Plenary Lecture
Consistent Aggregation of Information in Intelligent Systems
Professor János Fodor
Óbuda University
Budapest, Hungary
E-mail: fodor@uni-obuda.hu
Abstract: In this plenary talk we give an overview of some of our contributions to the problem of information aggregation. The framework is provided by intelligent systems, which are human engineered systems that show intelligent behavior or features.
The need for fusing different types of information items arises naturally in diverse theoretical and practical problems. The process of combining several (numerical) values into a single representative one is called aggregation, and the function performing this process is called aggregation function.
Even if the information items can be of different types, it is possible to reinterpret them in a unique formal setting by means of profiles, which are extensions of fuzzy set membership functions. Then the original aggregation problem can be modeled by an appropriate profile aggregation. Based on this approach, in this talk we concentrate on aggregation functions as mappings that assign a single output in the closed unit interval to several inputs from the same interval.
The problem of consistent aggregation, as our starting point and guide in the talk, is formulated in a tabular form that is easy to understand, and to justify as a natural requirement. Its mathematical formulation leads to a general functional equation of several functions and several variables. Well-known aggregation classes can be interpreted in this framework, satisfying particular forms of the general equation.
The resulted aggregation functions can be interpreted either as logical connectives originated from many-valued logics, with typical examples of triangular norms and conorms, uninorms and nullnorms, or as averaging operators allowing a compensation effect such as the arithmetic mean, which play a key role in probability and other classical areas.
In the first part of this talk we give an overview of consistent aggregation functions. In particular, we reveal their fundamental properties and identify four main classes: conjunctive, disjunctive, internal, and mixed aggregation functions. In each class we identify and characterize prototypical subclasses with illustrative examples.
In the second part of the talk we sketch someimportant and interesting applications of consistent aggregations.
Brief Biography of the Speaker: János Fodor is full professor of mathematics at Óbuda University, Budapest, Hungary. He is Doctor of the Hungarian Academy of Sciences. He has been pursuing research in mathematical foundations of fuzzy logic, computational intelligence, preference modelling, inference, and uncertainty management since 1987. He is co-author of two monographs and over 250 papers. These works received more than 4000 citations.
He acts as president of the Hungarian Fuzzy Association, chair of the IEEE Hungary Section Chapter of Computational Intelligence Society, and coordinator of the EUROFUSE EURO Working Group on Fuzzy Sets.
He is elected as Fellow of IFSA in 2013. He is Doctor Honoris Causa of the ”Politehnica” University of Timisoara (Romania).
He has presented papers at more than 150 international and domestic conferences. He has delivered numerous plenary and invited talks, and acted as General Chair, Program Committee Chair or Member at diverse scientific international conferences.
He is Editor-in-Chief of Acta Polytechnica Hungarica, Area Editor of Fuzzy Sets and Systems, member of the Editorial Advisory Board of the International Journal of Advanced Intelligence Paradigms, editor of the European Journal of Operational Research, Computing and Informatics, Acta Mechanica Slovaca, ROMAI Journal on Computer Science, Transactions on Automatic Control and Computer Science. He is Co-Editor-in-Chief of Springer Book Series entitled Topics in Intelligent Engineering and Informatics. He has been a coordinator of several research projects. He has been Invited Professor at various universities in Belgium, Italy, France and Spain.