======================================================== 第87回 日本知能情報ファジィ学会 関西支部例会 開催案内 ======================================================== ■主 催：日本知能情報ファジィ学会 関西支部 ■日 時：2009年12月 3日(木) 13:00～17:00 ■会 場：大阪大学大学院基礎工学研究科 J棟1階J120室(共用セミナー室) 〒560-8531 豊中市待兼山町1-3(大阪大学豊中キャンパス内) http://www.es.osaka-u.ac.jp/access/index.html 大阪モノレール「柴原駅」より徒歩7分 　　　　　阪急宝塚線「石橋駅」より徒歩23分 　　　　　阪急宝塚線「蛍池駅」よりタクシーで10分 　　　　　お車での御来場は御遠慮下さい． 　　　　　次のキャンパス内マップの40番の建物です． http://www.osaka-u.ac.jp/migr/img/jp/annai/about/map/images/toyonaka_map.gif ■講演題目：(詳しい内容は下の講演概要をご覧下さい) ● 講演１ 13:00～14:15 「Fuzzy Similarity-based Reasoning: Logical Formalizations and Some Applications」 Lluis Godo 氏(IIIA-CSIC, Spain) ● 講演２ 14:25～15:40 「Why Fuzzy and Possibilistic Optimization Should be 　　　　at the Core of all Operations Research and Optimization 　　　　Programs」 Weldon A. Lodwick 氏(University of Colorado Denver, USA) ● 講演３ 15:50～17:05 「Granular Computing for Reasoning about Ordinal Data: 　　　　　　　　 the Dominance-based Rough Set Perspective」 Roman Slowinski 氏(Poznan University of Technology, Poland) ■参加費：無料 ■連絡先：大阪大学大学院基礎工学研究科 乾口 雅弘 Tel : 06-6850-6350 E-mail : inuiguti ＠ sys.es.osaka-u.ac.jp 大阪大学大学院基礎工学研究科 松本 裕二 Tel : 06-6850-6352 E-mail：matsumoto ＠ sys.es.osaka-u.ac.jp ■講演概要 ● 講演１: Lluis Godo 氏(IIIA-CSIC, Spain) 「Fuzzy Similarity-based Reasoning: Logical Formalizations and Some Applications」 Similarity-based reasoning aims at studying which kinds of logical consequence relations make sense when taking into account that some propositions may be closer to be true than others. A typical kind of inference which is in the scope of similarity-based reasoning responds to the form "if p is true then q is close to be true". In the literature one can find qualitative (or comparative) and quantitative approaches to similarity-based reasoning. Comparative approaches aim at formalizing e.g. expressions like "p is closer to q than r", e.g. like those of Lewis (1973) or Williamson (1988). Quantitative approaches, that are based somehow on a numerical definition of degree of truthlikeness or similarity following Niiniluoto (1987) and Weston (1987). This kind of approach, although not always within a formal logical framework, has been further developed by making use of Zadeh's fuzzy similarity relations as graded modelings of similarity relations, originally to be used in techniques of categorization and clustering. A key contribution in this direction was made by Ruspini (1991) who pushed forward the idea of similarity as one of possible semantics for fuzzy sets, where membership degrees are understood as similarity degrees to some of the prototypes of the given fuzzy set (Dubois and Prade, 1994). The aim of this talk is to survey a class of logical formalizations of similarity-based reasoning models we have been working on (Esteva et al., 1997; Dubois et al., 1997, Godo and Rodriguez, 2006, Esteva et al. 2009). We basically focus on semantically-oriented approaches based on several notions of approximate entailments, providing some formalisations in terms of suitable systems of modal and conditional logics. As a matter of example, we will also succinctly describe one application (Martins et al., 2009) showing similarity-based inference patterns at work in a classification problem. ● 講演２: Weldon A. Lodwick 氏(University of Colorado Denver, USA) 「Why Fuzzy and Possibilistic Optimization Should be at the Core of all Operations Research and Optimization Programs」 It is arguably true that all optimization is satisficing. Since we consider this as being the case, fuzzy and possibilistic optimization sits at the heart of satisficing method, more so than stochastic methods. Real-valued (deterministic) and stochastic (probabilistic) optimization models have data and theoretical assumption requirements in any practical model that are extremely stringent which are rarely if ever satisfied. The case will be made for fuzzy and possibilistic optimization to be at the center of optimization curriculum by clearly distinguishing these two methods from themselves and showing what their relationship is to one another and to more traditional optimization approaches. ● 講演３: Roman Slowinski 氏(Poznan University of Technology, Poland) 「Granular Computing for Reasoning about Ordinal Data: the Dominance-based Rough Set Perspective」 Dominance-based Rough Set Approach (DRSA) to granular computing and data analysis was first introduced as a generalization of the rough set approach for dealing with multiple criteria decision analysis, where preference order in value sets of attributes has to be taken into account. The ordering is also important, however, in many other problems of data analysis. Even when the ordering seems irrelevant, the presence or the absence of a property has an ordinal interpretation, because if two properties are related, the presence, rather than the absence, of one property should make more (or less) probable the presence of the other property. This is even more apparent when the presence or the absence of a property is graded or fuzzy, because in this case, the more credible the presence of a property, the more (or less) probable the presence of the other property. Since the presence of properties, possibly fuzzy, is the basis of any granulation, DRSA can be seen as a general framework for granular computing. The lecture is organized as follows. First, we introduce DRSA in the context of decision making. The main ideas sketching a philosophical basis of DRSA as an approach to granular computing are followed by their extensions in the fuzzy context and in the probabilistic setting. This prepares the ground for defining the rough approximation of a fuzzy set, which is the core of the subject. We also explain why the classical rough set approach is a specific case of DRSA. The presentation continues with presentation of DRSA for case-based reasoning, where granular computing based on DRSA has been successfully applied.