%q~JU(fG9;S@Awo9B9Uz!1-qEbi{#lm zxC5rzvj`GQCewJE*sQ$e#FoWxc&$C_M-iXHn zGX%dPfA3p`En=A{5^-tidj~{<5VxM-WO#YnY0-m`f_@a7EW|HP^w^}y#ao;_ckkq; zf#`^XWxS?p2mB;P2gyx1R+Hpf_MbB4ig!8T6h7+tHN-5rORVNcXi2TS#t0&b3fa$h z40J;xvnY%hJlWhd26m@y+IGO?SVo#VuQ`@E)A;ztv0@j|EB)r6t`OxQ>R$82msWL+}9AYsD7pl3D;*CdEX|DJd;PbF_eLLm*awA2QW9h*ZdSlk(qzm7% z3Y=*Ddy6&T)ptUamnQgjfW!ENROcH3qS=s1F0K?5*)68b5bZhCLsO!*84(*AkPIj6 zsGCZ%6FQT&&e}y^iYXkJB%PAAe&-SsMl_DB5Jev{IWgwv7D0d`~F>YTv-{-p1OtUUPFLMxr*sf8Rh0kSM-Ep6hW25 zsw+WjtQ!(EnAf||M$r8JH9>g&8dnbl$^7ELV--ah_>%$jMv^6fO0Y#E^73zNy&{T= z9hye6b!Gv)_l83!acf81o^~9LoQF?6LL#*TNU#6f4KO9UL;UHkkcS2XusKh#WSw1c zq?Qd?cf}RF$QoC9Vml`kXnTV-wy6bA!KEL>xv&1fU^!!mj#%HQiq(#O?(k(Jl7a9C z&QCz+yd2nD;*SlM16N)U{)+4TrIj;;GHwff0d*k%miWJpX zP8H5-(kDRnSWXpKhLKg`36GiLa(^Mx68Ftsc44-ht6fcpDS%BhPQJGukqOe}!zdBo z?iE4UqiQ+muqL-}lKH(wz;?vF#P*P-0J3I?BBz=XjhtuNT>wqR+p}&DLb%;m;7)dt5elw% z^KFu7fYR}MQ>Se{r=}2oY1dxxfAmeoev0>#Z%%R)MV$8UeWN8!?)v>H&*f{4PL;RB zl)-;_N03bmRU&E(F7ewP(vBg=S(9M>j^WAI!kFyd*Erd#L?1Yqs~rl>u=k}dweNMj+CP+d+OLH2&Uhe}KV(yhxrK|zD957d z$&~eIl(L9qbNm8>Tg!Ih-Z1{WVa&kyLfsAg?E+SKsHP$|-cwT^F@Jm_=|2xR+Q2zv zVh%u!nQJ%&;4tCH{n`P)IzVJyA;GG`;66e$4h;~MB4($nLR}&QsKUtPr?MxEESpML zF#df3jNF^g5LVLA*)it5xM(}rbBx@Z>*?aJr)=TKn#?Rq*sdo4TZ8-3LRfbAM9gIN zOBBxBMzO;v&B_YKa|d(@ekF&WUe*Ck`ALDMwj$0l)NJ zR2bfWI$ZUAN9=hvTN%Ml@i3S>U?0qV1^4p5Gny4^@w$#?@M-Chd~SgU)o4~ZEs`bV z9dRAa8`Jo`49 zyy$TV-xssP{U(r?dK=~i5R;r^?9affUl4xPcK&>k{#}(e*qZxZdcP28WfKu`+$cNw zb!gOJi*-2NzA8%c*^kn##~p|-og*FkD#+gf_y-Yq?skH1rK962ca;9ekIqwvgltK~ zcb+<4c}i9S5_8{*KfT6Cp#>dM#7@9NtStISO4%2&Z+vYbv_wH%NbMj(c5&~m>YspT z+)i=$TS1@8T5zen1+2LtP9S6IRAsJ#`2R-nr{VG9v-fzYV+%z?L`_>g z=kUEh1=^XGG&ShrizXw`mNV82=&OygKA{TqP&R0T@x)yL^ z)cn`;Jtpka7xHDGo>=JzaIa4GCn6-x@82%ppJobYau?s5DTD6n2nmoBS`9JT8WnX9 znic-a5l}a030j1!=%qbk6E_e^r zqbWiFM|T~BEGs3-8CxNi7E9Azd0ko}ouCm#-4?Yz+*)+_>XD~+M zGSY~V;BnPnTa1-?s27R`bPTc@;AyCuA|u5*HQT|v#DSQt9O%;lQPg)XMDBC$|+TkUln!cVCmJx^Z+dtj|>A zL>P`nv*gM`C)0)-F8hx7SY$Lg>J0k0M#OBdD#S(Dg_lEsLTpLj-#GE!BATPg z2U2CGG$Xsc#cvn*)a|8CR~Ke}R{-ZQ(YzkPo3jx{Epe2cOHBq5;}JbrR(=lcQflu? z#EL(>^B~LUf4jh^xJOFa19L>ehIWlRYV=&y&a;EDq}9}jnHZJc-4O{FXNqRt1U$_9 zN;R;UwWavRz5xoA_@=XkZ#Tfe;1RtXCZ#AKBD?lwF!{Q}=WglaYY*RR6ycECqr{m# zq_p(yZDK=ord^O+`6`jHTAdNSNK2U`zw@XK=(6BK+R#dTv$w*(Rl9t13@@$B*@B#^6u88p!hE8>70(y5+ z3!<;Ck6S1&Oc(sAM}Yt<<}muA(vzR#N?6FKZ6`#TaOI0>6GOXC)`*cPp!q&gW8s}z z%OcTH@Lh@<;v`iV(=z^QL>y^@xwc56G7K;kBGITyDp%&@Op=ZRB3mY=`L`CyM@mRk zjYxLxaOY|@EvB?ato)Y3l{q3ah(C8g#82f-!oRgByMi3+52Q8^Zkw(%hvYQmfYXRr zouqi}qvu7|M9o!CS|5Ug-4KcOv79MA~r+Ak_^E*LYO!{&M z%{(#C^{4REKQiI2kTt#fU(a_rugSmjdq1W(8^^t$DK z6ySlVq=!5_XWJ5&(eAa9t%j_??-OVRU}G+O@fXsZ(sEKIYDX{K zAD;oqVIKS2y=mT4TkfF&_|*ZTyrgizAV5@;Ry&4mMh5|9O^(6V zL-#ONkeZxsy1A&ktci8SdzTIv{&mZ_F3E@7`ItS#(EaWE_lrLb&$sW5qA3_BMbGqR zo^`qF2`nmSgfEiNi^`{qkGF52ym7N*c4+@zT4?@>#k6nVF{{V>q1WblB&r>aB$$;f zt^@CT2T)KBSpwkk?&z@cDi}#xXhy-;w%I_U4v;iN5rqg$q77cD;$9>y@PWuZrYDuc z1u>A%pnBaMS=Wa6ULU(dcnN0Lp*uu^#J*tXJ*L_%9={?tMto7yLM~R?lU!N8@=Vs3g3jg$WhVL;unX)R$LD2gzoIcdmmB>E&Kp?%E6SQ>C#L5^YUy*xjDaUByQCt zqy^&SJGpg=ncs(SwoI0A0C-H)imJ)Bo9$XP27wx6nrIBNsgNO3VW8Q$jVJPN~z(Y6J>|6OiZ{=rbYa#6h{&oRt zJT^>W>w0-{#{BV#r2ljPa|JrN5}pMUl72W!H&%Y4kEt1Bo=_DE&TBqo_h=@@Qh! zSOQ8x2mD(BY7H{7A13N3%|lNxUrtmnr{EWNTG@4?sz$$@s2;<9GfKA}|21Yv$Z;h%_u|_P-rp(y^kdxM?c2`O4(3idsG30wIDlYQf6!14g%%BE zahFzgyk^d!#M8cQN100E=09+ol8v+cZ0^A=m=+0B2nHvd5QXpy3<8Tqt1Z=|@dzAYx6& zrGAb^0kmBS6e{MJadWwrK*sP3aaoKcUu|yi2*uGMJz(+e1U~?7?@V=+D(lVHsAoSA9 zh}D)2Ce0Q)j;wp1%IwL+AR+}qa>HOZlj@x!b^$!ZUUcAAmH?$LtfUT5h}Fea)OfMHcx6IbrSKeWvg?gFECwNEimkc zEzwqH5!npMM)_9ZcU4->N22og(&WoZwkJ8V&EkxF!n2?()0Kf3SePbcU+{@2a;jbb zzacJL%R&RUu&!)0pij+aMs2UMV~c+WD(?Ik^u&aVUR46(hn6@s@;vK?#Hl3gmG81W z7JDzAX`aL^O5)QnO{*af9!pg$YQU-EE+->mo{tt5k?PvxUo>;mY-cKcSv4(4`CrLvGZge~qhuITzFK)%*suaewy&{zsCm9r4BVpNq+ zz;AHcxfZ_p#A(-Fhw_2b9#4#X;I!vXBpf)6G->O5Cw2lpb6TXG&-Tb^l7GGC?^gc+ z{LaD_S;Yl|Sr_({hyz%0cFp$dO+l5`n*y?%bi%X9SDV5pUC4-Q+;vkRbt<_NN5O4V zAc-V_T|0atMqG-^hPZADlVTn?`7X1B+%i$sXXTJtL}A`4OaqZ5nmB4HL~lyd@Gg{ygd-(H z)rg_p!K5pqus+ze#STa*wTzppM$Qd*0^e#ZgwAwcBZ^uPQTuP5=VB*>Lmz77w&>ai z*CIEjtPpvFlNBGXh}=S@rQ(6u^3)o4Bh@h{UDgBKJoT(XoGm6|uU5?*a<|E^=8A|B zUNU-SkVH&&|Le@)5?A&nJ-2b$7X@kO>gc1#wz#G6-e!iE|4KN_j1i^!Ldwi1xg+j? z@!Y>IbbP;T_$owwE1q3!`Nf~&V~LASD@$*FA4^*g)J`bS*FRI^5qoAP^z3jYA^U z!tt@T6(j|ks4=}@SABd7UP*cYO%)b?gb=@VhtNehohZa;lOdDytvdD3QltDkm#&K*_igGWXr~U`AU5D1fb3uk(4^3 zQf-KH&7;te++!z27CZy4+31=JwZHPw5xcqVSUX}%7bnAYAw?!SX4qaMGS-~aZY|<> zB-45#4k?mI*3d|mZkA&L!dF2;6f^yvf{PRhqM{Q`PEi#qn%Pd(>`vr@HXjhNw@8af0qvzhKM5_?Fk|^HqV`cP0-Vt}eaUqM$ z6V|OU&7QNx`$B#~zVza+!?Y}9l9j-}{(a~OPnPfPdm%#uZ0q0ueg2MbM`_ovCAQ=$ zjhOGhSJ*yjXO9j3ojN+v0spMYtA6`%gnX#q^uGKL=^OZe^>VTe`pM{@a>{akQ0__U zn`iBRo#}2^?#q7t*U;Bo%*p7fP}$(Co(fjPVZxF|17alr!baWV$T%mb;CngqyxKaH z=!~C^jv_0#e_ylFKWRV7h?cL#F)NQDt0bYP!mpmh=Z+=J?@xJhWO1iAzW?YLa-8zb z(R4>42wgTQsy1YtMKv}400^>)As_at5*Zt#)fk2R#Y~GGgSc?gjKYwOq_$gG1t0<( z?~e-LdDE~?wL$V`7f)3Yv0Buc1HhAOYytfB0?b*^2d`knUtDiX73kAsK6uNpWdc-w z<=0}fF1}oP5V&7(QF9*Xi$M8~eis)knOrs7ftCYCW)toE`TI{2Ymv?)x>tUj<=Us^ zE;JaT%xZphKsB5dCZJ>3$a_lG7(6kEI<7HXiri>x3`{GCC8#iDa+Aek0)9*o0YObt z>|!sjc4_{_uURJ%L&)N(5K}cR=lF_Px){u_r(LXI6=ES)jJuAIB?3XaOW1~p_k|!k z8~`rGkLe#@1uGG4%P6?;0VnyUv+lf;F(600A{X1O%)`xvL9X3}vK{Y;-UyG}jyU=$ z#ABl#HDVyecurTu$nkBVXaX|vjJ)EU=&`TF)t@*g)2+Qm9XRLkXpYOU4RYo9*#dgq z7(_^gTn`K?xm2Re-TGN$$vNn?BnLYY*e)Lci2z)BWpI9%9bAtt&Qv0@ucPZ*#ucjv zNzTyRQ(p&ZTnsBxhJ1Xhv1q_l9jEAwT(jguD1lXVETT&gN-~b(LpX8}m&5N;jnF*S zECl2EzN2v{lcje<6QP$|Te5iqQaXWlEH)0U|?44yQ=q$yqe;aHWBD9S*d= zk}iKiyl8USoJh%Bf}e;dw!g9V0H|D$X#j~{BF~09aiU|9TKt_j8M-N4GRY^CuCZmJ z4?)-H3Fj8`t-Fr&Bq2@skP)&uaW(>wh1SdcMoXe-3Gp@xE;1s?)XiRxo*dn9(7cWw zIYVPu`#EHUm-c+UB8nm^ic(Yz;)$|wd9DCtufQ{XZ_1PUQbhKxDUZintwkdexIHh! zz9Y&Z&ZMLVajbY%;|}=wdL-M0LDSN#)z*a{3sDFoaK_xpdF0j72&Q)c3OwM&F+8!g_r1x5|tCzN$MgMl;|GWxE zj*YTxpTeCHE&O!5=X@%`YmAGxj^}d3)He>qU;TR~Y+4%MdwKF|g;Z2|QASgZRe#Tk zU@dz+gY*6Sk9CL!^VL2V0?%Y3B1~Tfy=+F2@P{pNQQ6Dj(^I-HHHbH&YmG|eXj@}n zp`s2Ibli^g$Vxd57t`v>oa%0pQjVLMSb$HlDJRMF>}I3$=z>CR-f0=qq9D$4-DxR= zzNr2aEmfpjh*{HOIWtOqJ2AF-E4zB(436`WO>jpGUXCFqUOGXtMJxmrS|ku@wYSz% zzE{o~RbZ=w^rlljDtQVrZwB65 zw}~F22$DV4CU@eVi%eWDZ!c5}(IUBm;y$6SU#)$)ZXFlV#yGujhalYu+P|;Ih%Pc^7U4Me zNo2|ip4`Vz>E1v>dyi>W!khKrQz}>p4(O0xA_$9gcdHA7L5%;ly1pHuqb^ZwkgNUk z_w6%DBug{gX@R(9*AA<}BgY_O+vi@go0}pfxrv+)V z@v_SF;>RV3$$ZIB#$!R8xoBHepWK=j@Q&qU+@nRj1^DUCKn8$4Ezl z`JkF1mYADMX{S!+n={Kw8o5D90QeFA-v2G3lRj>-9EwH907PFt(M+ej|um zL}JeDXe~BmOtB`!DB7N;j&op7f3s+-Rho5-lX6Z$YO<%e3AS_;@3Gb<04UQ;ctx5Rj7W_meK z60TrJBSDohle(=n->eUc8iOG_-l{9=`$?vV79gk4tn<#i-z3MtA;*p4L)eJ2E;3+# z@6raRaZw^>_0oM?JH@j6dAn3-&4UFSfD2kCMRJ*FkL1Xu8Nu?M)MLjy4KM9e4pc6U za;r8JP$P6ltgPn5kX%TzPR741@Ttz2%+1CDK$H*@R4<5Oga_}(ZdfjLQgNq#$j->* z()|dD41)K55Fl5ET$3xaq{!Io8gnzzNn4GpJR>;40kIk~jO8mS(r0HE_(UW=q}^Mr zQ?!+2$d$JDfzUlfSX4FQjM(0FmKvk66VVgS`9KN@C$qCVI~TE7a8dGVkmb~yxvTg% z$e02VC*m;30;+PB51Oh#%Jbl_3~Fxp#c0GNpztOR-Vw7LUNFE^oE3cz(=ZJv@;S$< zn<|+Yv*p@U#hff!g@{&XT5yyzqBrfK&?r*4hREgw=<-oDwF8cllaO7mT+J7@3&&o3 z+&Jc7LNj-o)f6+7uZ?S}sqlvLmBU)LL1u=MTkofBSuUI3eQfr7Q zr~oZ6T`syQ;iM@GH}|H3b+2Wyak~Y_Y(Wt|u$vI(g@{<1-g7x=+&j303s8xJJC(mOdOYdYf9!?Bp9DUB#1#fWZe)QTSLf-6ppV=pHl$Qbb@pK z04Q7l9jXeVwHtH5h{3mmrlL9uYh&5FI}sDPtipjf(Py|UQWF8kn=Iq_aU#f$&m%%S zOazCe@Y}o{qO!j3r!P$A^;r~9J+HkHNhGrIiYTX=;*ypwQ}FFI?tr62blCdTk54c} z{acC;UD69L$*#VC&pBom>GmDJHcaWp_b#x|1;aZzWm;Y|?E(#HL&JEp&FC{4=0e7X zypho~L^(gkqyg-yqSC3Bs6Yung%#E?uosa{483rmCb(as>5BM}EU4M^BGodDV6 zKwC6`yXLa8Y(7JX&Qf>c7Skvp!X0KYnPS?dngYmT4P1!sO<{TxO0ue{?;ZXMAN6tc ze9PAB>tXb!*insdfTKJT>ENRI+EYg4kx9ZbGm7`d5iS+_`}e0~qVP7OC03)zjO$f5 zUiFn#{*;^D5)Igq&uP@$kH1_tz|l^)43`T-!!7CM$+e@EN;J8$p_M7{a!w0)xkuw= ztg8B2I3OY=T5fnO(gF%%3p?9vX@a2{oY{Ns}Xaxq@uWzQgESk2U04GsPOSEa4TY2ULA19&d%%3XSOT$#JxtDeMH8&_JQmf&L6lYXa3O6beBHJkpXi( zMT+#>3dhp{5HWbvvT4e{jRplc3dU#(m${!W{p9pxM)9GL$zD-yp5KQ(j0iAxOmal} zaJ|4!?S(&%a&vw10+`pDA6s`azdBPkN<{Jn!tz8_o|KP?>x#X6mBBMcQC^KfxPGKu z<^{UwfpUunvv)<2BI=<~ZmXh&o5PY;z%cBWBL!(Y6-H>G*Yg2@o1M`kVgtF*-RB`; zW06hmG$usF3r+6sbVeC17lS5;m~iQYkt6Z+P8b}L5gU64$XlOF`rH9x$l|%7JK)-t zlhKS8eZM)3Dsr4}iMyP}AoXO)USkkbr`=rd00JKPv`oeQa(*mE{EbBjzS=qz=NL@%Qr&sB+;1JH)J0puXV+S{2( zrXd3TOu&x`s(6Vc-5b%|Wt-1yZ$urKy~#M>_^2BzWENi#CxDRH#VUYnURpIM8f29v z63EzSRbYb)|8=15;Z>vaAC_b4+d2UgUuLj<1CsbyN6~ec_(rT z`x+SC5y>eUO00@OnBe6UjjEBvM=2(IBV{i79dQQ?1#*!GIIM`y`lSLX#rxYwFW%cA zyD)Y_qZ%GB9#07YC&KY9;2}V;T_A1!{#{NnEw%9WokKR)Y9%gcHQd(lp4HZ*cv*(U1xL zPp;^_Ms9V2Tbk;^sbHPu8=nXybk@yUK@tYNy4MOWLPq!;2=%YX#l?X`{5!c*)>d8^ zZP8R!pPEU=N2)MBmXq{>Dv^|h3wu?$$Rmu>!pF+V6@!`bv2qbVEkF6jXh+aFWA(fv zY+@!1#bxMRB+ zsH(VpF7Sa!Kn1&spLax$?JB`rch2HwSWHc-pHe!uqt>T>`nDV9aOmd{m^--&aFnD@ z#7*#v{Y^&SCz}OQs!LL-kK9J1_!O0fcvbWJ_w#qr#NO-ue{6HKv6xDSI4j2Rm^tyX zLTTZcUkHC)D*jhT=zTks8Fe1|MhZ&qL(}ZTqqLA|HrTBCT=DCwMwM?dq?qhmdFh2E z;etXSOIF3=PGJO@pq0_yF#^7hhjAdAn136fj)4^j=MIG-V!en6n}8of_6a{qOI+R= z=;H`G)8L}zow7L(ABdN5Ms- z4b3Rn=nje|7BGCfaVAtWrm7KX01vhsVq`TWr~8>O=_-2cxYufVi&Rr2{WRs?RA%2f z(THg+=keR<6<*D2)})j|?6m`cx0Gy?ZmJRtAch-_%=&iwyd&ZaLyV;>A}Ph&=Dtk} zN6-^|Za+;6$L&lz0IF%>WG!*NaC8RVMbk$35UuF z+Goqh%CQP`m~!aKz06iiu-CnEB1%}3Kg$U0zgsZML*=}nlRA3mCwVec$@Rw1x!>%F zJK(qx1yMysHXSx1k56geh(y)Iz@e^dv)P!y%*=%J7C(y&P`Cj<#$>L3HJ^lg%wI&O6y20GTMe2LMSD zQdxRBu5XEwTw9kcHZIrwg!H#q^7Vqte&Pu8wzfJhmB<&ow7T&{wm*~W3DJ{%_6x+2 zN=s$0`Wc(y+CBr?EwcQ|QEdkxQSo}!08wr6Usc2<{LAfiv@MESUJ=JeLNq>a^&k|- zEn56M2nF*wQ%{3{C8k}dnwya=%})C=%P>WBJ{E<@4P#0v-}>qB?hQ$^RH{RdJ^@m* z#nu4tAI)i}!Lx;YK|}B@!?i|4sn1J|J7S`DLC+xOa8c6Fv4xoZcHG92ddH= z7iye$=%iCf3w$85!;&xQ!aj`NT(VA+f-ofXi~^8QMMV4oAad2R#1sH$)fbQ70g-V3 z;m}=ZY6qwb)kzd20y8MYgaEbSikJ%-o|8r_LVjK5(T-Sn4b4^ep~n1r-PooHwr_?w z+@I!&94iX`yoZU>ObTq#3_`kNSbrneVIt$Ro(~SM`=(1>5FK7R3w*iRN+OighB)ue zW!hrms~F_kovhOf9=mB#B=oc(S5}L=paCt=nH~gYaBT;yo$BOYDRgh$%_VV8(1?Zj zL*_b>P=OrI`3jL_<9K>p5WOLWTh>EL$`o$+?}&JrC95WyCz40b@#93}F+xt!^)S(R zj*u#}7?>N$PVNpk?l7W+%Ax+tzya(qgv4OZJ;lckBO66~_3YKiMwx19u_q4TvBMCc zJXlp-!s;@(SzR-m7?*Up#m_7g?c3#h-ryT)C%^k4UO!8ugp~Mof^B2 z$n~66rwI8UgYGQR=1S8(JD|%SLx`}@8VVfbuky#XgJm5BmyHS&yf+eFIzvgcP})oPhA5@;e?yfgDPq@-!EvYWKC;zEPW=ebdllcMO^BL~dTvdv#60q%aO-FEJX zp6FufI_+G3CIEP*pCRDli zN>*(xNzWd_%%rH;F>i#LY5HO0h@)U*+S!rkUW4SwdTGql&ZWh%CjlN&Gvu;JQISsDcIKh((US#}x_F(mc- z@@z6$gksdQ&=F*|o&d$af`cfkTX++5UD6yC8(U{`e;sA&$h!Oe@ z@Hm3xs2Alc5`|hqQH$0caR(elD}nT_`Si}r^KVl}p5on7MWlttHg$aO3OV=ks3-mV z_x`;-&G4PHPtuZVCAd1%*j@&^jpkr{|L?|$Y;fb_*stS*B z(v&Gc3kz)yRmsjMI(R$)WCaDEH2_COqN^!#MPKJkqcI35(w>Go;IG%*ob$ZTWh>P4 zUh%5)P~xx-u(p#){vPW5*SSzf4+G-aI*;O`yauxD>g<#fwL*a{+>TpR_v046VB zCRqD{0ohltwfoduI9ENc5Sjtrw2PT=02HFBTg3oxc&@7PBLj%g z;_7P}Yz$y_5NkvAV;tNr3>czQA?|>wT`Xczl(_Jotx#(h;iq(j6vd->ZzFQqDLq`O z@iq+FQp>+t6q@WFvUiph&Gk^sO`IlKM0uk1?>sX`{ar>f;aa~pQxh3z?X4M7#$g(z z@CZ3Yiry;!jGiJgY|+1_<*H*j2^fbSiE*D2GR z(CVgME)iczR5RgTd%-y;iO(Iw+Uk&la`@xGzLWcXz=>EAJ~u8CQ&w*|HC|C8PrRxa z`rUcuWNzgqGB;ky5l~LYSBHna=GCw{ zl9M`fE-;y>kvjdgl+r=Hwv2G2=>AAxAIz&CVd+MhMxX0JRx3L%L; zG|0IKGaR?kU?*a(a^&_J5g$C#*RO~KP+SD4=juZvb^&_s^^}u4;tn_pP@k|cIoy0L zuQ1_xg$bj0|DGK$iFSVP5;8`@BrVm~Gt6Qk|y_3vD ze2aCBf#ARkjjtN*6u-ECtbOS6ibU;!x$}WPY|`mr{LUbm#Gz^BJGlf5mu02NEgsdh zUKzls)dfB!pmR7EUJ)y1xhsQuKZzkOI?U&OioRco+oJo)Eh1M%iYmvhK<4=iRgU$F z0h&VeE#=sCA~@nIk?>9-67sDbA})wn={c|YATpc$rJwnL+Cym06CV@lLT zrUQdSzs&?ySji1}k+KA!u|^}Zo#N_r>nBOy>H;5#l}C^qZirqam9s@JF+%2%Z&m{c zejJMK+?2AHriL)x5qH2*%HsPe8^?v2S$Q-Gq2E9%R;$3R|yii5eAm-3%r8Zwf{IeBRgs;HVq66r`dIXLEI078 z-J2n&4{5(I&B&bhEi(R&#~VSxz4RS%p3!lk!sKQrq#63OBg+mTyj`&ipet~MDIh|h z06f$4s=$db_U+{Q6kHTl!e9|BRP}>fJQ6AF4ET0|FABMwpz{Lwb_2}oUu-JkmDdrn z{eedODXrBs&dc7WqiaOGA9&=yA&Tajc{x#?BMufh=QSV@E;}9oBrdcEK2?mpV^IL~ zNCMv|stHQPc6Xy12US^UMI*|*Al3M7PKxyR#23|wa$dq;^4cPa6nMf_eGsiT5t~N= zGM*aI9RQ*yV|1x#$SoP~+k%hIFZUG55IRRAmXZUGRyX2BGP|_Er^v>xVi}nYvF-LX z=t(O0LP6r)r3*nrvuL8#Mg_D4fk>4Q{NvaKK1=^;#1!rC%*lAcj6o0`*9|a6+A5D zS8^TAb4DU`wL-H5+%XGa!^34AU*?St-(su>QkSV3MXx5S_FIj%E)S71SJI&eBps{e z-i4NMkbu966d6kY#*?v=@N#YrO_8N4n|o8OA){-=BvjP@sA|NhskjBO)kvNB-|m-4o~s%dPv|kj&-9NgZ*l! zjl{$_lMl3hJ41$y*s2pU;`NDycnubP{wcTI&KOHHZdJ1gziK(reW*Em-vUqRnu|0_ z=Cr-5ggwJ!?!Bwi5~TU(IfYh}C0XN*L!ycK=oa{TDUTV_tvzDICxVJAM2BrrG}jiL zoGRjrsc)>q*G;BN`N%t&zZJ{_^W-=bUWPI|RS7$4yYpccx%gQh5OX2YcDiaP2$i;}pqN#MlC#0yALf5DtKZ&50glg0i#yzBiTgl(U3R z3wqtjP}%Sd!9wG_Q%dBniFy``@XMH(UAaO;36UDQ4}}1G_0Tyiq}#r|H-*J7xl1p6 z^oCn;qt`I$DRZ-k`7jYZ2_o7as3syA8(UI>o!luEH*yu=C@yOY5b^1)kkWW;_2OM` zO$FbkFu!+cl+kT;`CMNSRrd01K4~pK?nOjbLn`jy2{u@CJym3yjF8d=vrLzWL zy+IP&ol+4;kz{xpgUEl)@MOop1fyNB?--$A-sv<(G_K{uQ5aEVj$8XZUIjD$N3AL$ zJbm>lC^C`#;3>F>H2ICxcbc3m{Uuk1BBzul^)&`6AP!avqo#=M_5q`ilu&nB^cSLA z=&GEFmB@A)0KIr#?rjiYu9AgKVR%9#rK;HKT3qPjKeAOMVj|7jfvxO9zTJbV%H?1~ zD*BI=2brcAjvp#-^WEq%BecJ5FS&bPNp)k2p4g&t`P>Xdc8o+O$rC`f?UoNw1LPQ+ z%`%vOy~M>7Jzcy6d_+Qh5pgJ>B~viI?EnYkT3zTt_X~vciLIM$?w^-JFO%B8UeH0D zs-rc3D)F=*u$&f2NdmV(*dbA&n84)yj(;G~xfHheXk~xLu-+Kn7Zk>yH;fnk4-0PK zZx^t}!z^45L=WDNJLZp1B>m?BV*$-f8JGj$38QW%;8zDkGTvB?A)CD@|6XGRPI00K zYK#&*=|U6+HZ)GoLkfc@PeS2nj6~SE^v=5zfN&D{$EY#<_l`i?t5ot&-vZsqlqP3f zGnmRrSAQ%W@XmCD`qWZJ+EI9!oui%>a;2nR&%Zd9w^;w8Xz7Zc(#MMMS|_;w?^X0F zeXNM^{IcGEv!Yk&V?{hjmH4|Ay-J@d!ict#Amgnf(lWGiXG-}QuLHQ zR+I^&|GPYTN*^oYC?0;7M^EWfMWMNB>^!lXb~+yTfue zCs%}wP2o3U3ajoLs8%&yHw@SsACLZSsCXd6*G!lSov=?8#xFiJ;?WE$2>WNqRVuYFPk!!g1 z^i^=};@;kS9kB=edbQVHQ%2b)Y=$r>B%Mx!SIYzjt$gUiOjn9Y)+sKfFFOV*`?O?kFG(s{HX-(FjMOMx&c07 zSb^YXpfD8R#~tVU1;Er7<_`ytP0Z>RJ2!cOZbPg9h0HNkcbFjAF=UO= zT!89i=|r$ark?belWiuNPph zv)-o1q@fN>RiJZpPVucR4ft`#c}7zK_J#O93c9#(D)a92#;HTSFHW(C5aiP2jrtSd zuLm&ZVsBG66yDPY=3d|qco=d9{Z~I+Phi(<--YX=OXH`4nf-E+P_T=4e?-jEW53h= zo5dg9+j9oQfOJ1B&&!Z`D=?^9?pl&>fu6qdb`D+_U+xijXY(nUGN5ft_|A11o9JkS zuy($qWsz$5!0u=nXJ;2Sw9wJG?kCjWj#gs}-gaxvnnYaerMV73=2yh-Rw2K%%yhRJ zMUYp#|5S~>#$t2QfQVhI&AO==J%Q0~eMU`jiIdlkn0aSkxUtiio1w<6_%#8Vvx&@K zbdBV&7e3#Awt_fM`8o=!6svP-|LeYG11ST%w$_GNR6ZJuH z)_Kc+?f}3CJ|6(6k3_FjG=$=bWO5}Qm#XZmVyTO=esUY6uUSf{Dwo?dP%I@YF)pf# z+)RF%7&N(=r5QF}*U7TN6|dLJ%n@`^pjTeYd65IXCZ&=^CPoDy^Rq*EXaI2`OI0+; zs8w?N-gxa>reN=Njkvp$JwyS9+}oSl0psCq6^A)s_V=axNC+bsZ!iA3i1Z_4^7!5b zWIF>KzRUKBb<YBGt$N z%fwM`&eJL!*Qyzy1UjBOlSr4 zs)*AQk&-(G)&l8-5lNTYL9zoJ;RBpejT17e63bX&$YzdIKL9d`ShEHM(jryOE6lJ=F;UGwhcP=zu|G{f zhocsbHF^d%Uyo$;&^ZEsm=if7Ks`ZU6$Uou1S3=!CE1sPzhkh1Qqn7$czroU%~lfA zc>-5yK*Y>rsp_wnkeLFj3nh|!VW9f!Qs5xZ2vJqdAr{V|P*ZNdzf|}RBT8|t*6Ihl zkAKmQCwPW#7y%klf4*8byvx%_%xSz zVOziVA0F_p3%F7EV-mAfXa4vXVO_aGn-7ui-~UB0d*ANz|LZq9^^{6{$|u$Z!Wv;d ztjgjvEa&VcRUqdLHz~dr#O<2o?#-?kvRvzaWm@KNm}Cwm3dN zmEcB@mr9<`nC@Y6yVqh{Io|`QG~8;YWr>mNRCa^(jG&uky$HjY{9?S#J+i2$=5Ln$zLh0&h4gVL z8wpfj>oaYoti^StnVI&&zBrMsxmTtO%;j?ee*}nzM7F5;{rma*c+kF9_~T9^9x*Ml z#?s^ejfF){yb5E&_rI&K`OQ)uoAZ#Xzbrdv;UD)g|5q>HCiBB^Py7_}AubI6 zp$jy>%BR57z*r(`QvFbKG#erVz0FZPmIyQ}+gN=CRy&k`n}s`6t(s4*>{Q7H4?kx4 z+?!z$U(&8(syw^;~Rm*Iq&1c5+D->%ZVkB?3f>id<4FIJYd*A*Prn&hGPJ z)cDFXi=Fe1*rLti4t_x#@^44n0n^O)D`EBAHaxQGX@r!N&3&}Xye@V`ORRSJ{#1|0 zR(<($i6R%SV*e6`b)j)^!JAypEm_VjM$Rqeq~iB8*#EE2Ej$^eayzvgy7cc)2R?mj z8K=unrgV+f%A>gTgm3K@sgxjV zoT$9Lgra>%jnAD-6SgF)7in8x+qF0Ucf0<-+x7ooyOxQ=b}dRiRL1)G^G!9IAO$}HXat?jzanpqa# z=j_b7AqTud|7dTIueF$ou2p!zoW*&HmwoCrzYYbKEjdr-KG2t__#5AU>||S>foopr zCN5DZe~_Wnce;5WlU;HtZgSx!`_BvfjqDcC63^NW?f??n$S&$9;8zF8BID$)Fsd9f zVe1a?Swf58M7@o>A_lIUS39{i_cNUE>repqVjg=OLoZj z&|DWE%Ay-dee{&q~bQxGAG{<<$Oh8Io0``Q;T?{0i;p4ho*@1Dr~tnC0kp;Mo-m< z{hYN2K!t*==X6U>OYpkOoF?K`3=6y?HtW%r!~9HRdvr2ZvjaI?i+2Y=&3NN!0IOu3 z;e!S#jeXIT2G4ze(G135-Z+0{@I1lnh&$ltNmx2L_N7xq%4tw5#ff-`gKIC|rG;$6 zp^$b`#6A)AZ^9yn;ES)L+>{lCsQF|r>NR&k%e6g6)j?q5DQ^f`*@Ei=?gZgTE{y(2 z5HSK;`YA<_MJX~Q+UivgSu4A>Igr|faJXpRDvC`)?A~#lIlz%U{zL#T%HtOtKJdqK=dNETbjQe(c@bg zuYKfIx3;ZsM~p`FPIDa&NB$%44~QbF91p%X#Oes%7}U&gc(R!i)ojEjROV^@I2~KO zErO8f`AA|0;_azcz3isgcB_71^%&yY;!ZZYiHLSI!46snVY#CL(UYToG^0B=zGAx3%%F_ zxgzdDQ-u@&iB}GQO1@#Ab>sXRkiQ!S0aKh`8~@b-p~+i~fo&iUaSB70M@u51F$y6B z^a#p0)#gE{FnBahap^~2xa6rLsDeXJ3YRc-FJWWJE;I_xh0Au{f`}<{cv!CH(1C=S z8>nMf_)|einw_!Ed4$JrMc67t{mSiYmb!^*vAFkGoD}aGQvAc+-cJLhQv56$1oNn%CVY28d9Sz0~;w12B;)K4=Cwo^bZvGN$`?|tC7-U5d2oE0uI13q)M5*6F! z183VAy#)&1sFEF`xE!x5VVc?Oh&y2Fzl$^`CySh0zw)Az;$8pk#ryXvyKBo;)4xA$ z%h`hzrc3oEM_)U`npYaJ^~e{N;@SI-a?(*Q8hg2AC0fx<>(atli7idJ+nNq@b(2xH z@p(sNk2XtRI?Y3R3z}d1I>mY*7tI88iQ};V+!NfUMax~an4DEa~Uq;PCRdvmkcIkV1u|*>Z@U5v9vvrD)0;Ir|2r*9EbVrE7st zS%Wn*Lh9T{XAm|{a}|@S*z!%CTjU{yz)l*GG+jBlE21N(nayhEU?@a>Aq|iP7(+b) z2szLatZp>mwV9T=veO{at>WXT5Q$gR@I4CS22q1+UKh--4j`^osv4u^=3Yl&s#^k+ts*Fn zc^`{)OY&6dRwG*y9^3S}S9MF?8%Wb`a2 zh6iV;D}xdmwk;Yl5sF}8`QBp2SjiR5*+Og$YYqThD{@ePmUgfy-y03Go@4>U>$>4R zQjVSnM3jVhpHCk9+I5iUNi+&x)f0_%{nCXlqznnTBkq8sXbh>@GJHCTcG370?`;$l zbDK528nT0$sm2K}UP(?jG}42xcHYqJ(9u$hBdpLwP>Hg2ZB4fQWK2uy_>R{yjySO051#!w%Q2|bX#G{};bS@JS#@{x1a5aztL>Qwj^uutQUT;K-z_+2zPa)9%_XP03*hbynAyiHhzI>>|5~ z$#-_~%w_Uv^KH|T87dpXOH{Rji#i#|C7ze9vSD2CO3(!Z`nUC&cIps>HyIG2?nHthmm z#r$>uwG&6?j#95IZ{o}ruMrE6Sm=UhM4=F)kE{^Aj8kZi+2Z#u zb5iuMn2CD#X->)#s#yLV=A;*aA6}U+H*)c!t~OlzIpmf?+yF;`mP+7s^}EjvPw^qp zauMw#E!AL#ID{<;xCvD_(1v7^`M97hYMMy#kX1XVX{0sU8k%P@;fZY-Y4#ei4#C=7 z)hOq0Icu96=~h-mOjdWf<=$xe{RJ^5B0?EW3Tcj<@aX^8448A zQE)xp#v-h)n%vCui`~gxPO*rGsyRor^3pPR?*L(4@*!w|sIA=UDH>X$%U8z5Bg;ep zm1GXYJ5rU~JMpf?fUH%MlD!w@??60ftmO23AZ?W;($S4ZQC~}KX)!lM(Zo_s`G(l? zD2AXrCvwOVrsCE))SxGW=X)PK&#ah(D1ayEs?h^5tb=bf&WH;e;tn{<5RCPcxw=wx z88SjKyo!%IG?BfYWwCw77`t6%_wPd%SYkYurClqE(7ezwn7XjlC`6n)IWE5;`gzvF z0(yZ@>5{no4jSM&ezb@R8vq-@WDmrLh?UBnc+#N}TuC8{yfl~--4L+|Ek)BP631kB z1c;C^HmjzvbRfUeTT`XPxy}QjNt{M>gy667#$dv_i@LKwh|(W<8*LlE?L$$bCW@3e zjEgP8t2D((ve~cCJ(yGxyKod-F1E4|T-Jp$xd+)En(biC1ket*NkIFGt=jb=4MMU|nhu(LLAM*=d|{ zMK;78aO56bzsctMDuHOQ?ooWIRHR`mOEuAnPBbNuEK1b`4}%9MnujKUbFxj-^+uUz z42LPfyWDc(2KFgh*+vnvh*PVw50d7ID`#idhn5{tM66!jj=0>& zzDPeB*OLIA@8UzsDWtp1^ZQ?qiiE0Jw(_b8k-fbc1W(vwglL07u4G4i{2d?*%wb7u z05RhOo@^VUMT+c|EruMWL@>7)If<5D_I|%?P`$?EegMOYRp0uwgIrr&Z~sV0~m{^>}z*94wFc3rf2?OQ7 zo&a15)MUvGn6TjDx(_QFm{66!4y#B-nHPw|um)bCe72*YjYZSE1*Lg(tVG4;ne3(<>F!3+?$%+JsWxeXb7SqVnb>Xy_#9oUo|9zb@ok8~l5x^&5{q6I6CEqN4tSbtgA z)lSt8m@?ABiXs`Ye@+b|#h+W|r}$(xKWez}n!)li9nF(h>|?azKuitJ)GQ*Qo?~;jVMb!(?bwVn5?hN(zU;?#3kFhD zVyjWR^D4M>hww4EcT1Hl`KtL=E>y{SP1zsDp<|(tuj7Ei+EVbQJHZ96`19B9XjFsP z`=L8HZ+WhkcB%q9OViVwyAxtp$5o5SVG>=wBBq*(ii+DB3S@WAhvgV#^~YDj&kazp zR%pf2fIVOf{W_#g&fOI3Iu3-(4dFms7m2*Bi*c|bQLBshh6c4Q#R}#!4pDC?QnED? zvL(AiQ@kR&Wn?FS{ooH(4W73Xw*!vMkaSMr6inlj8Iqn3QS@E>`7SCvD#`r*a63+N zL@VhF?k~4h+p%S5WxmVJMT_-W*NhbJn%}*`@&oIZL3kvcRvp{lk!*qS`_i&^Y{=3N zO(XB5+=6LaPD;;a+8SR3J;FG^Q@7zDqnYC>_A6CLb}g0@RY{K2q*ywls(ofoJ4(WH@TLIyZ<-)>(|B?ZmX+CdgYSHAn}AQO|1 z`#)Qw41{ah*&2Dw$m9z>4-oQF?lIdMXY!Y=#vO3vkA>)HUm}3i5p3uwK7>WWVKJgl zHQ1K9DJ)ip3H?NSm>*IT>fKhIXZ?9J$#Z?4z$>_67(|aj2O81^kzh-0fe*yMl|pq_ z#6*T~`1FnlbB(cF5i@>s{?KB^ygb}nELeaqV=w`7%!M8;fZK6p1qL8zR6GH?4+OBa z#C4JmrfaVVUhGn=1~HILFU5XEgN$PkDBvRCri{$2`;L7LZj84^?kqyKuP{y#6OV~11c`oD<04XJBYZTGC z;SjgiC~Ji0vd=Z|r+^q`Bb65I3M7h)NaNQP7+9-_I!V& z#3IMPrtu4+zUGUinv`p3U15})>kJE4%`EoR{78wBz;*Nc?<6{uTZF%eu~;{F9?dB( zfF04-*l-{;h@zxmDg1$$dY>9HPK|CKOTC!u`0(2r6@}h}~=ENYe-0Sz6RS zp}i*H$Hl}uoX9f%u$pj$66L`TDJguqch?~enR{H#!@41It(q6Wj#y;Rv=woPTn#J9 zYHMtQ6I#92*m+bgc4@;7i;4i}xe(0-8OAPcO|ly|#HB^yGIAjMWco02?ij$}QF~K6 z;K;lP11tRMo65>N1|r$J`18F-bdF}eV?tc8h%^AHm&l6;Ky4&So`A^=s>5VwkQ&)? zQ@n_pMq#P((G$tPjz_d+u9oBLswaX+c$25#swX-;AiAY0k<7rwl`8DPCnq908^`%a zKUIE`+3<&cR*dD4@h&xDFX?=vcW9^?@w4NGTTJ ziIq_g0PH;pX|U6P;!-u0^+JOT#FTwpd*vwJ5tB@n-gaDw&+xXiFHPklrRa5v0TiAW zvIWR7#Z{#RkRmqbYKg-#SE;R~b^NT7bB1c!f;6WVh3gIF31E3Mywpl>M(;~S8 zpt!0M(N0{$QoMflsxA@{qDaWu)+o}DFves6fdEe`JMRE8-nnI3ryj>M@HXGjFs`Mf z&f5CdX~iN5#&ww1pt0ww?T79}O-H^{RR|;6{cXNsK!nh@h$4GwCes_D9L};%?Cc5W z%xL8X5LJykOe@Jw3pQ8=z!uT_{2hSvq+VlgR1fFob(PlGEd$yScfg4O**q6#1|$+u zUy2Wj##I8w+wVB!g>1fi60l?BZUG{yIvL3Un1%|?W5qrx)i|w4#mOP40M;h?*^?w#%`1Z4paA(Zft12u>+nW8VP)JYEHM zd2v6Lor@Nb35eqk#L3j_yYh{=>Nt@Oh*b`V)_y=tIrGxJPkRtvov}CiHtpDvCcLHX zv}?MSepu<4`nwvg8&jk@N5~O_NP_pCa=3K|qu|z)9oeYwF5X#y=#Gw@t)?lW7RZ@> zZEB=C+7WlaZ#QzI{g?yE1F{*oI_OPq3}+&^8@os6%Wb^O2@xENlCmBwenRn^u#3(GY+T^pGDdl=z@Gy)* z{w}uOrt_9pN3}y!#s0Dk+A!U&gA0p<>f?k?{WW+5efwP`KH*{JJFcnt9+>a1JD;5> zVY8$b4=Yu5u~fj%oZO>(b=-LpZQ6XtWx3ksxh^R)9qg18bJYkX=RrYa6D`J~6@>M0 zMQd`p7T*mO4KMC$zjqh9oG11>*UI)gVlVb&J6X8U75_PyrA~;iv@6`zx6EVs* zBkE}jv6yBmk0i-|&RL$SUqq2HOG8uM%g!|4JI_DeG()s3qUl_EBN7akW@ZpH7JDs`rpaBhHQI^&SWTmo8Q`@Xb}ZG;F%$ei zo6DPvED@>)+E7DdLvS_Ud9f9MX$Fz2lzyX5wxW}>)u5%;G#cy$jevS_&en3vgpEoz z-@6QTwL|AIPH@+;3!2d)*ByJ zo5%_HZN(E^U6@(3OFj_hA}lQqHrHgk`f{5A5mrRYZ3eAXREKy&Lt_(#Xz!nJbf=yh z=DR2vEmZT}qljg*03wSi9KNf*Ui?0OBI#vKr1oX}BANS0q-Mr%aJ+3G_U|W7NSq51 ztG?Wr5jqD)_jl>R3A1o=IX9k4B36cDpwY+4D77IGQ%9KW1qnr!8^s6d`J8dS3;Snc z-in;OfA%5*rRb~u-hBwO*OT8hVG-Go3co8DZk%c5;Q=28Q(ej${e65NS8kdlmu7g? zyuPe^6n`wP-{rOoH&HrFi$xfvs3YdP=n&i+nvT^3YpYAg$Y?b=DJCh{@cIHhI6i(`05MtCiU%y!qJhIeNe6SFL7UVpuiMmCGm?Q_f=sy$f;Vr^Kq<{1K zdn~xxm-hMYMR32q``3n-RT||M`3qkTVI4c6SXUFH^g<|-7IsF4QY`w4hVtG4*boro z(g0EssjWvt0s-PNf7@tStjY_Sk-e?4Vv*twG+?-i+-Ly8^$J_i0+4Q8vcw7haWeQH z#7W(Yn3bwuSHSShiGf_AU*`<=LW#qqk={Z$LpM~3G(}|gbepyi3FB04cjQneFi*cC zCIUG`yD|frpz2`~ux*NjTQmc}J1e2Lz79a)?!+AjfasQSZFRgaW=!<<4YsJ51%Iky z5gYI0$zKNm3k$AtOh0>-BO=;0g%gjAsBLOUMtf7cg!FBIhKOVbXs|Nd_y#yKIgJt7 zzF9Ur7m`o$AwH3mf~CKGCrE^7b@TfpB4N}clJaPltRQ}DrItHKh(W{Muta}DBj<7w zh2@h9HH6T6JPVLnth(X0hRD{|u)DgnMT>+;3_m6e5{#4WN^@vD&c? zu-ke9;6y}nxt%yfa$!N`MjRQ+VlH-d!9p_En>R$!FOZllDfmaBNg6;Q0EY;222eez z){CVqE7IPR`BKTIUcG5o6egV5eyFGMS$eHD)yam_3qR#ug;Pb*?w%(sYrLPQyZ7`WUE zeH)s*FoO2tl3NcvG3treYyq7&{9pkxm~sLYO9KWJDY)D8;4JJ&m}C$O85v8xB4YI; zJ-4KYHix^91$5bIa_risI@$JGyN#2`Wf9vjQ)3Wmx3_GELtk7}MBV|?4yNCRhCpB~ z-dQf))zE|WFu);YqO4yBSa?Xsr@?H10D~i{`?npy?O4V0*lqz#v!p|IQ}W6h36Ogm z(nd;fmo1`c+8)7A{j4Uc0oh<%fq~6{MZ}3eT%LPRX|kl?*Kz3ReUa~;yv zFbwZ8@@5%COwk-{aGQ(44GBD3>>PU0a*HZB7erx**)8zJlTl-3ToH5RA}*P}i~35$ zsRJohG#6~GZ#Bv!q+0AGaIKe!V0A<6Ebm}Rk*>+!aCUyU1^s>Yf-oC`v6`)iHAGb* z`Wsag>mi@0(m;L7u34;Lu^g^fs&K_B!7hImTIUY#hA85ghG8_b(;NC%HUP|LrLk}T z$aWwBlmQHmyAfvqM^;PJE{s@(=xCgkG8CH$upTAb8!@C_14d z>B>iq)4kUdwly|IwIKDM2vUAk60AGtA4aME<&5b0`=A%y$~tEb_WwCm(l^y zNcPceQzCwIMqG!SK+A4{Pej=viKFF0V`JUd1C1h*I$i=zjbwN(Dc)+tCA_PAv>}pc zh(IK_J_L342+Oj0m=me@2nGCY5`*R(=JR0BLP4NC#tzD zCl;w`oa{k4-eJUkW?4}$KFqiq9Y>8F7 z7Qp*WT5B3^RHS)82bKXLMBUwuDo14OQE+vml5^xlrsgi{t(b1}%<1)-XX$raZXh$_ zay-;n(Fc9DHL~sqK6x2QIYkK_?JOeE7W7W@f+)krWE)bH7wMh;R-<&9pcz+Egh`1y z%>v$Qj1dEY6QE%G;LPj^V<&Q4`Z{|gFyi^?Fngq*%5H6S_z2)}l=HP8S>$)pLK{R3 zjY9NS#ELuMVOhLPMA9}F%Nw%H6$yncV&{cDxND2Zks;XGpZAex#CadNP03^?Ta5;A z$nz;sX3)-W8<7mm=sYEKjrSaxI(rN0qr`&25 z*NG`$2*UA|=!+MEyd$|N55gNBVZ9FUMYE)%gZHw?WmAEt;`W&RD#*?&=WJ3{x`&cAf2N{3_WUu>l>o%60&D) zW1}w-a$vBSw!y(-%+;K}j*h0Ys;^<+hR1WOnnSRyiWrh{+cy2E=%L+x^rOT@H5BEx zFH303M56cP)UE@dsfK(o{R}*`=r;S(&w%YydXcvElN3pXpcgw;!4uDYnJ=X(F9I1E zTvD=BiHxU#EWc6}uqJ&C@lF*!#uzgisz}+=%5|0TD*JjW-4O|tPj=@UBC&*(kL-xj zN7>)*ir5L3u~0O|fpeuc_jMdNGg3|IFb+f#_iJ~jT1H+jf8C~&m+go<;K<}U5#PTg z%ob*bR)N6uEc+^ER*9h)fk0wDDdcQU+?UtZ9(~?=K;}RR5w<%dPJf zPK_RSf=JNa+?60K8p;BmOg5_3?l*!OSw~?|h9IovtIxCM$wfsJTbmq!Sk+!}?*bv^ zv8DR#!=gBGDdHFSWUoXIgU6ja?+Hqu7x?n#5ni0(f+)k_^;vdD#1Wq0fHy?ZVw_$9 z?TGS&*j*dq8gkLUj0E>lYr=a8YhnNx>GuO5h=|A;VQ2{IVpZIY_(1RreK6k7!WH#6l@g#WJzg-Xmmz&AX^<-dsjXU7P*|;<%hQc@T4TWYYK2*Vk zJ9k^E;p$;<-GYyLThIny%8i^0=DX_2(KJ&Ka|FE>W`S!b zuxvcs{F6WRA!ogcuz=#K-sza8@YBZu=|*FzMg7zF6&VOtNeHdLA=Gcl#b35Y4tx8j z;D_6s?`5rgK8lx!DgKSBMn8kyCmi+~MfWRJg#QXIo~V*7(gLa@2uUBo04GFe?Z;{JQk#`R* zVy<37FoTP5vW}h(?wX8;C#4t+MX%S~RKAkLC2kMPXf2l-BE9C*)XB~ubjA0%^8`6N zPo59R3ND)-d%b)})vTw{IJZ>Ae{US`A z;larus;6-%8{%Ss6^OlbK-4h5)sDkBXRu<(whwIeCa0bS@Z?udfWT}fyS0N_5Vz~z z)`vmdH4P%^0Y!Vb^;(n_&5dJ8QP-R1EP&uh*jSjRu<$QAU1_W&4M+zXG5c}tI>4c8 zKn}YW(AvH8?$$Yi0X71_?uf$QNicd{M3Y;;-*kkFxSE{_Lx)PL+}G zmAGoRZN&o&+18$Nb64eBHI}NW8DrUhP_u|QqVb+;#s}4{H4hG63GzR`j~g8F7x`sK z$m}%Zq!M*MM1$61Z(RIb(B#<8D%x&T^g@uEXufMgG!^)^c0957oWIgX4j)2&SVV5O zn-6zH9C?F3AoiC};ti2F62z}Gh&cH6+OCL@Q_F!=jh~k9q$c9l#AHV;JHquI0HyMD zT5VIJk@V8{;atqKP;FW;3scSoIZS%_x-va!0ZVa*rHOGGK_?8k0UsNGSY;v|AG3p>jjA{ho!waX&(MqCqjLyTl*=nH%i!@0~n0Ax-Hqi>oT?wNa2 zyL;xhg*MzX?LwPpfN6XK91n-c^R)l|5ScZf!&7{)aNMBKP*Bq^qJFY>oK@(;k@uhlbG;>bVZcKTnAz#u~o5@$f--TwgV|S1fDoHQv5(9;}!4q z2C-s467GEvrk?Tw9{{QFb)~o0C=wlhHKxX254`Xs#O!1OV_&XdD2NB4j>4-EZ9?Hcr*sy&aJNy6fcch>3^DYUh&2Yh2u}h?z%(WsyAN5V+7GnC-W5 zAQ6_Zz;YOe(CE-%J;?DxkofDg2C7W!IJ^ObZ^UxJkV`l~jXC{|T(Ug$n(Kkw$v!r6 z4Pc&u3u}=5nd&_;{C~ZQXKtq0{O{sp^CM8eC!cCCcCBO2HD5zszz>=nK3jQB6vM z)R$R(DsS)Lo?k0(?`WU4^5BBH^_rdq*I#jOE>j;!1fiErJECZDn8M%2GNjZ|5G`UR z08X?k;6bbehDhEbHiB^bc16TR%rEePD1?w<<=V$iri|3hD#rdrHd;OaWTIR+06a&= zx>Fvm;YB8I%LHu1Pp+~f?tmk^;jyP3lfTaZyHdPqO2bk7Sz-Hk-fLxg+n-rv6AhL$ zv$zX$YBDNZay_*?Tq*E0oaXnx-ro@)y)ECB6*f~IJ8Qy~?BzWvc?xpF4nV>qE5}}w zC*8R($zq74n|)Z?{$L&&tagAN9w39Qkp$;p6DK+jL`+0-iWV^w2orJmf{4o(QFRTK z4cVhqTVvyz_ha6vaU#ml_P$Q^v(m9lB&y;wSQbJ2yH+O2q1>D~nw&RMetoUoM( zMeRLpKUXe8E9XZ%Ro=KyS}7mA4wuZl&v2X!h){hn$qkW|8X79(vNXs7nnJZJK#rw| z6~?tivIVwgr)@H*j~crL9649V{qRes_WoN+=ZdELEds zXib#O#=YMMS|6NAq!7JjebAoogwy8gW&2K4@KR0yH4$EGdUmjCZ{6PrC(w$CN_zn2?y(5szd?URWaU!OdI5Og^^IucXS45EuAB%_S5oM6lpHN4^RT63<@&wzeNJ6^jOg{oQAf&`t`_Y`A zED&?JWQRNuwJr918#$8Xa!X?oHLz#8uFsIwhN9U%WCD8lY(`EFLSg-{N&0c*V3|qf zZ##0l&-vCYxdg!>F#4sRn2cfkY!IW1IjJIZ?k7gZAt$@E@S?c!s<*jQe}{uwr28RgoQX7-c%K8ytfvy;1@C5B9d>AXiv8m ziD`+`^Nz^nf8)09(jvy3!jZZeHN9dX1*xe~6cDbVZlr`r+VgcvAI>4$pd#)IK$y-g zZ#o5Q!oGeKTsTMeNsfqM$<5?$&DoNhi6zzjC%JJ!>`K3_u|yGNt6fi_$iZ3C0fW#~ zx!JuTY>jJdX8_o;I!wKGs^7U5!fo0HgD{9Y;K(^}*9)ig=8H>uAGT9Z@u!W90bfnC z8om4)@?DL>PlKCi&sHbPjhv+BJE;naGtKl@*2#}%M3|k7fWS%F*5`v}(rl(ePareW zYg`CdFD(M`8$=OsupZhFr^Lf?hN3wTA2B1Mh;*${lLB4E44x7Ft20_~Q6r10-gX*f z4RT0d=RvRbBFtt-tg0ThYj@;`MPIjwJWle?bZJr4_WUuxW9tjkBg6C8aR~an>HA?E zFa*l=Fgr-G8cgo8Pyz-)>98OJ7f5mK=VW1}X6I}_2)#y3@UQ)dXu-9e`jJGw;GWF% zLxjJMmF8`9b0n_1U1fn#$Hcn(N>#?;&!1LVU`;4j->Is^ImkOzQ`hF{H7C0~MI$X3 zF)tW^bEDb-j{Gx*mgAQU>zx?rH7v!S z`69gzrbST$WZ#mQw|Ba#I?B!2BDxB}HAl_QchTo?g&p7Dy~55yrH4T{S0v-eZe4h3 z?FNltVtGT8c9(0kR~m(#3}V+f%xLxskwwT8`&px3*dTeze6=+uN1n2D01%_U2MZu4 zI1v;L4Kl30_i*i1#fbjYR10J&I03LgSeqJJCM>$wuQSczuBI4vs(r`ecWrerV~nF( zx{W2dtP(M(c0~B4CQQBz0G9o7R-X_%>@R*Iw#jITZ4npv%JQU6yHG^^6t1^!ks)K@ z%-Fo8fqP6c-`#rM9Qp05MQqsSr*iFe=O*m1kc1t%^f9}W>tO6+*FzKqrTY_LFqe%u z12}S>6jdhtay8E*3mW}Xd`N91`EhYp6N%fOCriQNc((GaVZtugs%Wx6Z`rdO85+2w z8Rg~*CZbDQYko9C(XJfg%b6eyQLxGxXoKru3R;Vy)VomiuKF9PDmnI7%Nwo2@wwlS_@n z8ucMrL~KY2#Bo8?urma^*-6nHTRlfpqq_dc&bf^WR)JW-T11ir*Y2*Q$Rdq(Sv6Cb zL+Y#D0_04Q#ty_Ml1SuW0iwmdxYACXNSUIS+9f+)3Aj^z?Y$pYAa>zIO*yPUYyr^8 z4~=*1&+j8cmH?hfO?N8co`YS}fyX)`m71mqV5K_orKw(#B<(^Hmgthp*M$^ie_`lg zNYNC>AB%n5&q-4?#oI!foO4Iq0Y{c9!h`V6Z*K;vW1bO5@$uT(<&Z)^!*Mv=5>Pj_ zUgsV9DYpo6LG~+GB2Ueaz=qYfuM2@V;w=xPF|PCvma48kVM7}tH>=&Rt<9^mRP>Hy ztIrc6BrZMQ=#c4@0tfm%_DJ#G0S$_TgNZ@JK(2jE%|{n}5^2k?U0*KiBCiexAZ!2J z>q-=~SwSb3*eKgfA;K?uC9S6Sl3;2z({LuuE%C9`_pLmM<# zyhIRAHZ*7;#`P!KWKAMRk(Oyk>ybbz@z$Ehd|KAjM!kw<7%^EoJlcUOt!F}fY^47S zwnltInlBcRg>;;}OjDW6n7v$yuMVNPS;UHCLGf1_MadD?>@6ux=lU=ly@ix~G);+U zM-tI`qY**P;rC$>J+3#c-)IcN?s$5fh~o~;VF07OmUGAb@C?9>R}7=zdWR-Wks!rT zh4h&6y*s(`u4I=*$eQQux-3L3$NM~)k(JMAH+ETsrn1Yvu~fPu&p|f;RX~Yixf4g0 z+aAoCoqU40T=}pbCZE`A+yO@hg5$3oVP78@Uv}qD@yBNB37s^)Ki67W?On`w;x-0k zjs1J^xT=`%y=v4n(=KNIxZv;Q@odZbcC;bjO^juZ#SmfN6a3EtL`Ko$Y6nEF#FDpd zU>qJ}uowkbplXLt=9Kt?cAUtpm~itSfFYM8o-cGn$N+qQZK&h|aB&n|jT$_eOqyvm z@baGHLXJ6$p5=nyKc2L&o*YWNms)XSsU_LqImTtF)g&TB{Zp1u$oJA3PS*Wd`YQ&! z7XIs8SUV!$3j(+FV|;&?ujuJ%DgVcVf!K_#$bh`b5kUbzY57h#e^-AC%f=k_O^ti- zal36?xIG1D)dV{3Sgh~Vkfnn8H{|<>y-_@1#3GEZ9WcnbfM10L@M@5w0T5(aos9Q3 zNP58qt;Ha_HKDl-Kqeexaa&bN7{gd{v9xStN;enw{Q``IHx^8C zB*ZOGm8QV6GUHr65g(GHLF{nNDZv(>t~mmISmXnJ5g3T8E2QU%CF;0)hf@@Al`doR zHPC*&zem=vOpk1Dkuc8J43xvsJ@`@XS$E-gCnF-WvJuqPfj?~om4PNY7Qw)GG+E&p zN3mN+Nbk^b#2JVWATaRj$UK0EAA34rBsM^37Le7p-43$PGM|2_YU`9pTVz1DgTdoq z(T#uURU;|+lL zI&EE5e?NfA&Gk<8=Lq1~QUPg}S7%eUEXG>@( zTxE-2_C%DGg{9JAQ%r8KQDp!_T3Q)JcDxY%c$Xx_pX>Rj_+SO38R9T`s^OyZ{a`8J zqFI#cM0>uM)JfX3e{Y)CIGXRB$MU6_vGF4p-|e&cJlCJYlCW2UC%QShMZVE3i&Iv+ zUmE;10KLfu0wDVR#eR;0n7$Hz?USUAX98;1jVs zeSE+5Gm}HT9Y`sq=$)~tv5>_omFtjK0{Y}^5k+EP=&q#5O4W30J1MH`;ZkP-+#j`W znv#DKq$=Eo3xif{MJytL(TMndp)ueFfJU<~i=fE{Tl%^zq=KsX2kf#4aFMVsi&1&6 z#1{Krc~qm4c&a?MSL9FAWgOTD7JTb+Dw?tWfY{_r&3D91;L{pxcd}admG9ljqCS%d z_t2dztk|uU58mAg!EV#62SH_e5?6iWD;&?L?MEoiKv{4AO*W6MKF^(q#7H-aK%s_h>uqMH1raGP|$awNu z-6BCblImoGC!ARjYroOqH3U>lh7dH2+*V2k^DM+@K47FmHZW&tAh0D*xi=b$i~3Ap z5DPJv+MW0+(u!Y3Ws!$4#HzN&-i1>y7it;U-dQUbZA;xf(yQDh&{ z`1YAw5Gzp|>~Qj6seTj|Q1DG~7LbF6^RYA_`-*^lr2(@%k#TK}utU|t3;@N{d)rhc z$a^l=0p{hzWxHQwjS{;0x?wb>pA3{Gg%m9{9HvYd3d!GYHOjm+%!Vf7BP==sP$wNn zJo?X9kXo>lUEPm@Ct-1;$^eeMNBTpIz;D7LfsRC@3{ZP zwfvh_y#LnnNbJG4$*Yog2Q<;M^hWLztxuMd8MiR;hRi2>le_`|n(&{i%!j#<)Hn(* zG9Q^w@hJW5BmB$}F)tZv<<#MX4Yyum!z1hRfJnLu9v^P~lq0}-mn~u;gsDNCEovHt z6C!Q}cv0M1B=@nLzZS7IIkMulMN(g-zEr&M8Yk+s?L67y9ZskhQi2`^!IBb$6}f5< z7jZ@?sWF+=5s11-K&V7ERDUsJ)q=7`$ZYr&T*Mgy?p9xH^I#w<{ zF4fyX6gft8>T1WL$^qNc=h7>AGD}Pf`!a0<(A&#ma|g(ok<{kW0Q8W}!O$@L#V_+Q z`PGiN1CESLG!0oMCsWvAl%|ZVHf2EX;zNkix9j>*jc7OH0xNR-=ZW@QfQf)2>J| zyVzAyuAo%MQEt5eYwD!16GWbtlKPDxVN5R%+#Szj46B;bhWM~P4DR(<{6aDzf~tE_ zw$Y3^a#E#gZ=BEqX&K&~>KI93YkP~RhYWH>6b(Q)oVSR$GUL>CL#)0N@@gUufT#cs zQE{)4c$mcUUTP!>_LTTw?x~s@<;<^*;B5vV-NpsRG!?w8-IA`vbM;PmAd5)qm7M(* zkr&IB4=rNKN1l`$B5|u#y5AN$Hfp4I__}Nel0k;&!?GdDTjV9*P*pKx@pq~k(R2>8 zFUKC%ELTJ#2qOF##F1(9hB#Qxj<^Gkyr`#|%7=k0iB$8Z&^>NVPw^pCNht!F@87ZB z?8vRp_eXfjsM2L=R!*g6ockHrbZuH#9hmD=XgE#Iiz&{k~Vog(e;)Ta8n7 zOekNWSj41TE?h~GQR*ZDHl%1s=x&Due@YDRuQkeNnf^nKNHwln0EyXUm@V3*1`x3+!^mrIDtd?8*1oKktN13ry&d2vYGclZ#$q?QLEHeR zO`h>csUIer?@j)xIf<%SWQ9J(_sws4yJ=BH>|U*asx2 zCl2vgEI^JVjhSzn5-mcbR^FNtT}ik`vxrznVUu%1Y#ik7&|)L}e9hNU=_L5eb{Yr* z+lfGNYf*M{Yll%G?WVq%8vrRjTj{1LOoAP8%`Z&}n<7=SopPB`;Q)Q_+{&himY6Lm zxKh`3xCWaNm9xfTo62MlcJ(%nA!F><4xN+j6bN^Jqp^}qqB@KCddH)&m^MC!5Le$3Z|42j&{MUcS@UQ<6N%#N6s40J{ zv^X(^E%D?K#+N@|(xH}eZ##oG+J7>Qya+tse{J8Nf4sqeUBF7=ACs7^I``*)azj(y z6M^(^^Zh?HG@B1v(?9FfyA-w@+_T!DOPGalu$Z#srx%XASPwPL0X8VU78F%q^I2*- zqkpULf4vGh<|u3DGR8zV5z0zX`#bxOBvTNCc(x zo3@sVTwVL8+*|es5*)Y8eW$|{#!e;pPeQoYd)0wz}kE_<%K(lIvGrPB7^51O?eqkIgjT! zyjM6rs4nL*ZodinAQV1$V_iRZf2MV=&FVgr@7N4iAyXNm#S%W z{Z&>ZiT^)k-9IZtW{vJ+{h?5H-D3;Bmi^;3%7@bb`4rdH^Vh3XjB5SXMll7B0`*nE z69eig@Ser7CbwyJ#xSt0gGP$9!-yEh^! zT5Gm@<3Sv(Y)9My^XlLyzVnVs`(``uMu^R-@pT{e_3Qlocjo!Djat?V2;nybtJRzV z?^@o>vAftcj^p$FUA{vin_@4Il^7xHxALUpN!;7)2wwB?wQey2CIWeB`i}{!%wd|K z!HhP*?}tf0gJvdR46?5{wY70R?Ilz)x&nX18p_z1+CrSeV{C-=R)~d&U)59mQTb|- zCTQfThKx1KeSuU}UIFJn@nr8rqU&7|J?8pp(l_sAe07?-Xa@*~mvz={Q|vM>8E;LA z+?O?r(u4Tx0#IuJnVi$Q<cA11k>H>dVCxj2q>m1G3 zXHS?K9XmpNT&x1;UpotC!tZFW+Pi>dBH(>)Hv&(c*2>nSFRi4e5jF~Bn65^uIZFJp!Eg3tR>R^w!@Y4QOeoG`@$fJmF(9f;4F?3#DtadTAR zdq?D8%eh>MFY1C@?Og1DpCm7RosL~26hHMmhlor^dr#G8UlFpr3kD;S-8=VaFf^h< zTgUUkP`$4di-9DC1N;C?Q$ixDMsFDg!i5jXe~j4rTt1lNj<^GkTvG%WIUmcHm$bn( z-^HKr3})9P1@G^}keL(!VOEb*dR#1MWVAIA(no!2L6eIHGBg4~5YZQwrj{3R z4Cein23$ZeATu>4kz|S{Sl9t1=dAGnAjMicR3##%|LEu7u$9{Sl{j9)(@MkgsulIz zxgcV@6wFM$`JcwH-T}C8IHK1ZfQTREgdxYUQ*~q`PU8-STt10gjB!Y$qkDKql-h)k zms_v!j&PdhcPE86^t24e6qLyKUJn>8=9dDr18~FCbVnO;gx)3(?2R}%cT>@{E;}O2 zJh^m5)W}&Jvl}UM=k7J`fa8|#6_xA9z`pC2eixr=5HUZ}tOkF1KfHY~W+)8CgZAdn z5L7gTdL@eJ)wELdg!syB@G7@h<3$1vWQ}_Vg`H29lv11TCqN~J zNr(pkp_XH{G;sCVv%AtDN3AE`mqF}!`qXekB+3!-7Y!*y`R&0u3&4mZRZBdwlx!zE zfVhsxw*I6+7T_ZK5S!0VnJ5FipkJpfWX+Fd7lDX3EfGD?016?BQ?yN$##?i_HdPy; z8ted*1CKQI2Jw-Z@J)WDn)rEBiu5YtUvP(ANwyWWM&CC)@ou7v?R(|249Vqms=SAQ zyN2yq&l<+|IwT@_CATit!HgsK5C?$piJ%#iw|-z7KzOOI{lICMa8ieUL<~Gqx64mx zPI4e$_ix1By&hKjOP4MOKosRnFYouJI9&Dq>eAFWJ9fk!a5_{XiQVKB`F5yq7B0od zcT6^kZ}NP9Y`Am5F7Z!v*Dl=nYQv3pB+=a8<(5-Jjw=mc_SF2G8l(sk>2DHFI-05b z^45Wk_l}1#RUwh^C*st72~mania~@4y3(&Lz8=rLMTDajLhW8!%$if(o5AC`OxRxj zI(isiIo><$@OknsM*=bJREc84YxK1{UF4{;jML=$WG<`wIIPZh{uRe)LliY!hOq4; zzxdD*m|z*FDRv4wD9v3FBOE^zGk_tcmks#00mKlj(#NcShuhU&m5*-Kd0qqqtl(T9`%1WBwTJ>~XzDja0%ki+6Y)lT zm`@oROPTjagNW4xDf(`RnGg=q z&W9I8p=y}E3t}a(vL>F`5Y-62wL4;e`K=p7yb3iw{s-b@up)T*x-5|~@j_wnbE zC-EE?g?8jc^WJzzvtyt)Mp&bJX~07sU{Tar5T{xg>NE zA^76GOS;~H_OL{52RJRa>V#D`E?+*YTb=lNZkkx*6OZTGj8E%@4vpoGou)w)L7$9* zS40f2da^x`L#hRGPa8nVq(wepRE^7OJB_ZMas9aUn(G&??hIlNS>%-+aVoF4oN!q( z5XS>dahO#B5MGymcz|gl6%q>bUGVdbB$3>8rz&vO7+bw{2NRM=PDn!)#^tX>4jyM} zcaI<=or<}Pf@_c$k>6tJmMT#_*#_6KN_x$o&9uR>in4SAh&-qy3y`zQd2bgRr*G#G z^+rR(m^L4pYB+8LcK{S5e!B{JN=fcObtH;=U8$ZIpRyCkcy*pOIr4?GwF6?%Otj(F zE198!U!f)CG0EK*0)d2Fey!^tmnX5vh_)C2ms?{mU~51G z)s0eA>aKdrM(Xf>E~4R;?`;#rjc1Ni~!v52ybYS3&1 z#{m|ic>}d0jH9j7JWD&ENewt2n5J5i6;|v3O^L53o=;4TJhKtdzyi9Q0@e0P=FRHgdPsFi_3K)mWH7U=P}}a+^8a# zOApQ>%3)YvnL>`s0?jzZY-cxM^nzWj0mvy>RhQnIBEl+3WN$Q*+1ypTbR%8~SlC<; zIhKi7dO(!T@6vTWMNJO4z^7oCxjW00L0vgOI3frMtaWL zGczQW4aL2U@D3#ovr4M*T$a1y;TK>Fz0qr!5TrzD@uy(B)!X9J(qEuM$#_FwzB`@_ zRaml5EaDHpFUbXr+b%CH1MT@PSeia$|1Nq|arS#B-w2`Ux8^(Emjrh|H}82L!tiRx z^A6uI*4GE(XpbA>QtQ&+Ul+h{Zs&-Rn$>XXPhq#|{klMLri5ruqn^JWKz(_!?>*%l zc}IBPFjB^OCs+)uAeATY7`eU^;p`Yy{m_@j|CpfcIu9sj#NdR%$LmnHRvlJSdt-FjsY>MWTc&zjQh2_Vj~Rcc%0#O4@jF!rJ4plN8>+ChNX@QSB7*AM`q#?wn@G8z zDi=;1J|^M;RA#VdI~I*tV{IRcIYTU#UhOIi9V_B?U1c2lf!(%~B1%SUnlEGr$OwG0 zfCmsUKFZDl21nkAGl1VGZrcfl=uek6BX&~_^8xQ&eDG7XHi}C%9)qyK2QfQl1#D;{ ztT!GpKhef2=v)>?t(B}FXQhpDg$!7Alw0@$Awv?f^wzSRt+HDB-t^=P#8}1+Z6H1X z-k%6tG625@pvkG>zfy}`$t&NTB^Rsl!e-pbC1s^5w%bTe_DdeHU=ihr_o3!%KZOf7 z&3n8faubfr)(uhEW>hY{kArNHsJDvW#-XY>8BgON@|B#lcAa4r&Hco697GW!C!9eH z4)e`=?G7dxzN`Z{*m);7Sg`;umU=m0@Dk8dvH-jU_1o@nxaKmj&Kw!%g0}Q?&z5OfzBTE*tZwn{JpUWe9ZjF8+Ln`=Oug z-#tq;;26IHq+ZPjX92ZoKDYrGd;^;SoXQ_kBUg-m3VH-ZEeAxdxA9V55Jifcn>V58 z#D(~0`zJ#0XGN@8eqFaeUdp-$+b_W9~4b*G9gNYOtUxq3)Obst}rixhXStI>#h}>lo74kL?m>-t< zqIy{CFLvectVi0*C5K-my^HFMUb;RXk_D^ z6k#yForov2W~lqVKwuNb>J9*5V+5ruhPa6R77w47k>%mb^~2H>fscfnpL3a8noJ0L z<7TI1hKbl~+yO@xAx?oT0c29poHt}>Zq923QaAhv897lEtv zrOVq`A#BB7?M@^=4-DptNRCZj-}XV4QQ$`K1kil*)dFJT{?AVI41sf@VX&|raR(fE znHIl;adlVm+avFbX`Tpzq-YWN}@*U6Av$Y@PQYe~Y>|Jh7W+&DL#6LAZ z>TDsbeVc^KdG#bLQr?LKjC8V5Dx8B9$+kkT+1N$){j4!h@l7*0MO$M+QY5PQUSq;_ zc%fmZF>6$sWD$#h-Co^DspzQ5SyCEdMnbRv+0ia6!?oyYJgTJ|jbvGfwRMZgWmOA1 zVqRHX)R5`hlE9%r6zErtO$5IC4IAlKdK?9p9UckB=Zd8Y7fH=^a_LUO-X(jBsFT>r zwL2M$*TZ^3EfiTKTR-L?MPZsd~JE=P~68Q6dJOB{; z%}sJd6_BHAN>QF`pk(j}rrhE}fI-Xzrr!yHQ>5}r6}i`lO><>tKWeO#lQi;b91yW@ z>&RsnQi{hbuUNz)A2d1A1+hjozOslF2Sp#Rh>fsVF4@kr?YWv8fJeK-QC$Np^9tVj z^_m&GD4CHy3NG4R$86sphMd4gue61q{Ro)ylx^hVs`e3q*6*uKpB6}Td#DkbG;P#* zWJo^VQn54!(#Yfk@r0LeejT#G=30A2XWnT@T&d=LA+6G)oK@LQD_K2Wi4QJ_C9vwd znLR8!NoMkO_IQ8bZRId~n!=_xXQ|Tb&s=Z&LJqYm>iGdtloiq=-nt{Ir+?k7Y$F?8 z_k{bl0AeV+&<+4~@wc_zE>S;S)d>OjLy(pznob z8QcN*mf)3QXh>xFT8P_e$iiNoKOp8RV_WZtgkH|6Tf~YhG_3xDh(%ySI@l3Kl6xlx6>gA;(EDkf4ux$jg(#L1dYR3%J#n;m4NIU<5zsgecRHMGA$l<8R)%7ZxG zQab{8`c5^Bmgmm{)r1#JJ=h`97YJAWLFIUq>UXwmrsw#=~ zLj6uv6&<>Ff1-+9c_|#ai~+P}6Zlp+o`=MlJXDUc0%qwuyFk#^prHF$M4UZp`y7iW zcE6qla(uMjEFvjW@npOVVkGrw3 zj>kqOP;dMC@Iq`R&BFOAKBRTh^B`dE--YRDBB1B{b7RnO-qCtU{&sx>(3l1qxcF6W zxiP>sO;gFfYaRy?-pI&4PmdyjR~CaF5%-D1Vi1GlUA|rsW!J~Pz$aoy5bIaO*ooD> z51Q9-p(H%E0P=>o6=MkcyhCVK+jCQ;*VhTS*GP6QcLKEdir!=aP4-60X6j^1Wcky- z&YftVWI2q7$>WliNO-1$K~vwP>sSU}C1SMG9E2Jh3oyJ)%`T&?PO{jJNv>)Wh9bU6@kMtAW>j_5Wdt<07juBY{yY6m0{@KqGu z>E!uDMbU3`N}--!3uvcPj_9tpYCD}GfQbxj=S+iD6Bx;0X#Rs{r$*RyN`Bo6^9m;< z&Fc8A12VA~kow<^d*^zO^#!U5cnAQ{HMHXe) za?@2!ktdU}pB8p6%3I1}KGvBysT1Wdq?6M>Cnubsj z-7~2eNOrv~>tx2h#6Z@95o5A23eN6gCFo5mLjXj$Xs+I<=6dIqfOAJAx;oC6H$?f? z2w7tgWs|CX<_(dAV5;S9vl%L(yuDpYp`sg|0QyX~0v;MVKC-=BX&^|cugxGvLMjm- z<${Qby66iB#K&;+hWHq6?pt~oZho8nVYq2$e=^CPsvYqAzyJ@41FaaFhWF4%Rxns! zbvTX8Y9iLzX|S>=C+ePMOulpH!=~83d)#_BLo5Ib+7>Jz!X8R8l=+yVCXo3gEe=FX zhtiTdaq<;eJTBV4@>=JSRxX|Oto&N4t+8VxvF9mNfK*2Sr=7;Z9IEWIro&N5_<36% z*mS4%8c;2`FbCp}`NVpc5&^+qnq%B*7#H;MW5X*{r=*!_VZB(Y?M_ z-%c(!&KyF9Tw$6-;lGVOCIXm{S#k@>FPa^FVry2?Y~N}mFf-0Z22l-u>rToXIX1hm zcClg=mC^}-GjCdR4GutV0J}g$zk2hy1F#`<1Qy?ki=JG^{o1QYmE9GK_(X@g^lB!j z+Yxubk{R&F%ymy@z^{cyW$NQKj#Jfso<09?{zl(d!_&D zReI!^UQ)=(XVEhoMq#5CPpn4aK@T3bK`7N9x$5+LaXj&U^$?CT*%%8IQy;;OlwuVM5 zf*;>9q%`5SWL>>WKc9KI5r%F)3+pFd%>KFK}+2}8MUz{$z!mPJ1WAne*) zGv8AuvZkS$7<{srx7C)X^dmbR=Ch-Tj7xCt8`aGfM+lA`4Y>z@{R>(rfm1QW6Gw_X zC+0i$`Q(1tnu&y~YQ<1D)TWpT=Jajp;CvZL%3+qy;9aKS&+n_s^l&NC+V^yNah?9b za1!DP`ls|mdLcAt&wsYq@FciVE^=4Fm-}685-tiuMo@gJeXgK15zzena3cI`lba-q zak(x{)@zBpm9hu$bodA(W308Obo*S`s_HhuMt&k+<+dz1%ah)7iW+X@= zruNwz`9xLHHT6ON)cXk0EnCT;#puYfck%?-zf0vJ?wPwi5*y+Z(w$&P8VsUWTrGc9sj@)Z! zf}8hA#8VPk)1WmXL#A|Dh%)OwA(`xGGBCLJ-G;8WNJNqyO%7io3*1IL3hPSPS#KjX z7~Wpp_Wi^NJ2$rZ@B;cFuH{CUFZYyw8YrYA5^;W1g5P~|g%aLCxd2U1NtL+;=}J6YTJ z6KC?I5cX>t_a}+h$E~(328(8Ox}qJJ^0kpjqnYp6 zx@qF`y&{p80*QJphUcWq*r#=X>6i-Dw$C$hmHXHh#0E&<3j)3ycFPVOEHkZO890D@*r@=F_kzZS?? zYhyuqoZE7IzpNih92TRT1-xAxo~Ui12fa*)Xge6U{P;wjZSm!aI{&VOn2}@ZfxbE@ z-w7Hoa@fFm#MyB}!3-rr*+zX{((OudAjrJPKX zXfC7N!ZD+0)?ESw;A(9Z^njiM-+jzXgJafDWvvHn4s>YFiIq@ zwE$t<4mTzPzymNXO${1SBG9qF5ibWGPH4f_h|SPVPjORu8iHYHdZWu9sySfT+8rm{6<0Q2x$7)HW%=@egik;{u5 za>Baxdj3`X`Cdfb!{BRtfA27CO?u>Em@tBuaJH079y{YvZtgG92#E+kYkruJUZs~w zI0AScyGyn~lznN?I%cdcZ)kOHbnI0Z_yCaQ&a=vPKuG!|Y5_8SLVGkMV4Tc7*^P9p zD_-ZhLF5r6-VyP6t?@t#CSD@QSwO>!+(!%O6<>U3Y3MkEt^>Ge!!Qh@m-S-sf>Q}| z8~D~I!_XN0cHp_?=os%AK+NP8x4jisC!5FH@e5f)Rb0mpKiv@Hfrgrdv98Pe0VuyX ztWI*uX&T1MI+G#F7~)yP-tYx+HyULFGd02Nw8*7<&*ke@iy{LRqCDHZHaW#cl>r=C z2oiAqISer@-7=>boVegKd< zrdvS3@rTGBI}O3f9ASoi87EnnjE7x*vb-X}mu|#s{GWi{x2+cg_+3RPPy?NJY~W@2 zXAKBeh?PF-`D@5IuQ=nTubEp+-zO(~RHcwL5roOLta9BIMde!G$VPMda=+Rj{{Q{B0g5E87uw zz>&w5Ot^wWOr7N&M|fQ68W-UyKE$N=L(KU8h)G>mM0R{HyfoYfgg{D(fqR$x5R(%5 zkVefzSK|&aE`;|qycaXJcBsk*ts#(c%6y=W;|t?{<<(n1I>JCrU;DwkH7RKh9q}F+ z)fYn&Pc-VYbm@xd6|U48w?PL$6u@`@utIqzGEHGMD+lzoDekbycxVt4@qt{qA`;tL zZ%cMk3R&xOv^7>wFlTq036IlVbB8{(0Fn1Th`ycUYb||{67-4J?Ta_-qqHL_tOW?C zt0xTMizS}`sKTMy<%;LG24J~RExfa5o<|HKmYL+Ay&>jKfV^%=spxCo&ZQjaTqD)n z$cx&??bD%im=}n;XaPKQubrEKwDy8Spi4v}~KRoLfTqw&5hAlqL5+gT_1E#aaV=!pM$K+A# z6D=%~A4XKs;K?~h%(5ME%2?!8DSoFv>jO;-Ztw6*YE%C*3rIxpOO}R&$2UUdq(Me! zUe(CahWC&~%)JFY_b$DbwM-5QOA7Z81kE-z@_MB=a+l%YRX;~>5y|a>#p(qy5FMNs zqn#9C`xUEr3kXDv^Ys8Aw#f+)i74OgIE56Ee6Nd&=;FyAdW9H9YWLd!s~=TK2U75g zBGRC#u{GDpk#{3S)cr!FK~p2C3VTcjfHOi&&aStJvr#N|=^_tej&A{jA8b?`z>!hl z5Yt-v#wc77HO!6TPvdwY5fyF!E<$7G!E=0nbjOr>bagfhb=U2$a`s9SLWvclRYHkiE0pIt`2AH8>yVGeO+8r;=t0R*cyv)Et25g zNU5xa!|IaR3}ke;l2Up3&CQY`j6V{wX9q;Gh?92L00vvw0b}=X#2LVmEyzm6p6Z6g?%jUwZh%RAk6mg~#cj#6lzN4Y~17a6mo5Qsa*y5^iu#-}W zX%Os`aslu3@@EIg=y%do8X7u=;c1`iMQ?col)c6wplpCCpcud-pzMIbiFUwXI2+Xl zaNiK?X3cnAw$kx<+ znkJP6eiH_f%%WsmxX>6dD|YK#{u+QzG{LV6kMo{f--nfly;;Db!cZm0RmUv(MwO`J zWS_T)1skO9uZU!%!)eltY=faByIw?Mks~}GMou~#$n|q+it)kg-qeUMa>BaT0^aQ6 zv#DYMM?OM~t`JTa!XJuGl=aURd>0@5rN@ka#`r$u5fSu6f|=D9Jk+N=(hK^>j>d43 z;PrwAFTxSyWe1HkAkx z)%@BNiJ~Wi@UCKaXAp&RVAFY}Q4g>$z2jfqP*1@2%1PTFpJnlboKR$ zNcJ0IF&M-^Hs-uf=%Vg)Zmfdog9s{`Uuyye11V~<2OtuV{?0%M9OCLIxH1sUX=7Kd zs4a3jcdQ`;R9xN}|hwL4VDli%E$c-f^IHX zs)z-E7->eoXRKlR0Z?UDQnvk&E!fHW_N^b$;Tsd;p&t=12p8=tA^>t*7xdPT$>?^( z9dP7;nrks0?u0*|{oi-OPw~M4OV1<+*}r4=vsOPiV?oe~vsu|5ZtAB^MxMLh9j$d) zW4YbXIdHpGGP;YK#Pb0y?nfOp~^9a zi`+o`bpiY)Gy*T~PmHD|xv)Mj1>8htJ>zk|qd|E)M&wZ@JB7g@NKOfM$&iI!<2~#c zT!*%}tc*pxsv7KN0IXHfK4J!NWG~qHrS!9Zx3HHi3?4S}F8;KUJMLR2ps%IPF}f>OTlt%#sM*ji|8UAz`<; z4#;OuF!m>W&_o<8DvH1qss$wPHI#s9DiQc+scNSY?@Zn0JEBb4N2hlh_4#uh?!k$I zM!YifK=vCvdA?4Bmvw~c;bA6-VcY9k`QTQOOyT=nOmc~N(7jz$0dwkX=HkQ12qSsZ zPYqFCVm<&8w>fqINGyn5(ANRLT_|EFX$OGxP9!e0{fOL7nZ-C6N5*yNu&V+I2!)R8 z01yKxHfQ^Z=7Ra6sj(5tCZsDVjYG^IMvf4tuATsm*wD7ucwEG_s*U)FS8sB~BFae$ zufEYZRm_c)$>=_6EEaI&fWrMt&WB6$4*@H2EySjL7ayEa^Nq}Eo|Tn(G3#0$6aV{V zx;oR;QEs_RClEga{Lvz^Rt)%}j_>oiO^#qUI~2L6A11&uU$wE)x9ac)$2G_T$lrts ziv=VC-&aF{#~!8D^g%)mbzzC5H%k@~%~^GXGwpB#kyo8XBzPb$N}q_!Yn#Y)t=s7# zXqZ>w?92RMIU!r%G(o<(mkXvDAQ@{H(Tz89VZ#-FhD$$1#m%yH>;Pe-pvh!zh=F7) zsT)$b7jk}C6$?X46>nGz=**&B4>09M@M=Sx`o#NkC616cq!KoWrM}D}Z-|{fdArof z9;4RG7sTk4;Rn-)=mX}*b{9-mu{C7?M;?L?r|_z;c8jbdxkbA2rR0U6q2^(zU;9SrRP!*` zU%-gS&d{P``G>J08<&IrM}`FS({_leGes@TxR>I zlf1_33&LKb#upCPX~Zr@EdvZ<#hlphh>HOLmiEpr02yoSPpbuR7eJC?3n-*XH~_fj zDeEK*mXz>VcNPGPBk#c8oj|VPP{VB;5Q~B<(IOI_TQ4gvh$0q+xKuo(IGGY6O?_P` zGW@clvO6pkLO6GSqDmA`L8-Ud2@N~YWH(Fb#1E*S%U-zaWKuk^7K_NcQNY;pG7hpO z@Lab`Gm==77~=*onAqOb4mff*!UWa!?Gm0qo2u5li;wSI!XLg@GR!k>&-Zc5-q6rz zvv7|zZ_^2j!adA})-$%w>4rwy$K*JqKFD~z--(oN>F_w<6Ts)Q#;K>taJ#U;2j7Z3 zK(S>ifmq$$PCF9wEu~R#;akbcTCLfVE6SGdYpYWf!mec}cM)Vfxjvgk6hWYp!0m?E z8=`D5hyk-N>|QR2!RfU_G=0D#m}doBBNxuaU;98HE@3*h4+Qxu-09p9Mc4}}vhIkP zcd)i2uJ=)mJc+gs6+gYhdctR^zGwflsfO=Na90{T9xd)L&K<|ObRB0NeR^1D)@=K& zMgur*duf$Bw{P1X^DVabyZG4NDM!`rx9{A);7GoI@1iK+Z8N_A^+*SVUUXY|eLa&S zi^t6!fDd$zdkX-B9$5$$5Q!$VRyB6A{Om;=moqwCVnI`|dxF7@g6mGk&?YPcVYJg{ zpt+9}HRv5{|4`yENxc)`f4RRKfedV+2fYml2Mf83Zi1@@GN`ur63C#A%fScw>Y#jY z()H|@MhgFY&FqdRmP%Ul8|AWi2otrC^6L!{Nv*rCq@c-OlvE0mG%(ju@L7J9 zcMHt=lBKE;eW2~#nN$^?txLCv#G=EE|AttZhf?jxN++@LHvkQLO)O6X5VZ#je+%FZ ztS>|Q5ElZ9y-PRZG=vaI`3$1^spRYOl4BaDssoL}=7f6`TVpPSxpuepLUO?F)|L4n z%JG5c@RgK~N5KMr){CgcvAh@3kL!ixP?-p#cD-oyD16p-s(Am*ep%idRgpVlGO!s$ z91;k6d_#P=oJHIK^ORaR7@m3lOT@E=ime%vpW@GuaeDFc9f{l2ZgHukT)<KI{71yq(k!5Ll28Ta`w2--UjC^GuaV$z>%56R}PhhnLK%X<~?5H&c9(3^%Q^FC`80z zq@QXC=TWA>qu#s8pb-d4&eR)PMMh_5J!tRmSWXasXaA0ua9+O;#v-dCm40hS77c#w z&LR>~6pxAF~~bLyX;VscyR9lg35r7En0fn((g@>H4WcYMeBs@#MWQPw6@Rxr1kqe%HQCOE;JPrAXqzJU8ev~ z+`UH83`x*;t8wo54bcFOtRN9^x&5sD3M+7Oc%b{diw~Vhb|JY{$M=UtsJmiZt0#Hh z*z1WLPsyNROc`JiA83xsv=AujL<5Oto*+MQT0k{GJjS89b(5#b`8$Q6Y>sXVun8Hh@vK`v*?XRLKVpR zs1}i{k`VUJE!vW(kA4}o2@noOW(NzXF%hrX0q_G*xl<`TpFn$X+MY*Qh=cBX3B%oqWlN*B{lv3Sti!@>)Q}pbXA&oh6&+fcJsQ7X?jcZF1T!U-(D$lOsl-2`<2ci#}Gy!g^&KpwwcU>7k$?5 zj))?HN8x2)gn5Lh?0F}JggL(T13)yqc)}oiiCZzotqU2OL%Gt>crN86+7SscjN{TL z;#9zz2$3#kJ4rj>H(|Ii>U<(EEryfIy;@YM_^Icw2SD_db-4|G7I^T>1|yKNjIfRd z!^;XyzF77`WyVZfpSlm>>~2^cdMj3?9}oX1@$IQBn#Co(*E=#X9sWVwUYn z{v(6@Yokw6a40U}(wU(UEE{Qx~m)WB{V5rtD-ZdBRm1Pb-;%}?vs zWr@<>I}B$eGGi~25~W|FNQoDsEPN%>f__PtKpis0-;> z;(gRLkyF)ZLi8|G^NsKoVwkiPKzKIavd^Rd+26@ce~x09@EYo)0S>UYslm?_mw&1{ovH_bIDc&q5hLrwm~ zD)aTIC{B^RP`NwynBfm)6~!2 z47rCFu@G_Zu8NF$@+H)Fvbl6Jst^wso$Ooxy|l8<^1C5H+HHeyQjjaD<1IJXMU)E+ zvW+1^TX6)t4a740!7U@1iWHa8LAU5h@s<3w#ei=sI>b&G@eGLWHaTPl^1($-NGk7a z!nMcDxw}s55CQ$|?b2iEPj1`yp#DN>JNc6jV#$`<^egp6T1YGjv3IVh;S-vka~_Tp z!^SjqXXDC|0f2DK&$x2jyR~-<%~in8O!~7rYLpQ5c;kv%C6u2~gt0`*rQ9Z2c*C); zXhJfI#AWvP9(kaQUN6lSR7fJ^9Uy0+31oNCfKq*^dYh9zGvl_m@(^6mGH55<0Y@3s zDt7pfovsSyFUadZ(5H@pFHM@i_lqR1PVD@h>_YN&#zjyjw!bJcQCCFqE;q93vujUQ z8yA%Av-Y)?W!odGE+h+payI=$yf@a#A%_XC2?LIg z-dzca;qp~aG39@nH_45V5|Q;ZD+P!0tUqRR8~0v+7HR-^LLx3*aR*JlMBSo%0cr6B zD7;)9x{|(-J-098xFn;logx$)3mCe_$jauHa-RpP&ZiR~8!G}{e|9n{o9M5fomvL9xrZo1u0qm%W=V$H-eq=H z3iIj?0Fv{)efFtOx!6wZGmovjPcQSMha>Bloiq`-!xZ2^bm#BBmZatCs&YQn%BzP&Ca*K$O< zpa~N`!ogg7%n>&|y=>4Ki>OT+P!cIzqV7tBtmPi$#$y%Dg8CJEizu7Ptny>XSv;|$ z9@ThHxQuER{@w|9z){W;BqPG^vwN^7?|Mh?zvJrLmR$e6OG5#Y-)uJT#J~%SSR+7o zYpw|;R3Y^{^=X9xrvW03vs+slPl^Gj;>H0*VLy0hNGZv0195$4s2l=Tjn29jaJ|kR z>F5?Km^w3}uPDM`NV(>2^8(X`6k2LHJK#qlx=&7k;ifV_S9^gZRbf=cpUf2>x>5iil5q7}6Y#%l_)X!x{rzQ`Sn>ZqsE5n`U}zQp1b zFG2dtQ4)A+k`l{|aB_ zJ~Uo}n5%2!^MYTLA(yXnRb`oz;6@SFf&pS~b2UZGMyLQs+0!zUr1cJ77xoYtLli4I z%!58;jOg8za-Yq}D5eoE;v(~Ewv*9=5p(fE#*>wtev@nn1CmYiQFnltF@}h{y(7wK z6jSz|7U!u<{I}JF-b6|;y%CnWT*8x8Cs$2K)siOk&SMjy>xi+^W8fvetbplUWvBvu z9fL>!vOHbodQB)sEU9p>JqGGp9hNz{#3(6u z`*LFcX3}uz!BaoPodCI#ZBH)}eAX?|H8&oG#Z-e{B*a1D|9>O&l%=nS9YXl2Nj}}C zlZVHO$GRfq3Z|JommVE~MOIJeNW(J--@R6cyi;@=;3xt+)B0pXB1g&=~%*T9Jbwi2i+aPx1Q85m;NmA%{;5_6Kcr>z+0vQP-V#5 zLswobh<5T`c$7IU#LB4_smtW^0T6N9^jUWm<$v2syNoL3%Jl0N3#AH+UMHgL5bl0m zC}Nw7y}i!U5R-IXFiAJwoVom zH3=1=^{*1b7apfL+6i~SQBH}%EvMAsW#avHlvwm)2l{v&EhO?xvl)@6#*HtMO-)Vo zcCzuzDG~)cBjl3ACc8g#T5`z@yf;@?BW*av5uwN>p&Fc7!Ye+J6n7BL%Pb?3*=qm^ zhKsAZE7As2qv*Y>MG}x^+DA=DG%emYHy$Y~6ii##)VBqX*w zB0u-RSIU8Y$7i=oO;erd{ddYznA7i>fHyaP?-rpU3tU*b zm09uQij5d+`CE|{BE1pM3t5VTOpfo|W@sa;B(SenGSW_nzMK^uMIYtQdL^WOEH5ta z!_Fg*I;*PFwJcoi<|PZ6H?>J$NK~pVxX;8 zOVd?~B)9`B73Vm$>BnMM@I@WX#kizrZU8=6?8IuDb~&jQJ1VWK2QlGAPP%v(6CR&_ z;){27J$XP(H~~bDs>bN9dOZ5At8u0&!VPdd>{=!=`{`3kKkOd#aj20upK0j74{M#r zB)>E->(%-`uRC}c{{+Ngt@q!*?Y~>*J>5PfnH*khzVPOk%&2+D%8k(+L}u}w-WZ+t zu`b~ErpQP^8FEtN>!+`rI(Y$))7SGHYHq!WjZZ#vr{;y%W|F6=24&Ba=R4mfTESK+Op}WkQfbd= zJ0Xt~dKE6gE3%WE(OCJmd>!hln*SX)2m3P)p7;~cn3Nbsa5ncG4UYFcv5BM}dPs0neTA^*&6 zI3#s%mq9zB=zozaVz(YEsixK*hu{w-KJ%~vy5QfsQh?)$P_XLVr$HuB^mmlD2Ys9f z*wRZg(=3uPu*SCEMY|E5D$U=AeIs81Y0%m6S*-K$dpS>l_RVhD0^%TyD$#Z;ws|`x0U(O zI6=j)x&E+mlKDB3SLnvY5J{rh7Bv^Ln8jKMdM=zSi|G7yE~wBcYM9f6gwIik;_|)F z-g!gOgal(`e_bIIwIX~)6UzBZl=HR6Opao!_XWv`W&7&o?O#huQ^*>wmE^-a57W8l z%={EHJK+vE${$M6i)sF0vEaqI=Rxnk59gl#`{T!vMP<&Y*KhdvGKyJ-^xc1#kyk#z z_IpL~LoRK<^Kc_msXMm#&fsaRC`dTj9JqD>1|Qif6hP(yw_**zai_heDL|#9q@RGF zuW6c}qE9!~wq?=P7vBdA{i$Tc-0dt^>m!?=MiTOWYz-7MVX}qg=(Fz){P7heTH~L; zkQ1c*@*h6@@_+uz|NB4x2TJLWfB3)t?Z5r&zx~61{m=jMPyg+o|Kb1m@BiPw{PTbR z&;Rw`{^9@m&;Rm2|Ly<&*MIsS|L~vW-~Z`f{?mW?FaP|@UsC$z-+lb$Z$kO2{b6Q@ zK9Jc}4w?STuiWoHSmwzp_>lb;vQJL-SO5Ln_Iv-A2mJd5+y?f2GP5;je*KMis!aX4 z_WMuu`){(dj>7Q;9E zIpn{;kns%g?T)h{_bJVIXrr9J|19SOm1V}yPb0O~lzHoXetsAsn24%fzfy`J9j{#; zJeka9K)&$WG4z7G9%Spzhp<7mYqK+gjYI8dh3_eKQ554;NfE6(>)%EH%0b~1zFykf z;(PlJ>ghF)Z`&tZ>3P z3?VapZBQPdH~gul-=&XJ6>xK-8-_* z-@B*UV_3WIWE2PDKJffGNGwlQ5=ACz3;XKUTO>?0XeP;pN%52m8KoD_muhOGE~#fc0f8PcaKj`GK@ zhQo>*B4=l4`+cani*TWeTs>wo9$Kzoclq&P5BVXAycdz$8zD1rPXb=awAYM|PSoXx z{7>I-fM5N=hqpUzw>Ci45~NBoJDJHT(n%VW^(cE{QD2F&K$S<2M2c~f&4AX>S*Qn z{=0~K%Cm0FzQP8VL3Dg7Y~Xjf2)S zK7G6}Wd?ly8ZOnAw_cHnDa#YnW!9;h)@t(Z{1sVig!Syd*A_~pxfAQ7a=&@zjwfmCT~JEB!*VMqaMqVwRc164!l&BZp3t*a)`nT6FkXe!byX1 z8C1*TndYE|;P9s9G?t}F{;BEnIExIClr>iwADwVr%p$L3e~VaDgybR)-rN?m z7*wKSdA;AfD5D+=OaYoVP0(H8xFd_f-guP7!-cT^R^Cx4J6^xl5K4-02OPzfoPr2T zo$CBuTs`(uN1p~$G+}K)?;N{JY__0S;q@PRnF_ZAnm#JUz0tXd@P zB;i2{AksaO`bl*a#o8h$YVFE}c{39$ez;W=k~b3v!P}@)%ixzZDdL{(_liS4k~10`~ z8zIRRd=vHCa2CM(a2!@CsdT)Xjil5}aXlh$M8h@f_WSD%Fhnew_)@uT7E03)2c!ud zGEWf}En;`V9dHz}u8`*UaNqqdVh#7b6TSaVvDLt5*nV$T;`)0R8PZ9s!!`Q#l&hK+ zmu!_7ku__BP4x1T!^yh?e6}qiwCXC5ATo!ot9kbvh+Zy@m4Lg%)f|(<1y_NAxY%Ps z4tB9E51oshhpYK*oHEE3x2p(6o>ImBN|?C~O-U2xVpINb4aJi8U={5bF3bmOt2@Z!&_IuEO$`Uo^pCRov7zL23~m-4I z$do`ZJgeOZOD0HjHay0u07C*xgFG}|mW{_DsOcFSAtf$LT-MX}NeCm4Z|iN|tOrMm zGzA#)YVT?X90gfRHE>Fk{4U5MCP^6R{j+9KeW*Fi@1k?Nn=r^NMX4QlQx+61wOL8M z-4Zt6P+QWY_31{PhB#LQKAB%vOEA2|zy=b^Ep}n*$kv zHyyEii#Jdu*zM3O*Jt@KGlhcg?tNxzoHIjF@xH3w2rK0t>-8gwf{YQT7Y$K}EhqnF z<`jIkDB=hQgni~F;FFod@xTzFwanbONAS%4IvkR=EpHl`sx^`6q zi4uw?6cf@lqi`eqHsF+CYV#Z^cX5BcXc0MiAg<3B_zY&fW;wAUp5_;>NqXu_(p-D_ zrVh`x8{rV|JK+vE%Kd0W9XjQ~o*60DJAwy&+!4s|GbZY2hNK&F2b<;)>nGh1v$eFV z^SwbXdt?@(YVofZ`0^COM=BHWO#y5TVntkK8w-oP2_12fZKA+Ais&eIJ@FjDUMGNj z%E3fU7-Z=bw~uqtnyI52{DH8&RfKDgu?B;)y&_BmJFsmogd%ZbyZP7&bHWL~TnSsb zi3PrhVfI2*EpkB=NhNq`5e?o4qo(15gO-dPV67&$oUc zE@n7cTHAiw$B8SiuSK)lRDWBq>{6CI8NIDYnw_Z|dm|*>e_7yDVK2>-;kK+~<>Xw~ zPci^xg)Msfa4$h!8UR_o%Js>`;b`Sj8i?x!26-{^;d{!N7|7kbvis)rQd*^NMd&SA z?A&!RdyM|Dz!x!WTf&VSA!!Dw@1h91E8UH7NZXxo2OPz4B7x2^75JMN7NJShR!1M| zH`U{fEZX$CJiC3P!aV?6{J~pHaZ8O6t{i&PL+|ym)r6TWitga%O)CEUA~7}LE6+mQ z2;XiVOw)U}!C2qit<0yR!M&~qtPwpNG4#mksJWuRRprFG42RVFZh_C6h2}ApuEU{1 zdQQtc47p1FzIA5jm4bZZdZ4JJ5aHC^0M4|;_WC1H#FilZ+o&A36jXargbx3(di_Yw zC+c@*%}V6i40UC1G2-iA{goB55lD2RTNEKGOVFtGK*){t%V)kypj=a{kEPcG8PZw& zn%!!A#lhk-0Yo#WST&(+H(T1X?0K9-y{E9imzQswYuLZl?2+g{5@GK1KxQD1XUu2w z;CVAET=uDX;2@{SlbU@mI-&?(yk*iBi>Jw>Pv~jXR$xMBRv-|2se?q7Y-Km=m>Lj+Jm76_u28IgBc#5QxFN z5#kd-45nhG*cgbxR4t0R$BpK-MVTe4&gd2^^}-JT(q!WW)g`6sKcz704v_1fUt z&B#^eSZbuYI1m?^V+&rNdL7|fjo;yIoN}TT5le^RP+Pj))9x)&=*hV2w;(pw6nmJ5 zUc<+jvexg1uyMTg1HYL5-J)%O6l>?xVc~y7i z$*by|d+*9m*(|b2#M+@Nkx8Z5t2NoVD_up%lZ?H5o zPdPb|xM4i#Ct2Lm3B3(`sauY+i?~O;1 zA+CM&ot2j7C?1&47)M+H?0|y88T2%Xo@H8g!yEI|av=+vBCMpS4fV!6cYl$zh&wB! z&^MI0(EwP_Sp!7Mie{>rcSQ6xd-TkF3yO6$YC^G9a*pQ8e5|H<;3`%UX>rT}kg0qb zoCXx;7u9&@fr~53q}`aWWovQ1PiJaX-k7IyR;mkp5*Dfu+FOrY+i@MH#^9;$>i5mp z03u54;gijszw@vIjv_&R=Yd@N=GbokuGel4cK1T(8HO;^@q5=?TrN%L+R^Wog>nqHo=zBgD6YjKtVwNqMC$>mle-CVJX*S|Q)P>6wHkfKj0uz%53cSy zwEte)g|?Z$|9pPO)<{chxQ@8m$;gx-Mq*IyiN%^pm3;uT2do2tiNnQfX91^{iVpYA zLm7nLChAydLIT}L<9cJB-*}``ti+gGZa@XVxk~nl2IQQ`prQeyZY5GADCY3rkR$ZU z94Q2m(mNpm9iF@?m+4^Jie>HOQ4)&WeoZ}Z?y4QFGnA^Jkmgn{@dn1E`lB;plUBo!3B*WTvRb1Chz z(}cv1RI?L~;fRc6uJtPC&avcl0-y{Yc9Lj`+zIL}J0y-`ZU;=cx=0g|*BB8W8kFuw z!6}`Yd(iu55ur+C3%`db`v2{#kV#V1^no`&3NIU#?AfE~x!+q&-N~XS+9J*UPQiuX zG@+QVCi=V)#${XASokPAp(r(^u(B2AkY|eA?ZfjP$-S^TTI^TjH&({R-3fQVaYvISAcSj3`Lm;Cg5{+31AXlD zMAq7xqevmwgSUAuii6;1&}0uiFwUa8;vc4z1GSDY39? z@!l9xLdLRw#Hkr1T*LfzoEctT!+9SkH&)~<_;j4)@}OM8VVrWlSZZ{5~Z=Aq2JB(ASJoh5H1iPCR`L>0X; z_-!a2+*3FJnZxGxO>wbqo@gC$(Q2sR>Xylx%88uBH1^KbR8B-_Mq&{ZS6uH9w)iPm zV$TPvaUHl4`=|x7DXwyfu_^MGxvDsBqKoZ>qDh(T^No<4qJ^0EeK?{v#`1if4~N4W zUVc8)HXO(5Yu47rmHaHs26yEO;boVlF$k$~Ru*=5LQmP9wa2M01&Y<{774|P%gpU9 z%8?wR*UKJGLQ@UT3&mww_{Hw~tC>)ERt62Ii6r29O%nDX%5;3|N}8AgN`@^E)ncpJ zNLgnEAT4?^dOC=DNK_8LOozu^E?2$KPjTo6$=)I=l`7iB4BS8>f09`UU%EW29Y< zh;mz(7Q0ySEh<~SREpAYJ-e&CBpTN%bs(<7PEHqJ5A2HaAOh#kZhR&2*6b#+oQcR) zfJ|-Wb*6^|BNWr5t*beuCFb@dstBnq(rmX2kFqy#HP?h48*8|?n|u|x>w z)8mx&Qh7pE>{78aM|&M7I`aB0r>aMAzWV~ZtOlIR5A2f2Gp!3@sB&T-;jsQ`m$?}W`lbKyE2EE+0zs4)~y+jWb~ z%eUmp(&E%G1MYPDLYjAIhP0&BDMHe6WE_+)gY$T+6HU0YLa90#-vK~q@*DHRXnqnS z-FO^QawpsYNAX#T{I`Bs@Cor*h;{J;eJG3svAMOGL>cQvOxEf-7&~3xPt=M1o>)nlAk2{cwdIjN^@0*2vo*;76RjSq~LfNEG z6@5UbdKgRTLGK?y#BapU)KRQcOo!Vy;0n@nzxm+JPtejbV+rjoi_$?2?9W;*M)VXou`#i28xz~M?M!Ujb~fLC_hDb}+dbXCuIkg( zbrkBKG=#1K3r#jwVK72qXVE1p8|(dl65C8Uh=gOM23f8vgEVG7IvhWoXPl zPYQE$BPaNCn51R(IqtYulg^vV3y9dMhW(l)(=~?r%T)ZD{)f|e9tW@6`4^5!0>y;#Mfi-dU6xF?#W4J25&xKVFcU9rmI zW{u#MKx;6NpYLh&hz-mk)1w(!<)S=Drg9Hfmt_;476!^Q|CK*(!tH{{ksmm`&|;oO zjI#@mMYGvVKH`c&NCK3 zHtOpBsMd|nd(H-G{JjPwz_LptL`p4O$X5tlzuqOZJBd3oMHVlGLFSj(fjf&ZBx1zHNEO$r zDX!?S;L9nJZo#R(Y}(^928&Ic0bKBpEVTmQ8{s1win&)4>{M<*ZuW``k6d#X|=rSIBe zF+#U8YMuX(9-jOJb`07T2lI#t*uL2FSugO$Z?Fq2de~UrUkMjt%d!?)U-#{?yN5^P+8)YB%gLPKNOp*BbK5}OH}2;`m<)f3H&I3X{BiuuQpTku;nG@ zeY!mz!N%(Ibwp9{W7sr*To@_~guPOS9W`v+i$fV6pi~)o*AkGFW2D`(N?B#(4OsZu!E#DZ1aaOx@dP4C-;)U z1uvh%n}`n}FoB3HSRqsjw&f5b*!n0QqRulW*;loQ6R00`+kx zxa_y048=r**!6*n5sSWT3r5{m0}JvmgD0a7vRfXc5Zx*lgJEdIdjX6!*{w4Lty1(p zYNZqoO(~bzMp8-fd`tH&B^WZM07-pF22r%e_K#SNB$e zk_S_6u5~Xu0!o(UOMIV*K@96j(BF*k6s&%!=S8e_Og3pEB&^DK1>EKKsof4VJMmCPc-iK}U3 zNdnH-Zq|7ce?xxi<5@(Z5!k$5uv(x-M4duD*%Rpr$F%Xb&c&q9T}v zpcsvMeW)E5&-Kkv=0ZD;qObjxLrr*z166o3sic)|BX5_wKo#%x5v^osw7!a5n!U2% z>?olYoDmLqM;PK{MtsEtM4OHX%6I!sdO!9mYPMdcbts&=D;o+!Yx#3aeQ+BM&D6jd zfH*Z|!4K`=iI(5K3uF?}gB%`rqG>PIl&+IVr|try6b;nNKb1^WNVs?yw}+nhR6kf> zU3k|#FyzxyjAa$Zb=EEKAewq0Rb*`FpbRfh=AU@MQ44k8_|2Rqwc8@L}>Haw(7 z;*Kn#C~a4o4tF}r*8DcE&8{F!bx{EWHavOTsT=rj2%XZLi-o4NRZv_7iL<^&=t_z{2|%@QwKq*PW-$(;s*t(a-&>NY(D+2tUevm z^^`g63V0GMSGYfM;DV^kPOBJVl8C|xg|);iDnwo7KA74ZL>D;9^#zEqkgzhO?~QA1 zF+ZwFH0l*z?RbCBI6H%{M!nKN2bo`)ArymkM>Hw~m36&YJXcUj+hRl@C&17tkzEOp z`DN@{7Ncq$unGtlM2{zdsRj+??_w|i9kbM!tbAO*f|4t2((Inv>2k3fpxbx8uH2#9 zkMI=2#>*u_zWopPP_QW=PDu+$Nbzs?ln;*|JuGALH1}KutCYrZ``N|GE{PSA=N^1DpRsaRc6Okz1;n=gT%M%c;zmTWk-b%z^4~)H8w(@ zA&XuqA_-aEdWaN&;zBg(HVgY8 zyj?L^(bp)56P)v+)g9faVs$HdCmEv$Ox4nhq+thkKW&M4ks7Ifb?%XfwkR#%9PG<- zGk(j(9=1%%Z@`7qQ}*Qqegg5cnHils$hCD5IlM)voWfD=@sKi4|1jPRo->8_PdF}Q zhWqJ>d&YooT}#DIW5f?J44j;K(=7F=+?>F3mj3a01Gi9O(As%YS@HSmqTj0hS$duN zFV8KGErwX%+TfFVc&LYLC-uo?mYoC8n);m3vbAaNwweG!EA_AOJahBwl^yw+pb*h7 zf!ro`P}TF&nyMvD7Zm&TU+&o@1=h&a(LFc;%m)LOWL`g?$QUP z>IEA)DU7;R`%+vuMCDx;ef?_b>Rn6B_UtL`DEXxa(T0iQwO!OKt_Rrn?{gW=g2+p4 zJpNC>H94a_F2#)>vx1(vCGgmm#x~^gZNFajlT>b+GT+RPgW)(V)%Kr;>NeRwdUBGO z``>lz7;McxbgRcNAHBmS5Nn3R)dSGT*DZUdIGTU=q8a*n8MHrcs~Hp;zv$O@D>oe~ z8w9qFb3TQuN&HOb0YBc8qs;ff)xxy{e(U_GAkJ&GfF<751>wOF>5-UY$oJ}^WYl>? zLo}n}`8_qI-&L52lt2bQaQWj9W)3l;1#%K$#m|uAhsHS$mi;0sN0;R31*$R$^`aT< z5&YmJ5H;_Q(x#{qIQ?y%u7Ip8k>T7tDOHb{gwVE`LnweY8cof)7@PQKDG!oBejZCh+>cP(#;j#8j`%vU=D`T7mw z2*LH_nYcq(Bb}N|YrR^aah*1z>{t)BH^si&4~M);Vlv7=!B0YvVo{r1$H&daDV+S+ zy_5MUW>Q4z@9O5d-ZXeM<9k~0@k(kbgxh6xXm}h4*{?a&Cd!8?@ z?S|*n>125JC6k*Qbm~6p{!66%>aJgQ04C?8r-eUYjn!vGsXyvobl$~TO^wiA%dtM* zFiaeWc}IS0;vttkHA!{hd~8kK|58uLY3hMe$mbMEMnD(9$3GVa7l<#-ngS>)I#kX?t6MW_nmu& z^Xg^z2t7|Ql0;2_qlX4rQfIYM%e?k`?esT z;{CS$r^ijE9c))@_f2m5zmp;F;?r13UcOdN5qY^=?6n<0m@E`}{B|BX-=c_DMNTr- zv)YHS-DKbrc_dEMQA7=7Y-A$mf>4;}P|l|cO}U1u3O$(q>{*bKAXb6K!aNm#Up-b^Zax0DXhJf{`Ip|vQ_tsQ}#9`=u{WI7%BFk&Q*M4)_xU19ma?*i zn7{hjot&u0@hDJasAvU3DMvDWufP@a3bpqH#g}n=OS69c%MHvyobc}!LRlbUZV~Y{k}h|PeFcyUdM-k5ci1+xs?GFLjLXINefP`r*8U?s;J~O zjQPKuO`+-uDPDFjDiN^ty_*9<4aCUS^7(xY2lcN}J&r=wwA45J&E%cSX-4o)Xw)9g zYc`ZrRFd_{kvANhJrh9>b97cwg7ywqu_dJ0^C9zFqq-KkKje?|>hZY*{Dqa|r7b3I zuX69i(bIh^zsEsMszVZ0V3(>~DtXc=$6a0TMiW6ARTyv52T_ip;GH{>>>_fI zR2rxwnaYwvOt4J5e$*?73_C_MtGs@i=Oe@Vt{tE%<%=AnI5&a%n`7z z+I2|3xqhbp_v8|YyuN1V7>C@zFqN&S88Q2RJ1X|^BId#Ag9KtX2VR7Hd#eHe*@z5 zXqUsn`7;Ma2a<~?lH`5ywUp!E`3tRpEjRVx2QPozgP{j872mZ_4TR>d-%2-bmjppI zmjv+VzHSrnWi-btJg|SGhn_C#GvU_0>x7BKFxa8@BG2+5`@ZM%$|C zb1B&T@WH1n{0Kgo8Z7*(LM~Jmhu5yQJ4PuAqc3C>rO9C6apaPy<>;2)g@U?T73K}g za~vs+gWj&l7lzVpwRN~LI6DZ(25MIsq>iRhLo;kJ=A9BW8-InV{@2#?!98Pup6 zJnJ-l#H*b*)J$4mB?j`-dE2-qp+|9F=x2teGE+Xj3U5@Z^q_DFYek~Qe$(HR0Tdl-W8uHpn0^&prDFNgY5NsA6c(S?a0ww7+ zw<&VrHNa*~y?~cKJInGEM_l^1dbhlDM-7pj>uPYu9A@3nou}`{5q`EpRP@k>X=c6X z+l^sY9!7Zr-92{Bab(${%P8s1x@w7jQh&jUd`U%n-*E*k+MRDkYc35J~ zNz%rFQI$CQilnt!RXAmE7bj0xP*46tFEHY0C+7Fd?WOYA(hHnaNWZ ziFI0_kl%j`QOd7d`(`YO&pL^pOJw6^_Y@5l@fV4Hd;kfjml@Jgu;UOR*aFOI91$zq zF{u<5F0N3G3@kdNUZG|;P~Xh7hR*SKWI|qlp2(qUSY7=CWYYrv)??jiBIj{{v~4Tu@Il&kBlV-e)|!73>$^m$;i7wD*N)_8Qk;yyVY@ zp7Oj!^p+<_l>takcbpAy5BsaD4YwN%wtiu;le?bd#1Deni|W$kN~#(UGD!apH&_RX zfO-9N;UI(m@>dtle!Q%>5jc`Lo)KWRItCqm?|wLv9Vtl>-%TyVA_uz6n<1=KFoDqH z0jwq}*y^m^ZWkez-AWcBn3-(32bf*?q3kEG>@u)mYs0+ad9$!i_p~ zn8vWoo7Hbp8dq0b*k7&h%~P@pLZnLG69Zkd|9U=!NIGSwK;!w0)bU?ZG5vPxk#Bb# zeUXrTgzO=hNN6>OBDMIH<^(|PWqSmakbINcuA~-H z_-~}qb`P=wU5JJ3r>JQalVst-6{r0k1!TRH}&rZ*cP8p8wigX`h7l&K< zz%M{-gaJh5XjgM@_Zmrq9$g~!;x^V0_s5bGDjF$<8ILs$9pc*dj81ko;0w(QQ$o08 zsR8Tb78^I{{?U_fb9r=(HD6@>n}bN8aoL>H#g0cjLQoCX`Wpn(&4zg`@=VGczd>*^ zl*~ay#FrnUo=o4dm&b8;6H3jvNam6mhT%g-+Vd`adfpl;KUqXPH+te&NPNPKFnf@FB$L0V-Tb4wuhgs zU^5GIGIp5)bhX>kS7ImJr{kU_wW>i?s9T#8OU#SWUi0dv@$Iu~RS2@LcSzNUd^76Z z|3jsU+Rfk=%ZmS@5@D33)vCWcY_&|`{crPw7d$yw%arK{=?cowx(fdxISuKpiW{om z0-qP~dW)46GbLO#iYM4=Y*Ixd@lQ8q1%e>8?>rIRE;=>hjrE9<{m*=+H z|5%g-=YK3p>$}4_JdLO`J3SCeI{OdERHkdq?K&7#8t78*%If>?|8|@yohvr2YmS(z=6Whjnh> zpUC7UKbhk56QYPDTHNw2S8tZmos_B2sG6_(j`e+EwAWlp z!9uRn{ichF@m$U{Vvy@nN>}uM6r_S-Be>;E>VLrw9o&s}m`MTGnsT$fWm1?1FFG{n zYWUo4F6Xe?gXhD@RR1Q*fp?f6Tqd>QjAdDpEZeAPU5%oVzcMJc-@tc@2UOR2n?ToD z7nNN(i8YFa0srI6)&E8M^w)opR{OpLJe|u}Uxn%hqnG*!e2HtjC-bzi|7`i`s+3Hw zF0Eie$^AzlCX!T`g#N6=hf}b_ize}3+Rao!Z5Ip)nirM1;7jYWepO%)m(U)ILEJ-3 z|8|i4yn8qkMq?n17Na3)2|8X<)f(7Vp~y;ZEB&>=aaeZLjjXzOyg3VvK$D0^$SMdb zy$xF0V^w^K76B!y2NmVC4FqB zHio+2;qr0w@y>NF=b%LJo!{j7fA=@_V|F;ZRZLbWZC|nr;BRfqo5;;D4nxMy<7#e) z)nddkTLkaBoyCtY(R#uw#b!^@dQ^kiPH!xMs@AGIl#!x7h)7K(xrVj{^zjTel*jN`;!!kXd1Qu=1L3tF=As*RukT3{n&UETCih{#sl-ysZ~~) z+?&H{&~WlkSMD6p$4c`}V}ySVrkK!75dXdMRYIC^gyU9$HbaJKj8Iw#N*IHmg>RB}4{9@m}xzj56=wQ@F7N=Mebqnt3z{r86VV)CKOvVzBm zBE=`2F~h1Iwy!PM->-4@Df;W7)2E>FPLaDa+q|>b^GA^PdEYHpwWpYb%^Xa9F`4&l zxdg0z@t=7C+R)k9*HQut{urNqpJrqI;pQ#^To>d^1Z>Unt!^KD9c+GIpfsd%S_K8j}gL zyLO~{#Um?nURpcnL6vupC0kSA}>qbe2(niUoM{?P%rj) zJ;i` z{~4N86e$-|4e(<`l$*gqJoECqT8Blo_`5rj!e7&rRH`;$u}U(HQ!{ozbU+J@M^J>^ z@fIYHs8rygAkf8M%`ZNoubJQaOZmhtVTg?MgDjZaRv6FL%-W*mR(M7pM$G{G=IBW) z$Y~37u0Iz^6m-I9-MNSu69x8{lMv{2cG4#kwo#e8;6hSD?^sFsJ>IDe+ty&gC~5)Nr-{j1ac3f=?z7sz3$#|X!VFy4 zQ!PmO!5JsL(&awvls7lByIA3l5*PeNjK?!(ZaeU5mp&6wfH+$Wjl=!9y)pAd*4a9!q_j=&>o;%r2_1yV{B@+f~Jb@r%5V)bX46=-*REo@P>aS}Sg9sr| z@nf*ra-+htj}cu)Rs3O7sUOFYDq3Qn!Ei$fo5k(s+p8AoYST-n_(&MdGF1$ICcj|) ze&)2K;h-vzxRiG2j*A4K`O@kMAWf5d^aY)lL;{14tv=1*UQN4uHcWLek#&+gLw^Ry zGew9Yz!SD@;J2XXr>{v&3ME%cjdtJw!P{3McwtH#OdsXqA#<&K6CLI%lob~HKzdA) zh?>l=B*}A|CIDP9@UoblDRdom@xm>PSV3`N}g6c!v|H1gctyQh>Ig9-W6eIK+8O#RYPIgq6R4vo=o~*b-yImmDxqsyW zjERMQTm^a)NuvSj2E(B#Y_gu{X6rOZh!W002=8d_azPBce}Yt#?gYi zXBTBE`O;5qwozl-B|4_1z}motWTKFeyygD3=4^GM;rt|NxQ20$H=IZzC!+RHlNu^#L&FL&QJVlE>Xz4F ziEKIkh%RbOK|rMstTeM*`A1tqJbt>iNqP=tnSr)A1u51^brDJ+_xvR@q=uED%=6p8 zU)JzkCo^U(hlXXL{5dQb$gBJWuWXVz&Ek|Ems`K{XJKzw%Mv-1q&lzamUe47SBMKfFp6tsT|~KFYOlT<{2Ajq)h300B0UjDLuVu?t3g#8k^~Q0dzmJO{!k`~2}} zc7Nj>u|VLAmU>$1Rf z@jNW6c#`59w*jU5hngIim5q5Ts$LLF!rF(6>O%W`hFf^9j+jl#9P`X2tOe4X{Lr!t zP*r%PBI1}8Z%8LseYK7IUVV7>{NB9ff2`tbf#O!NWnP**`u8%zMPprIt|jQsZ#7|_ z(w{3%0(-?pW82Ll7F*3weqgTvr%C~^?0gSEmkxkWSPBhGnSsZrGfTu#y|*AOKKH5& z3<0TjAT{c+bGN1=x=ATC7;_CGow%2zEQw{7xvrOIb2MGF$(Z!2rRXLv#C>*#CCcAZ zPd~As+O}pkZ@Gv#3w_yg!`|8CMtppHt{0|;KtbD}b{#!XLX#US_+M6HgvH9^8}u5T znz(@uJ0FyAL5rSfEgLly;&gya-ocU7+dHk=Lo=agL8{y{aA_fH}v zWw>@UTcYbJ{I)$^dO9vCX9}q7@w3IVg1R;(iL<6naSqB&wn@UIX)%Yvvv=1JM%z66 z@RVBP_&PJ_C9zRF)$1BJU@U#{mO3g`C7}|Mritq~h%&GDZ^A(b6AzhpMwE^p;gH-o zyC6KPy@Xg0r6r}%tvK;fC!16ob-5<6UJ)#A6oqXdCm=@h1q|SIm!f(6=-{k!TN$k? zr4;&BwrvNS4+R8XwrzZ;%-|fI>7f?|D2nx?aB3*Tzvef?Sx6m*iGC;j2LGU1a!t#o zu&1l)Y@40f=nR7r@1+|*5c*T7!F#f)^KjeT*rq_yN4Bf-m*{rK`UeH}yJ!9#ROrb3 ze$?wJot~8iu#^oQPa$BGfESSBy144(|Mw-;(|4O@**}pvu$}J=0g+y9@R;K#Lr!Z2 znCirmD%wL&sQA@t85S;~z`i?s1%B;V9$DhfVit6bVu^h*4%y_-!69bwQj8CSr4a!P ztH_7V4Mfp!r;yCSM6t1@gw93fNunY>rJmj;3T;yMCgH7m`T=G9(ddk zwFJBCz}daq_Hssrt7cBZ}zYGMi=qJ1y^?e3)Uv&3@!mW!fTM8uR|yg6?PO^#Qk1c4 z$aILPgbiY{n>!0OxjB*OU9dqEcVfz%H_@Fkbn<+Ua!^C+9!h_O{ z;CE4MAsz9-;Hige4t1r4YQ3+u?>o%FaYT$-Oo!~v`$_4aQaE{C+uDna!+lPZOFXx* zC28c&>YfZINf(88!4Mf;x_NlgbP9M3Rz|&WmN+LTo{6`xmbM+(LVMVFFx2b>jzN6bo{vTY!Xdsg5~h z1b67Jrc3sBXIvvlS3Jwes;yjAkheW1fdPC+C7!!AcS?#vm(5RTs$j98>-3Zv)pGWg ze;`(ET)*LC2}*}mJM64xh;i6BM#L?f#dAQX;)p|DF^@wo5&0zydy?|4hq)#w3|O=D zocsswG-Y?lJn23ct34v=>01M&QJBWdgV@(!zK&Mj>mLOv1TM96Xd8@TKG{13N@4Te zUW@<c(h(+v%d<;t@5A%&@m!eXISZUI9_) z_SsZ4=Z)tM|GWm%;m7;z^BnK9RwUO)YIkJ|V47b}*%bLC;J9DfVtzKex+8<(5+-I5 zWeWZgJEY$`Pg3&V0@GCrpARb=DE!PyLs?jTn5PCLg5{BVrm|4I3`qz% zU#J9z0L_BK6S*yW*F38Gs@1tDf`6T$<^6sR^3^9a2+~6ojt=r(IG8!%ej2u51RE_- zmhq^T0S*|n>tfnGn1NQEhKPnJD@9$#Qf*^|9lOCID&YYsO zTg|Ze{A*A{iB2!Z&l^Xs=9;D@aWIftaO?QPh6oSQhQzLWKGwJT&uz$MLE#AER@>Q% zi}pAg13^h9_M{)zor77TntI9y>pQyy9L7Eoo_KlhInz?wUR6c|HJ7KlDI!~uzLtK7nQt^Ohs$GFibK-DtP0qc$SbmNZlU6DrVp2zb z$klIDXgP=HunC54+x%QB6|=48P>9NPog>e67f|M#H!0V%Mc_;OA^m_X({~VUSvY_e z1?&K1YkKIM4hdOiuhUGl=b5|72k^~=%F=L3vCr%mH6DP;Aib3;VtN(tKn9&5$n3fH zQ1jxOY2JePY_HlgK_Z2=vCK;1{&7erk5}c)b{!AgFYZ^fi<}?sSBuRO1BpB^YGyg8 zEPA=Y8J;9*@N|?#so!ePnx`ADO^D~@{=+tg@~vJr=rXI3so)E`?sB?ax_!jvi+>El zm5C{95&dUz4{|J72FWeC3-D|Qc*h!0Go0RdOzYX2M(ykL`on#CVM2EtowH&@5EWGE2b0 z=*Op@TvG>*aw;NLvFzjEnrYghmUf|%INU4hjO&OEC&U!bCy|r=O0Fh|jaSL+5tPC9 zA5MFYVs=W{XntP0r`D_%A3Z@q?RPxXW~>Xd=9my9YT<4X$|q)m*=f5a;VVp5REul+ zLBw&K+ktrP99Q2*zjh@C3(m!DJ?vtd4qH6L)n9AkA5Kh*5UVC%sp-gn4c1t^7_9C^ z&=11I5~#8to=TK{B&3%z!{o9@st(l)S{fYHDc~6)kTjV#5%`XahReHAu1j(S~38;Pfw;;c|!zQsMz63b^=4H5hxR0NM3#y)iBR2@vE_Qv+K2qUf78;Ia--f~k;uzdca=p^(BQ`lV>Q(rl4#?uVt(*j{vK*n|N0!rx3Id-42mW*1WULi0EtS|* zZUb#543!~Eu%|_3o|M`H9x}_x05?_h6+fVYXeICUoZmrG;g6Vn1rk|N8ARN-UZb&U z;tMx+nB`z>Bo^;KGEyj6qOFKeueMbv854W)j6!%Sr;D9FH#E{mBG%{YG*!Y#yq6xU zdObV+VrEd&A!eyq7j^AoN0tOu4_^NZb|68xHS_tV5n^@|r%+H+YrJfh-*CknSR(#j8vWKe7ckt0B)*I|h<)SlF$5GEaUkCKbvvf#OwWws6| zGN53QqK>^Qajo}n6K~0Sl8;EIZJbU(oX(>1N3dp ztlVSJ^hY2iPJ4!8Ve5mUuh&%~_}Q!&%0h=-+F{j1wn zd9wT4!1@fw9j84iRT?XdLwo#~x7D<%JEb9xp$;zLM&%RWB=0CHPM>+7HgVtb^&Wk) zj6tYxk73xs^?KM$ zMI%~Q;?PwyLsNv}-t61YZlGO$A|4kxuG_G}LwB*%Sh-93tO6I^T z$cqg_Gu_Kd#PLS1qV#^J;l2dFSw9KvZz&E`iad|4KTq%q6}Z{4Y~Frax{UtPo|(T{ zSYOBKYoSU^7-UE|r0Qj6$`9EcIW`cEgm}Q23{Ee1b-F;6 z-wd`Pl2J#gw~z&Y`tkzyh!sFyMRcM5t8;d~j!iippYB}aBD=7Mtei2|pZ0N-lysv6 z(->S=bmMwDqH9!?^#>B_72?eMZ-Kd<9$%0+bndL+z+ZTenuz;h7-^eMy*Co@y;8tZ zOTSLPW-%G;Jg?G2IP()#1Zb*+l`&j#6ky#b^&AH1r9u%zlQej3gQzacHtA)kn=N*C zDLSYwy*A4*HA)|+9E##vLO+P&?2P0LVn}aV6n4IC8P}|vnkep&=)@<#q+nKTU)DSX zN7q~!0b{0?wGO=bi?txHfP;PZo3=K=GwJ~9$yLCthE<5A0Q21dq6>)o)rfsSbiJPvO=irJI;|4Wa z+|>4%6W6{GM{JvLl&@DAy5bzip`ChNcINCJ+AP(6AT1j zTe4&A9uQJe>$5)hh74L{fI<-vS5-xjhb@KN<=5gFvf3n@8TGXZQXniPQBN}D!BUtN zV@n#ynqrYY^d@@;YfXM|=_lIdQcJ5V(!V)?WsAdxcnj>*d^6kD5BEEinkPAu88VQ~ z5PI9(V@m9v={jbb_3*IDAtsX9EN4`@%r}3rm`EI(EX6l7>@}UjHB&6YTrF*sE^*+P zq>Ebx|8eZi{!00~MnLliTP>WoZIqNgmbBxOJaf{~K?ynw$YNo#;VFn9UJ539l@BkM z@nv&YA#J1eL82EY_uqSXtfMoFnxesL?_&eZx}lWb1f!#p7qf5lqLP%7W5q`bnpzoCRmN<;FJ zP3~rHUW9Kp@H4I&2lyYk^Gj0Klo^M?p=*kDYySDE>C!@~KtU~2mi)8Q)DX-Mu3@B4 zzJPe$c&UYL&U~oh4FVY~(JN4!)~-}-GFO56N9sX7rMKinb^KX)Lu~cD55FYnOba0` zq7Cv#>8zyvjagLyX|KS~8-B?&=Xa^<_iOCnfG@v9Vx7wH=}I+F-O08JNK)`{=}c@i zA>zY(5WWI$*hK>DD3AnZeC(tN?kQzBe=oG6D`>P3(?*L7IIOVH`RK~Y=FPq0nwx$WurA;sD<9!D zS3^a7P(wT|!TI@ffLJ0l>QKp36F_kGvc0f|Pxhz2p`IftzVewV{R zS8FO^r~K#|V;mrDxM$zPgYceZ^ua)fW#j5Hj4iFG_SJ~4{2We1?bXV1-Qws<%9tNB zS_tsN&8e2-3hjI;DrZ@7dWY}~hoS27c4q-TC?r)b_z8Ck}Rh zVTH3K3%!m$z06`I*(3ozwA3IE@GegnD~W3emk3H=^_4td`Xb_ArMv!PeRL0WCh%G zj@thgOmQ|H_et$dzQoGR4oqGMWcMAXxJ;o#(Mt7BC`u@ssf>epEXcIHwwdKs&`e=6 zkE>l`j4QX|tJr!eUMt+}!J}FjaU4n3#s$7~UKAKkK&YGso1P;p*Q{%Lvu$m&W> z;2luOl*=mKM~k_h(O`B8Ya%Ylhnb_ESRtLOrXfQNmuZ@YG@R|^R5V^b9%WuMje>k% zU}PVUZ?Mr~#0GK;=dYz2HPxH1e{3$1IOF-9OGe4T7g3)GS4hkQkEq;dA@abWlBiupMpC5Ifs zwO3EDZab9sS(xHkHM&EVLuV1)sY=f!eCYTX=W}|;AKlbKt}3@QL~h`I#Ng#fUCuZu zTXnC}&$Cw3Bg^sR_NIXdFR_)Ye{Xw`Hb-$+52WO^%qiYUnNSzA+}^0O~x2<;8As5ysrgv6s9-Q0T~;ZxgV=HsMH)g59WQ*zs2fIM5 zsNY*FuPl=;_ydjDNlp&x0}XXzo19lP9HxYlC>k6y4e4|h&8z4Wwu7SKO0kJHMZ-=k zdUm+;72OR-odbl_OXb9y?97T_)BVV?R0!SBG+CBKN%la)8SgQ$6b)6R{AsQoPhjOV z*RpnfX_s#u^7C0!bA^sYD0Z(%k*C!IMqj~cK0OCkcl!U^dzYR|b0g1hU!TGPdXWjJ zM!a9%t-bNi8?S_3$kW}WMyaJisoE}J(BIwh<2mHR!-(g+nS-%`0Y)iEnJ3~Oyga-F z$+%cvwVRt_ahks-IV}4tL&p};!eC#7rJpOp)Q>0O%MIb9;!ZWby;btJ%h^-yZsf*; zZBqJqA;x4b%^7tkW7k;rCSf9E?{W!84h8X!pgX-Am_S67IC?6fiD&qA$=oLd`mw3-xjJ;GZU6w#fMiQ$eP zm&v1b&x;(qjqKjuZ*5+oMN=-lVOw>t6xMI_hB89#jBq!Tf7OtDog)PQmJC}&wN*op ziQ>qSKA{&yRE|=jL6PmAd1e$hbSzQCUyG|3G!}yORYPdzx;c7-BJ*TqnAa2;8HHG) zn$oaP&)%qFE1W|5eUEg4JoKW4ik?AIps0n05vHVgCG^37^`P`5363uYIXt8}%&`%y zb?8b*sl!H)@ZJ_BF?f1A>jR|{(=^|Fhd5q?OWUD~BIl?^lm_B7@gZv$(%&kan+P7*Zr%|qA4Gb4*P_ z^@raH_P@sV!ym!>M8JiGBvY*jyJTHiPDWscr`EHeKjI&czkz#AiLz1VU!Dv<-7<{2 ze%-63gu37V=!Z+nG&K6zpvatzBOl2>esdESUn0{`x4xSku)Tb`U~wm`_WQ;x`#C%6 zR`Xldxv*cu=iAwTo9vmC?x?&E_N!Y&BaAMN8{nrPav+xEFYM}%3I6)pb*oFR**3c* z$^I-G{nC<0MM|&I0b!VZeYSs<#!0+R_BF+;i_}ZALIUA^wLQdg?w7U z3YB%F$r0jJXd)%{2)lU*oiHp`STrVV{*EUt88!_$%mY8pgs?x7K5-Fez8iM!V0(+8S zSaqvKy$jhJl^;2XF-le;u0%@Ju8WeD38%?)$8yE;@JIZM(Ks|;!B5>03$%&LLF#WGlgF1xWa2$4{XC7r{9=pShL z{$^N=;2*^AJ1dv3b>&rJQG^!_crGC1d+*7(INN-AF1vB269OWG4v^W7i0uq`9x(v1^mH zzSl57F-!ecDf?Z>MV0M9RFtHyLh1+=#c6ZOoxj&nzHUuLNtPW`@i%TCo#JmK+6htd zlcQROqvG%6B$GqN$&Wop;@aQ@D8ko^gi{NP%tcz_WIgMCBR$4lQ<#?*$TBj0NSAu2 zfe|;ws|T>8>eKBUdg>9KUIPhANSL`D7`77m6olq^?UG?N+yg1tltlc;2s&|B6_G~J zCoPd*`}Vb#L;}Y#*KX>_1#jWtbppdkdiydJU#dcUJ@O)Ipd7y(*A90j=JC8~lqkEY zL+QCIG0R%U^Xu7oGJzCaY?Q>23GwG3H{X=&7lJUY?1+`da?B9Y@ zaK@%KpOT?qMNiJ=L-c~m+$^jI8p;iqYw>_?rOMd_e}fBpQFur3$qcx>gqK15D# zr5VsFnwp%r@9QJ#p~Li3*Ju1YRZZc>1#1&s+(^h+^L&cGVl$w=VN{@F?;QcJQ*9mdn4zJBW3{DR8ifTXQPsbrY!GKOBD zamHkUVa}jeXbqZTPp%-xpWxodQe$#qxRJ`@f9#^weN@uTFasvxwFyhWK&p z%f-OR`Ii6-1}4?cCH6^qB+C(%zn{|lBiy$v=LSU%TZEc2q*Y=ykjq+OI9K?#)mr!T z>`Whq9raTUu~_D(X#V;X+}8%8^S&{mt5NfF+8>6K+K|m)zX(r_R1v0HxCmcv2p`1z zKn=7@{&qQ=TDp_Nh#*)`y;N9udO;6dq89;oMi(@pKEku;)>hdwvX2HrG&T~WDFpfU zMmF;kfy-+MMvrz`?ug;65W`QFjBN^{DPJ!d@BAc@-e^mroIHs!e$gJoG-=caS8aLD z@nMyRqf|AZ9II>f+ZZKwNxGoiU7~CrEgT-MvR=h*2+dBm-IsxX2rUk$FV)fMk(w@g zXHmjPjbaohhkP7aH9A8y=TsB5J*t^OTi7CgThANaU3#q#stHTX~taL`Zr7|f*vpx*soZ{_Z zW9lZ-(rIWWW`DgOY$QA#xdwH8DyeCzRi7%Fjm+VDp`v8;{p(*)yu)QfxP=l;Ou$vw zM_}ABb%U~&q1WQ7rq~NqZZ0LFaJ+esVu5l*1j~#6Sxgg!RgQ*Cz9k4F%h@QQcid6h z1)PMM1CQsLAS+kToW;+g(Xt9z?Paa_*Tm!Wmkl&8LujONSYBVVB!pPs0OJUrii2NBstqcqweTADDL zl+-y^8eJA-lE-oDb61?q2NK&FLL1iMS&b!P6Oty!-;JelA%Zt^GoW%(9uLcR(1;su zsQ#=^e`aMWYM%AS!&cs=#YQ1iL$@gNP0OIn<>saG&1;Ii5EUX;mHgR4R0FX@{NjHe z;ra)Mk-#~15@yZ%IBI2$=$NaPq7kAm%KpfZQ54U1D}UD%Qek6Zr;711H0b~XBLmP4 z67r(qrfVHWW}R%blvJICa0Uh?5z5IbCysY)3xl@+(L@(TkLfvV z3rV^B^R?9CSY!Tr(Ik5+9(H$Eg%AE<@XJK%TT;aB>C#}o2=jjGB&_?PN%(R@U;*$W zT#s4ux69e?r*^U&b*d<>=lzthkZapBgwMLmd?4t=MVN?67#arSiBN>EizK6$r?;In z#dXFnqhyjEV}W39yo>E$$C0<;3FO|%6VSh(HH1Q9QZ}GdCb6J9@!4L5#tTYZkuhx) zl#_C1Q8}OwhG99AN6?1FKpuvtMm^}mpi9*^jRbQtUrh?}DoqOhqro|YA}W`B(qd35 z2Zh?s#*ND;QDw5f>6gcq6N#k^q0*t5$OxK(X?-b1AETT+VXU2vGuw{GbssO7W2l8sbq3#v7!< z$U4`^;-)bqr^6^gQ^-3J%^Hj#yJig#HlBYJrtm7>vVkF;#JIRnl*-cDO-dgrE(}Uz z1T=3@WRS{%(dg=cA%SNNJZl|@8%L}qjd?^&GWw0cFEM+grP7;D)Euh@DQ04j8Z}7c z3fQVasr+JmpFj*bSB&q&eMKT8hQ<&v9qV?D!EHHutJXwiCY86M5JcU*d^tTs*CZr| zT6wx+SsZf)_v z`yM&FE_IPTp`c1oHfX{^z-N3jm*_>n)l^ms)uCDud{+m*ULaJ3OX42c+4l5;Wl`CTAk2G-dHZr&-y}G84 zk3t8f<00pdLi>Vp;0+@a`3Yrrc6bSzO@)C+T5>I^`iR1u67PQoMULeo8@Hkmxh?IV zcvU3fb>`B2MN%c;_Bno`$i~IMu_&1!+UJNGl-BCVk*z2-6ij7|D%Qk%XS=T=_s08R zyr8lZjTqIp(2iU0#;`##cTVTWy`bb8>?CJVOa56(n9?!GwZ!UqjkSGUVwRg>|FpYBboJg@#^Y#GqZ!)}gm zpv1lM_3UFTeaPvRfr&|@Ie;x12nnLbXPNBLu5%@{+n{xNv(#}Ka0CGMa5%c5Yx!m9ge|L089S7-KrDUn6Ns9GV_ zW4xQRyqx2n?TxV$u%Y0s^0q@77nR$$y;_CWyo=!*S|#(t6BW^Dl_WgBqJZwA`FGD zhmVdAGkU-);^Ay93tQ3^jPHajXaiv+M>9_KPr2 zF(+XiGA7~64dH`ORlG;oC4ak|?G&?qLUFF-wN(ivXEE0hhtRU1a_T3b zzfse`b$!}uj6qL~FNad66n~%3(+ZZki z&-Az&8QKJe}|+! zw8~B=3X*IDrz#3h#fvj~qXjZoMk0zuSr`m#2tjRNf$<|n zZL+#olfLf6&c^mayoBb!DRHag&kqupO34o&rfVeet~-OWz)?sLvln}T@oJWF^adz#U-&wFhx;B!R6|9qHE5A1-1Rz>s4jW*OeT<4ob)GKDs$PFw z2pO18SWPZeG^U5))SxMzi-xhAX%b@<%V}K_*&96`UqN)>lkJv_QhzRm{Z=REm$7cLpjE zo2lsTg~YpeHp|L9d?&Lksa+j+3zPKo98A~XoWm~@%QKZJ*kHLv@1+Nk*p;-JZ?Ax4 z?bU!POQ$0ZfqqZ6_&3UP|J9Va_btkoJBp}Zob}vY>E9vgW&geC#DY&UUG@GeEgj z$~5A6^-W^~$7$Zo4e&6c>aX>6Gt1cEh$B^bvcR&5@C>XP`{bxRl1F04T^zEzx9OKL zI7qxL{h``iF*t6#6snRTMHl-n_qE!UGb5>?b(N<=V^||gX-ALq&9IJ>X$%rW!i|GntqEfIHGf*Xsm;H zq7TXvp-j3)3+4_ur7z0b0X0Q-Q*!F)R&yM|typ6^Zm_8s7)ANI(Go_JFSI3*Th~a! zkwttADeo2Rb>ktPD;c3M!`4``=aszoYLFh9(ZmM_kzbl<`39If`V{gDIKO}~D8gml z0F%@!up8cji&VNqvt;?utcnW!Wv^8 zR*{8tMlRfq4&tyoeDBR7Wo>^f)0)&r&O5V}Jq0xEB zN!v!bue-9&kgWXeV-J8Cx@ZFCJP#{&(V|>qa)v0DivFYZlOd!PFU3%s3toK*^p1v2r5oOkDySveaM{Qo!TOe@MHQ@ zrjgh>WxstdW9Y8I$?S{s%ww^vu^qN8zhK2$npOh0Y%MYlHK6r|DTjHlZ18WCsZ%J* z+RrED%N=ESo#E4dQ~GyEdeJF1x>Qp}PSH9=`H;g`9Q)Xmz9`Zq(k;$to2;ss>=_tY zrHQz2U}VSWT!Lu~7MqQPfCh<^0>Oo}JQj^aZcYP*S5|G_LMhVMok%>+{zANj7TosJ zR?Fx3tr|C?<4lDbhDsPHZ(yh!+SHVNDX*lw zYwryl0hD3-)D~dQ=Jvr8J;~@7WmS)eWuZrdg7;`-A}eV`-IP;_f$g`X zFcu00U|d{j6nPwo#&?oNn1MF3!=3_(jEd1&trjYll1SKNpvawon~;IZxDfSfp-Re? z#EAx~$XSl8Sq6$|#$??#NR8Y@oRu^QQ>&rOO_DK|=4c^&ik#@4kL)JTHQKg)y;WH@ z-1``kvV>^>8HQbnYwO;!ty%mVWhxGevNr8W`Ep13Xd{6@y(|4YB<+gBP8VID7AVC* zc;tL4XCP;^fGo12US5=iX(ORnwJ2m;+xuft7=z?;(&g|lHMU)iA<9a(de^(LTL%N= z$#4=~wNPx)sXGf5@U3B?)F?&zx)GxcHWI^b#7ih%F)3_Sc@BDB_$v7!u2vJBIIf!k zII!UEL02{mr(0PqO;Tmo4#6r?w9_&!@3$h;%#;4$=%Oc-Vu9&{x^=h$2cjUis>jR7 zJtbe56OTcPGUl^YU)heN`Q=T017>u@2{7amONfAuhFm--YKd>;MmdAE{i9wanNzv# zO}XUaof2HT|D?mXEbf!wBFVZo@(CJpD{hF~Zrs)vd-9OY#*&V+)OU@Qc^QXBtiin# z^8Xy9`2rIM&AEM%`1vA{7^Bzw?zk3p4L(U9_KM$I5`Er^*|_T$0%(LD3=42og4D41 zVJ^z51F2zqzA)S$NN-%wCOsQb+ZD7C)+6=cs?eW6e@@wC4g#2Xm8ZfeSAsF25<6h@V3f z)!m2NNJhlkPViPTeiqI(tJ>GSZjbkfr(e0s~#l{OVkwO4XHe0Y8WJsbBNbN6_PA)ICklZ2JX>YulF6m z;nMS-Z-4_gXD^G+367%qC9{hHc0Q5aa0g$S9~=e6#%xh0ctq9n1ri?N-t0ZHP!a*& zs0|i`qJ+I^iX6*_HIo6xk|l+5G$n4}K2~${3v|M6`>6Wi$w3rdS4&UZHxh+AEDwOe z!P9!zbsHgg<|LNO)i*k)?1_jG!F zQzJYIGIrm(N+d48yU8NO5yaOl5@YdR&ia-OoFn7PYAZe?axi$;STbVBPgbRBk!_$S zOD<`jiuZ5iwwgvtdP8m}07LACTwGlvPqnvl!?+Cy)+Z0RZ=NqIJ zQ*F_b+q@%O(wj;VoOxdpHZX2DzP|yJ39*K*=Q^>X2(Pq(HxkyIvJ{!5xqhkIDALW+ zMotXbJwIRC@MziqS4pCOHawp%QAEq^d#gy2uVM;+GZ6f^MiJf58VF>B^4N{u*SC}3!mh8J zW@OFNvxx{zvW?oGBjUbVoF$O7tic6qWJN9(2s3b7(G}Urkh>PVLvGTfo3*fTsPX>u zweab{H~||aliDH}l*WmYDKROl)s5Vau5QIA6!EKNz000~%BG-lHYTh~xW|qwQXELm z8V&Pq(xi2ZB>g;Mgr7wcLL~K(%J`+oWpp(-WIOOpcMrjz!nmojNEsjh-mM~wW2pJN zRt~FeQ3*n_z@lyu9`6kpqwSQ<#y0MMy?X=hvQ~}T1l1xEv$X2;U>I+mWI&2Wk#ZbO z1Z_Kt+`)vb@u65?Ou+i~-Bf)!9FhLos+61_Wzn-baVY4ez3Id-qK9TZn8GliPbHW| z60MKi8Ag%##}hQ_wn#$E4O?C<_nPoZt|Hy61+w6ZcKgY*2*LgFg;r)$l%o(<=B6(UZ+21#1M2&u#0T(zA-;Sa>@ySt% zsG8WcHxAjome&IuJg{M%b5^@@SD9W%{w25fVQX~v`6`Mn#;v+LHjh~?SCRHj=UPeg z&=^YkWYg)KsO59j1-L5Fa>U-}qWrWd6$j^HVB|AI8Ii#v$%tz`TJk*axm*gLq>-Jk zvbHLLysZ#^r$-AV8ypF`EtGT_;him1sW@ME3spx>NN1rc_JP4$C|oIsR&9|;gOQs- z;*d$!x{LJl{8;Pbug@!-CC2sqXYt7J_{Zk}NtZuKCQV=W<2>j`&N?=hKYtq=68ZDH zSZc~Y{D(}J--%}QUt|B_kN)>{Br{mZNQP_~`RkXl!9BU=j`oE8$VN)`;49AP>+ka~ zPx4PUrQ`(+9)sf(iV(_tol! zwGgddNQ66+**^80gn5HVr`ho1FLEMPdLNC-EpQE>>dxH{H!zd15b}?L;fX?UT=t=BL?`G` zP8#t@rdBmV4k~>FIZzNq`MT2zmi;^Jk{GyX@qW=B%V8SE*}3D5u!apg?%YRVhy*HX zs$nCJf5bM(5K5}y{7z^gxv(z=w2=2JwR>pgvAQP%Dyk`gAn&0KXSwZDY!sZ2L=DtJ zyL+SrxfLoWV-nIDLOcCt^^YiVF1_n?QRU>E5V8Rcm^9?0f8quu)iWvYRDFm=>U}Ue zojZ&YPoaB#@Q>-yYDz9L{mYxC$TSQjmeEkz_$n^zw@}i$hhhvpE?{NyR6}SB1eojh zLc5dJMB`IS_heVf{ilxpq4{X2=rw)BJ-UVB&+MHAl>=AYmKqHu4rwh#_rlsew9tdH zDY)BC(MSiAZqrEpw)p97QDmO{R|OhRO;XAvtn#@ymM^hN@I>o#7Z7)T-rp4vAD7(n z3ovd#&ofk(fp{Q;LcW+0rH4~K$9#<(`A&*YH?<8Tbfg?`+SvhNjzG84_uf~GDr^<90k^eoRkx;~lA$wl=! zlyQ{r_d-djDTf;a+KEf={B2VrdE5vrt?I*db|bT?5*q0roUrcd6NwI8^`wl~+}A=6 zioGhxfs0esr>jECy>T=4k#we@cr2KSI7#KN=B=YuX3>hkL z>eiUbV!PGt!c^>=m6MjNZnV zU&l)Y;Ob;pcc)%m<&N@^ODCAYUFqK;=|w-k=$*`fP2bCorja?N zUcwipWl;cRiPKGMg}f!%z{o;fF`CdAoNHOEHO7y~0c6xvAQEW>?o|vF@ftjtfHg{y zzV5_a3z=#*;w3bqxR8f;5}!k~YTU<<#9mm#o8SzEXmiN|^JdAJ<{F_gc~aQG8FR$F zS{q5uZ1{f}a(UjxG4)<XV_D=z#TLBR^L>r<4h2(#*!Gd_H4-wE|F0CjoiSBQ>(MUGSUk9Azr}IZ(y$b zu~RcD6-k%Y%}OQmo5XjON>%jiWPLsoJEO+EsT9#UQ!!gq(>Q1=@RNgWD>s6i~1722>ktG$_W38~RG)X`=_QQ&}D^b$m%2`Lfl zV}le%D7`2JUKG)0$6~4B_g=B2~xawMjDaIpi9(siM`8&h@o6+lb*w3|nSP zE(TA>A8%Q-!{1Oob|{_ za8|01%M27qOKz#f5s2wtE)i|JF?{;SdrD{pOYVrjw|z?`H@cYIP_&fXDF0J^(PIx1 zX?>j@d(Y{N5iIe|@+Ef!OIs5-rkNk_^&Z zcV1ihvLwwm*h^Bx=hquoAtX!Pmn3eiSdyNAr|Op7W3eKG_@u1rc@7;FZA5{{%G}6x zW1yt_^CO&zLWvSM(ySOLsUNumEs`7s9PVHSiA=t@7A=y|fwPN5ClGbLt@;=|R~K)m zlf<^3t#p4Q=gkukG+;eU7;)hY{)@7XKbkU+CX@2zj&eOt*rk7mr2Y7_)1%Mi>`~_* z+V*^(&@Gq`6pOMj^MyiV<`ad+%r^>+nU542v)&>!NRz(q#OFK3M!ba1hl)jfj>Hk1 zA>vAOb*RV9Y)wwgs|X11va0OE@(Q<0OK$8V&)6ZiZeZfY%TE?K%S05_|4pSv%o4UZ zU8zEhw^S{;MUHdCTDzAkt4)hBSYWb1g?jeX-l8r`@XN5hz&!{l)ccY-l87#8fr+I} zK)p9$ve%1xVsr6(4p>Zg%Jk=fK(J)|h&1Cb;JDXrNm3PNGzKmgQ$(2Eq{k~{zBre~BUJ)SSsZrXSBBLDiYSG^& zZTf^kn)}&NWYMHupV&4#;UXl(Cr2CskQ;FWT$F~!;;Ti;_?23ip4~{3v?xE#CK2Ze z^W76OqJdz?7%UR$3C1T?@`(KGl2N@hEC_q&dPb9YLBk%;5E$78ALIK(i9BhriORB) z1shecze#2_u)!!cY9#k$sn?`N;-J0XuWB^RIzm#UJRA(k`=A7pB^J|s=7U0^W+$>W z8YmfgBb=o{8p$D&{$7wme;jTNRFM%Fnu!vnx%qCPis;2-y){U~-PJ5h#5vr$7Ri`d zbM9nfI91i2`rEY;3~*RUTXIKOL-$5*k(CQ?6$4Dzg;>4nsVB$dN+vhgHjX5x_oW|j zCBw~0(c&SEX&vtB#?6duoLC~AZH6RENv(PUa9fk;Oq1j*fApBbBGoxG^SfaI2Qm14 zw&W68xg}FB0giLO9jkrQleFBm9>2;hUXwOGXIazO|&8BYTgQwQ}((zgQ3u}VnJoIB%jI#O;-f1OppY&&F%~=M^G->k8=G6j&c&M zt(>ialu1UJyd^j1T*gb}U3)M_iQ@rFgR8?9IlTcBnhob$qf$(S?xQ7F&g!|oe92R8 zA7RMUr(>45Le%bRk%)IsxLb=9E7`-0+5`&kMi)KDrQ?{GGO2Sf4$()-9Pe5mR6;uhho zu8&gN;JqnutwOq@C+TK^!U<0n-}gQs0J2z&kZDp{8Vm=jz5wXJ%_64f1=V z^b$0Hu;LFFWqti>@L!avlpB<%LcAzn?kFFbfKc5|i1K}scBOoybA0V7Y2)<5Mef*} zjTSV5WsY>*T#vm@e9Nh|$1!*%}&NsTw4cE<3g(14YyTp79NoB7NP75$F3C z;w6+Ah_XUl#OGLy8ux1AQbv{}*^PX)z%odOf%5{(u)Ur(;m8DDb4j)I5t|qDU5kVn zE9V5G$dY)4b#IC!Y6X+rkSk+1ll{_=+gfoW)Ag;~9xHi+EV**Js4t&-kphV>><%Du z8tICS+Kj;3rM~IdFs>8#!Fqus-lQ!YV7GxIFYcnb8kI_pj0k2^$c<9$rldy5)?(Er zd*Z0nH%%##z|qbEizd@Wb9ez?$}$|SjQm~}xRFP0o9Vh_GE6&j2nJDeN3JPDZkJ)i zKVB=9FR-kh(pf$lVBU6naZQ1r99+M)*!C8E5V6Ov&&Ty^6#c*GVGIfHxUFOnpRSOF zWxJp~+q_-So^q=ueb^h)u$U{iw?w~90J}Gi1`}{myc{mZVcR0LN={>o6oxzZCUA8L zsm>$EzpHYhk%z5IxFnTtqKo8`6y=mLq~#_e(X@KVZ9BM=rM_8mGcR1ulq>&6HGETF zJ~5%r47nwbYYdAd-CATySfqxpjnf@F!Z>%r5?U2)#p50OrlJfqS&S^XT~zQ8tu~0( z*$B;R6iE!sLNMRkA|LMBIazW8U$R<#L)`U5d_{vK9Wy*5joKvMu-jRiwuA8tsAf}m zd>8HU6G)bIAy4ibV zWot#ZflKQ#h~r)F6 zkL)N0sdR$vIg3=Whem(UWsxvMRDTO4O9{b44H7{iJ#G(LU?8$eG%HU;+!|~Y4X!py zK4&5{mqCi84uc;J+yAys0VkEI27&?kFFe2(WIx zDg8Sny_`2UI>wFNX=e+ve?}fftI-1aj*YZWi?T4HxT2yPnBHn014>~$0Z{VZi5JOp z6PCRp&kw_cm@HHrio3H=iu83OCeE7eg?I_YjGPEX6%S-H67t< zWsxY+1UGI_4)C(7EF%jXDZP>fZltuzC%Eq5-b&fW$Rf!qlnHBU$d%g)cjb{$b`N$Xk8k_RcW4y2oJR7YsqlF2P@fH@W6{BE}yj_w{YT6LvZ8F8+#NOCm5 z2CYpvDbleFm{`$8;>w;Ksf-gy%Wpha}vz5o}c zV>4n?2wcE4R21(pcrZa&j{9-zqDfc@I@&ZztGT4@b~wXa4dQ#mtFPo>=ScgkH)^kV3_2&MCPg+H51QvLzR_W*qJ9 z{)M9xu+yv*ksWGyS1CDD^O|dcbFCgvuzRp<7QJ4+73_>UQHk$jBfrRl5`Bwnu25$wi3!b5hRlMK0{sGVaW@Upx>Pbqb45EU8a3%=vCHIDisk)d5et_bw#J>x6Yx|?D!23m6LS}3RgXI* zKkm{L$m5uVnYe*Up^UsSi)1uvRojiYp;O_sc%Fff!Uet=p^*>S&tPza7-G#t1n)Kwlj~=E&V670BpM}7>W!%2lTb`{MtAaB+mD|Ho z9l~BiH2vdQpSMHk8{??Stp4FYWZ?YofBi52{wp>XAAk7wzyI=wzx?e#|LHG(`1PNE z`^&%n-~apv8RC(D{M*0$_)q`(_h0$zAO7+u{`c44{^f6f{q^sE_}j1k*WZ5q^*{Z~ zKmG8#7G)^<{~3?%|L1sY#`5vmVEo2VX)nAh;Le+JU&rlR;*9xrhTfUt=Z95Md%mg; z>5fB0xM~*qNZ2T9Z3}spB5Nk`(Ah$q5oYpg_CrQ0M$oZ18Fr!hKHlAM6wLMd4qt5N3am3S09=# z!;?27HjR*F@sYoM68^Xlaxin-XO&-a5&fi*;_;eO7bEspE?$fyyyez6L7L;r#W-g1 zBI42%^>;;2A3^9+_jGL?4+|WVXFZk#Vq=f*`e59;K9nB)-jTCgWSWIX7~i8?2v?ed z@v$Oz6Vd4UNIZV)^D4r*TXW_V22IXArP#1~)W zH6OWyi(w;N@h@dum#z^c`*2&;FGOU)NEhLc3(?1m|J7J~#BGdnoJ#d}#NomohGqSe zaUXG>{B0wf!>q>=&z)bHddpfTR30(7<9LgD7q)3 z>l2p*pyz^iMkhA154d^1GVB7dUsF&X`Ni29^2Duy>B9JuNWayQ0obad@; zoKZg&j^pD<(4~d9apB-}TjlQ&lRni}IbISW!9d95eJqs+qWcmbOXEq9O^tXuN>s-t z$~T1|VZsP!aw23G6F}6}4I}omF|Y%<&C5Qp|KW~s?I0alT9hIzM0BL3G7u->+D2sI z$`;}zEJP+Bl7%=4Yp=*8$TPZxhtrhwhmw`8jn@&4I0+MxQfNG=Duf~|M2;gUr$(HF zi6Esn_P2^qD?u^!5Y>c*Xx!Uah?6i8g!`<)D7@tq6DiA%>ADkmnjdbYr9HLP8JF2{ zm*#^Sm74I=9YJsK$8N0l_v~b7g|hGOq9i%@`&LC;THCv90|ht!#a^6fBl+6i6~pWH zF2@eCdM)y`S$B4ljd1M{D2PoXc&+3%O(F7l0W%O8tveWqQe*&EuMwqUbn6v@!2h^r zD+K=OF~n*hYA<+^XhfIop=^YIxFf9fEEil^wJ1fnwDMshXdq6)LJ&pCs;(j|L~8^w zG7u+WEn9>=wh$zBIc|B?I;zmwDNe@)?kUg!P~cb)s@2JO7neOK-ood3U-_>)Pz$3zTM z{3Ed#&#&srIg%b1PB|rU5wRUD@if?22oNozzL^uxzQI}ldvLmMVM*AKCwI@nT3b45 z=;HjkG>U`w$H_eGgO+xXF=fn?YtWX6vsZt74Ao7^xi~n6j87zIp^-VyJ4~)t)zK;9 z#^}f5O6-Dz{~t{LIiSB;%vo%UB%HbU<>=^r?!S+bZE5z$!G0$tids`J26+|oi_I;4 z@)&EM@&(rOUMAmN>9QxMa5dq!v4e3q5oEl`{yQ+bZ($)}wR=~u-bGuM4eerFL+)h# z@yWXn+T^$cN!M5Us;F`RTh#e`lA5!NUZ^|D7yI`SGG6rQbTfG5@YCeHf64j9{+(V7 zgn?TE{&>U4tzjcZJqeRWTp>KK|=&AToxyq-1(h=kJBV z7B19MEQoO@OC1T2o$PiP*k<`R%g)$Vo_G?mbHz6KE>(!PfFFT^)ce5+*NZ4&*uo3|5*AgT5Cy;Wbp1a0Wy6R) ztKZ8AP+auy*9x2e(SR@C#r@~II`HAYwSx0nI+LIvtoF^Hdmh1~=g(8gBzSnPC%5QK70_93I?64yqxX+~iUwRQSP7sQpKBtygx`F5=5*e(ViXRYS=Yq94*D8H-3RFavk! z?RoxQcv(uwR{Z4K6ewae}E>2V(k0bgr^5kQgCn&KA)xQ(o72fyra1^>H~U#i$@R6P(dr^YW|{a%?cUHx8p zjNa<=i~T#dNMruZX*h3FNlaX?(f^SXuh{F3>ZI=*q@vB=M{MKZ&pJh78gsGzcK!Q$ ziFswew#d2OqF9Iu{EbWellwb`{9` zc1>`?YvCq+l$IIit|??WpWp(vC{4 zHR|nr*Gt+}^syRsQjTe&ojT>P=fR-kw*uwfA87x@N^bWvZ@gt-mW#TjjrvTx6HLuBVFDt z5c_`|J{NFKD=6NN?v;U+{jHwC=Zz^W#!b-(qJyOBbgjr;;YOY?yH=4G3Kumn%2}`V zcJ~<`i4%;^=LtHB`Pp5G<2`n((PcU&T^OU+5=$mcSiY4gBPQ{*qkt#jV|&QHpQ)lS zaXa%mXHp>kKv?YSS~KNR{Vryxz?TZ%Mm(kymZjSX7jpxvo*;f%!G??KR*MrR1A4?{ zov^8uV@K%uBt6T98E@5cCQE9I?;2n&`Q%zs6%x(257z~|LU?NMY<|Lek^@UVVQqWc zNh6N^gssVk^+M|gf#UL;0BRrQB{^h%f- zr7G$q>KI*uAGf8#ZocsR+T$(&Y~!K0Bho768-s(F_JOZ zAKXTz9^5}Z*oT43xmA%}R`#ZEJnb!ZTS>gAwpMO1zCRSH>gkSsa+t@WrO%(|qEpdd zC!J5u3;JlZByCTY7jeG68MsNhCFsLeOFVvr?afY(v*i1bnU-X@tGC1nu-_6t6?eWR zJ^|`2@l!!($z+dqBXH5TT+2-M?-g^N87KRvXU56+Emt$Ea_11gE$uwktyX-~?aXKF z=Y+kgVNKZEO3v+ooi1*zoTF7b3irkJ`?}f17%wP(BG-&m94CJto_vZQy9v*Rcks(P z=0>uzgMUQaZA}9>A+clWLF6K_j7Xb8>N&EH-d)9J62q6pxW6P4+{I|FXo8}!8SVAN zWX{2809$qSMgv4obFn+!fS5;qgU=g4QjF4*)qot*2qvzolSxvOtFDfm{7NOIp(0=z zOnOc=TXEKFqyI!T_4oA@or-6?#kj`-7y~8hdiGvZ9(i$It3&)JkI(Ok;8+}MiW)({ zu;z7NkgkeEpc+A5!s7H-4!qN*mNbFGmfLIf%MIc4opCa9RxBy&y>Av_Zj9%AUK`_4 zv1?m;yC@p_+g1FzY*(Iw_wC~S>i=Z_UNPtDKiNNB{U_r$t^Tue=fZkh+PV6#R(#v) z{|WoK`d`(sR{z^d9@q1$mFql`J2toH?~#Rap#>i0^_(5UOvh!8N6)V!A4MzrRp`K311RmD$4z-E-&Q+3W8zvBmL z^e2G_AG7Zy_>r6wj^Kv)+Of2{=H((C1$dJy2MPzZ(3qz<<8lATiUtAtX6#667vcB zxx`%6u$GwHO0LH<*J6&UmFMrDkJi>>7POuj0>P zCc3c~B}ol4QO|P_tBlJQ&F&$%$!?YVGKiPti+UGJA-k+gyA4o|){ zg0NIWHJrcCT@^>~&=h|PP85Ex_(zz0Qav90IXXusj+4JsuAPheUDWhE9@!2#Rg`u) z>?{5j{YVpX@{{3sXaxs9mYX96#3HCm>)Defk#+oXsyG?!JMA@OZtk@A?}U;3?UnW+ zNL^{)jUH#+0bZqkatF9d{iZv>$^GUIFmTv%wS41>zu3)DzER)b0d9=%S9VoBwRles zdu5-?ky_c0b?REF-!FH@(mo}bcYyQO%-#~pi;AQ%THx|dIJ_X1cfyaM5lL=@WWD?o z5AW#BLwEV6-^L|881_r-|DJI=Jw8i6{6`8aQfK)1!~gmBfBxm~fB2_A{qyhyV3&|NhrsIq`e{!|y^Jr2hrbfBE&_|Lx<4{}-S^j)8_-4*