Informatics, TU Vienna

Algorithmic knowledge discovery with lattices of closed descriptions

Lattices of closed descriptions started in early 1970s as Galois lattices (treillis de Galois) for describing classes of objects given by binary object-attribute data tables.

Abstract

Lattices of closed descriptions started in early 1970s as Galois lattices (treillis de Galois) for describing classes of objects given by binary object-attribute data tables. In 1980s the Darmstadt school started a big research program on Formal Concept Analysis (FCA). Since that time lattices of closed descriptions proved their usefulness in many domains of computer science, artificial intelligence and data mining, for different types of descriptions, including numerical data, data given by strings, graphs, etc. In my talk I will concentrate on algorithms and algorithmic problems of computing concept lattices, bases of exact and inexact dependencies (implications and association rules), for different types of data (from binary to graph-based).

Biography

Sergei O. Kuznetsov obtained his diploma degree at Moscow Institute for Physics and Technology in 1985, and his doctor degree at the All-Russian Institute for Scientific and Technical Information (VINITI) in 1990 for the work on algebraic and algorithmic issues of methods of data analysis. In 1999-2001 he was a Humboldt fellow at TU-Dresden (Institute for Algebra). Since 2006 Sergei is the head of the department of Data Analysis and Artificial Intelligence at the National Research University Higher School of Economics, Moscow. His research is related to algebraic and algorithmic issues of Formal Concept Analysis, data analysis and knowledge discovery.

Note

This talk is organized by the Theory and Logic Group at the Institute of Computer Languages.