Introducing lateral thinking in search engines

Abstract. Too much information kills information. This common statement applies to huge databases, where state of the art search engines may retrieve hundreds of very similar documents for a precise query. In fact, this is becoming so problematic that Novartis Pharma, one of the leaders of the pharmaceutical industry, has come up with the somewhat odd request to decrease the precision of their search engine, in order to keep some diversity in the retrieved documents. Rather than decreasing precision by introducing random noise, this paper describes ELISE, an Evolutionary Learning Interactive Search Engine that interactively evolves rewriting modules and rules (some kind of elaborated user profile) along a Parisian Approach[12]. Additional documents are therefore retrieved that are related both to the domains of interest of the user and to the original query, with results that suggest of lateral thinking capabilities.

• Type: Journal
• Year: 2006
• Where: GPEM
• Authors: Yann Landrin-Schweitzer, Pierre Collet, Evelyne Lutton
• Keywords: Art Search Engines, GPEM 2006, Learning Interactive Search, Search Engine