[Return to main page]

IFSA - NAFIPS 2019: Keynote Lectures

James C. Bezdek

Applications of Maximin Sampling In BIG Data Cluster Analysis

A. MaxiMin (MM) Sampling defined. Extended MM sampling (EMMS) by random sampling in neighborhoods of MM Points. The iVAT algorithm defined and exemplified.
B. EMMS and Dunn's index. The MM sampling theorem for CS data sets. Approximating Dunn's index in Big data with EMMS.
C. EMMS and the ClusiVAT clustering algorithm. The scalable single linkage theorem. Numerical comparison of clusiVAT to c-means, single pass c-means, online c-means and BIRCH.

About the Speaker

Jim received the PhD in Applied Mathematics from Cornell University in 1973. Jim is past president of NAFIPS (North American Fuzzy Information Processing Society), IFSA (International Fuzzy Systems Association) and the IEEE CIS (Computational Intelligence Society): founding editor the Int'l. Jo. Approximate Reasoning and the IEEE Transactions on Fuzzy Systems: Life fellow of the IEEE and IFSA; and a recipient of the IEEE 3rd Millennium, IEEE CIS Fuzzy Systems Pioneer, IEEE Rosenblatt and Kampe de Feriet Awards. Jim's interests: clustering in big data, woodworking, optimization, motorcycles, data visualization, cigars, fishing, anomaly detection, blues music, poker. Jim retired in 2007, and will be coming to a university near you soon.

Kelly Cohen

The Promise and Potential of Genetic Fuzzy Systems for Bio-Medical Applications

In recent times, there appears to be a growing interest in “Explainable AI”. A requirement of this approach is to interpret the decision and action especially when applied to safety critical systems such transportation, aerospace and medicine. The current state-of-the-art strategies in machine learning either provide high accuracy or cater to the need for interpretability/explainability. Seldom are we able to successfully address both criteria. However, a genetic fuzzy system comprises of a methodology which provides synergy between linguistic based fuzzy inference systems and genetic algorithms. Upon completion of training genetic fuzzy systems, have the potential to yield high performance while being scalable, robust to noise and uncertainties, transparent, deterministic, and computationally efficient. In this invited talk, the research presented concerning genetic fuzzy systems at the University of Cincinnati will include applications in the area of bipolar disorder treatment, mild traumatic brain injury prognosis and sports medicine in conjunction with mixed reality. Results not only demonstrate feasibility but also provide an insight into the wider applicability of this approach.

About the Speaker

Kelly Cohen is Professor of Aerospace Engineering and Engineering Mechanics in the Department of Aerospace Engineering and Engineering Mechanics at the University of Cincinnati (Ohio USA).

Susana Montes

Divergence measures as a tool for comparing sets and measuring their imprecision

The comparison of sets (of patterns or alternatives) is a crucial task in many different frameworks and it is key for the development of useful applications. In this sense, we have focused our attention on two important lines of research have attracted our attention. The first one involves the comparison of fuzzy sets and of some of their generalizations. In this problem, many different measures of comparison have been proposed in the literature, such as distances or dissimilarities. However, it can be argued that these measures may be inadequate in some cases, and this was one of the motivations underlying our introduction of the notion of divergence measures in 1996. Since then, we have both analyzed some of their theoretical properties and also investigated their usefulness in a number of applications, such as pattern recognition or decision making. Moreover, we have also considered the more general case where the comparison is made over Atanassov intuitionistic fuzzy sets, hesitant fuzzy sets or fuzzy multisets. The second line of research involves the use of divergence measures to derive entropy measures, understood as measures that allow us to compare a set with a certain standard. Depending on the standard that is chosen, we may consider different approaches to the notion of entropy, that in turn are related to other studies carried out in the literature. In this talk, we shall summarize the work we have carried out on this topics during the last twenty years. We shall illustrate the starting points that led us to introduce the notion of divergence measure and summarize the evolution of our studies throughout this period.

About the Speaker

Susana Montes is Professor of Statistics and O.R. at the University of Oviedo, Spain.

Janusz Kacprzyk

Cognitive biases in choice and decision making: a potential role of fuzzy logic

Cognitive biases are a subject of intensive research in psychology, cognitive sciences or economics pioneered (in our context!) by Daniel Kahneman (Nobel Prize winner in economics, 2002). Basically, they are some systematic patterns of deviation from a rationality in judgment and decision which is usually postulated and followed in traditional approaches. For instance, people tend to follow the "status quo bias", i.e. prefer solutions or choices that do not differ from what presently occurs, the "bandwagon" or "herd" bias, that is to tend to do (or believe in) what many other people do, the "outcome bias", that is to tend to judge a decision by its eventual (future) outcome instead on the present quality of the decision, etc. Though the cognitive biases seem to have a clearly negative connotation, the very essence of the "cognitive bias" does not mean that the humans make mistakes or are irrational, they just mean that people make judgments or decisions in the ways that are systematically different from what the traditional economic models, based on traditional concepts of rationality, say. The cognitive biases do not necessarily lead to sub-optimal or bad outcomes. Needless to say that the cognitive biases, as discussed in, for instance, psychology or economics, are inherently fuzzy concepts so that fuzzy logic should be useful for their modeling and analysis.

We will first review some better known and more relevant cognitive biases, notably the status quo bias, related in its spirit to the well known Gärdenfors principle of minimal change, which will be illustrated on some fuzzy decision making and fuzzy optimization models. Then, we will discuss the bandwagon (herd) bias and its role in fuzzy group decision and social choice models. Examples of applications will be briefly described. This should show a high applicability of fuzzy logic as a tool to formally represent and deal with cognitive biases.

About the Speaker

Janusz Kacprzyk is a Polish engineer and mathematician, notable for his multiple contributions to the field of computational and artificial intelligence tools like fuzzy sets, mathematical optimization, decision making under uncertainty, computational intelligence, intuitionistic fuzzy sets, data analysis and data mining, with applications in databases, ICT, mobile robotics and others.

As of 2018, Kacprzyk is professor of computer science at the Systems Research Institute, Polish Academy of Sciences, at Warsaw School of Information Technology, and Chongqing Three Gorges University, Wanzhou, Chinqgqung, China. He is honorary foreign professor at the Department of Mathematics, Yili Normal University, Xinjiang, China, as well as part-time professor of automatic control at Polish Industrial Institute of Automation and Measurements (PIAP) and part-time professor in the Department of Electrical and Computer Engineering, Cracow University of Technology. Kacprzyk has been a frequent visiting professor in the US, Italy, UK, Mexico, China, Japan.

Kacprzyk has authored 6 books, edited or co-edited more than 100 volumes, authored or co-authored approximately 550 papers. He is the editor-in-chief of 7 book series at Springer, and of two journals, and is on the editorial boards of approximately 40 scientific journals. He is a foreign member of the Spanish Royal Academy of Economic and Financial Sciences (2007), of the Bulgarian Academy of Sciences (2013), of the Finnish Society of Sciences and Letters (2018), as well as member of Academia Europaea and the European Academy of Sciences and Arts, and fellow of multiple professional societies, like IEEE, Institution of Engineering and Technology (IET), European Coordinating Committee of Artificial Intelligence (EurAI/ECCAI), IFSA, and Mexican Society for Artificial Intelligence (SMIA). In 2013, he becomes the laureate of the annual IFSA Award.