Speaker: Chris Cornelis (Gent University, Belgium)

Abstract: Since its inception in 1965 by Zadeh, fuzzy set theory has risen to prominence as a comprehensive and competitive branch of science dealing with the representation and processing of imperfect (e.g. vague, incomplete, unreliable, ...) information, the seminal idea being that of partial membership of an element to a set or, equivalently, partial truth of a statement. In adopting the terminology "fuzzy logic", one should be cautious to distinguish between two different meanings conveyed by it: - fuzzy logic in the narrow sense is a fully-fledged many-valued symbolic logic centered around the notion of graded syntax and semantics - fuzzy logic in the broad sense is an umbrella term uniting a panoply of approaches and methodologies tolerant of vagueness In my presentation, I will survey a number of major building blocks relevant to both directions, including a) the use of various kinds of partially ordered sets for representing truth values and membership degrees and b) the construction and classification of graded logical connectives. Furthermore, emphasizing the "broad" direction, I will show how a synergy of logical and relational structures can help to model approximate, human-like reasoning.