Hung T. Nguyen (Emeritus Professor, New Mexico State University, USA; Adjunct Professor, Chiang Mai University, THAILAND)

Title: On Conditional Event Algebra for Probability Reasoning in System Design

Conditional Event Algebra (CEA) is a novel approach for quantifying probabilistic uncertainty in rule-based systems. It has been investigated by I.R. Goodman and Hung T. Nguyen (back in 1994), to provide a rigorous calculus of uncertain conditionals (uncertain rules in systems, ML and AI), but not very well-known in research communities. As such, this paper is somewhat like a tutorial on the development of CEA for probability reasoning in intelligent systems.

Bio

Hung T. NGUYEN is actually an Emeritus Professor of Mathematical Sciences at New Mexico State University (USA) and an Adjunct Professor of Economics at Chiang Mai University, Thailand. He received his Doctorat d'Etat Es Sciences Mathematiques (Ph.D.) at the University of Lille (France) in 1975. After spending several years at the University of California, Berkeley, and the University of Massachusetts, Amherst, he joined the faculty at New Mexico State University (NMSU), Las Cruces where he retired in 2010, and moved to Thailand where he is currently an Adjunct Professor of Economics at Chiang Mai University. His research covers horizontally a variety of topics under the umbrella of uncertainty analysis, such as Information Theory, Fuzzy sets, Statistics (of di¤usion processes), Probability (Conditional Event Theory, Random sets), Quantum Probability, Belief Functions, and Data Fusion. He is a Fellow of the International Fuzzy Systems Association (IFSA), a Distinguised Lukac Professor of Statistics at Ohio Bowling Green University (2002), a Westhafer Award for Excellence in Research at NMSU, a Distinguised LIFE Chair Professor at Tokyo Institute of Technology (1992-93), and a Distinguished Fellow of the American Society of Engineering Education (ASEE). He has written books on Statistics, Probability, Data Fusion, Random Sets, Conditional Events, and Fuzzy Logics. His current research interests include Quantum Probability and Mathematics of Ambiguity in Econometrics and Finance.

Sebastien Destercke (Université de Technologie de Compiegne, FRANCE)

Title: Information Fusion as a Tool to Estimate Parameters from Imprecise Data

In this talk, I will argue that information fusion tools present interesting features when it comes to estimate parameters when we observe imperfect data, in particular (but not only) in those situations where we can assume a deterministic functional relation between observed inputs and outputs. I will start from the simple yet already interesting case of set-valued data, and then will argue for the need to extend them to more general uncertainty theories, such as possibility or evidence theory. I will illustrate my points on toy examples as well as on more concrete applications involving mechanical parameter estimation and individual user preference learning.

Bio

Destercke SEBASTIEN graduated in 2004 as an engineer from the Faculté Polytechnique de Mons in Belgium. In 2008, he earned a Ph.D. degree in computer science from Université Paul Sabatier, in Toulouse (France). He is currently senior researcher at the French National Research Center (CNRS), in the Heuristic and Diagnostic of Complex Systems (Heudiasyc) Laboratory, in which he is leader of the AI team. His main research interests are in the field of uncertainty reasoning with imprecision-tolerant models (DS theory, imprecise probabilities, possibility theory, ...), with a focus on issues related to reliability and risk analysis, decision making and machine learning. He is currently associate editors of journals such as the International Journal of Approximate Reasoning and Fuzzy Sets and Systems, and is the holder of the SAFE AI chair, that focuses on how to robustify AI methods and make their use safer.

Katsushige Fujimoto (Fukushima University, JAPAN)

Title: Interaction Indices as a Tool to Explaining Machine Learning Models

Describing and visualizing the internal structure of machine learning models to make it understandable to humans why a particular output value was produced is referred to as Explainable AI (XAI). Recently, SHAP (SHapley Additive exPlanations), which utilizes the Shapley value—a concept from cooperative game theory—has been proposed and garnered significant attention. SHAP values reveal how each feature contributes to a given prediction. In this talk, I will explain an extension and generalization of the Shapley value concept: the interaction index, which demonstrates how interactions between features influence predictions.

Bio

Katsushige FUJIMOTO is a full professor in the Department of Symbiotic Systems Science at Fukushima University, where he has been teaching General Mathematics and Computer Science. Since 2024, he has also been an assistant to the President of Fukushima University. He received his B.L.A. degree in Mathematical Science from Osaka Prefecture University in 1990, his M.S. degree and Ph.D. in Systems Sciences from Tokyo institute of Technology in 1992 and 1995, respectively. From 1995 to 1999, he was an assistant professor in the Department of Architecture at Tohoku University. Subsequently, he joined Fukushima University, holding positions as an associate professor in the Faculty of Economics from 1999 to 2004 and in the Department of Symbiotic Systems Science from 2004 to 2013. Since 2013, he has been a full professor in his current role. His research interests were originally mainly fuzzy measures and its integrals, cooperative game theory, decision theory and machine learning. On the application side, his research focuses on the decommissioning of the Fukushima Daiichi Nuclear Power Plant.

Yuchi Kanzawa (Shibaura Institute of Technology, JAPAN)

Title: Fuzzy Clustering: Unifying Algorithms through Theory, Parameterization, and Implementation

In this talk, I will explore fuzzy clustering from three perspectives. The first is to theoretically elucidate the properties of some fuzzy clustering algorithms and whether they still have such properties even after generalizing the original optimization problem. The second is to achieve high clustering accuracy so that the fuzzy clustering algorithm obtained by introducing new parameters to integrate the various fuzzy clustering algorithms can be flexible enough to include the original multiple fuzzy clustering algorithms. The last one is related to the second one, and is to achieve efficient implementation and stable maintenance by focusing on the relationships among multiple fuzzy clustering algorithms: relationships among fuzzification techniques, differences in the nature of the individuals to be clustered, and differences in the degree of dissimilarity introduced, and implementing them on a computer.

Bio

Yuchi KANZAWA is a full professor in the College of Engineering at Shibaura Institute of Technology, specializing in computer science education. Since 2018, he has also taken on the role of Assistant Dean of Engineering. Dr. Kanzawa earned his Bachelor's, Master's, and Ph.D. degrees in Engineering from Waseda University, completing them in 1993, 1995, and 1998, respectively. He began his academic career as a research assistant at Waseda University from 1997 to 2000, before joining Shibaura Institute of Technology as a lecturer. He later became an assistant professor (2004-2007) and then an associate professor (2008-2018), eventually rising to his current position as a full professor. Dr. Kanzawa's research primarily focuses on fuzzy clustering techniques and their implementation in various computational domains, including recommendation systems, classification models, and regression analysis.