Professor Elie Sanchez
University of Marseille, France

Title:  Weighted Queries and Fuzzy Information Retrieval Systems

Abstract:
Information retrieval is facing tremendous technological challenges in
storage and intelligent retrieval of multimedia information.  Boolean
query models are used by the majority of the commerical information
retrieval systems.  But these models lack the ability to handle
imprecision and subjectivity.  In fact, fuzziness come in several ways
in fuzzy knowledge/information_retrieval systems: i) fuzzy indexing for
shades of "aboutness", ii) retriveal status values, iii)fuzzy (weighted) 
queries and iv) ranked about output documents.

Interestingly, fuzzy models take into account the lacking features of
graded relevance and weighting of queries.  After discussing weighted
queries, a model is presented for optimization of queries by means of a
genetic algorithm.  An open discussion will follow on the future
application of this approach to DNA and other large documents.

When:  Tuesday 8th September at 10am, 

Where:  Commerce 2.07