From: Rolf Fagerberg Sender: daimi-alle-admin@daimi.au.dk To: daimi-alle@daimi.au.dk Subject: [daimi-alle] Daimi Seminar, Dec 12, 15.15-17.00, Aud. D. Date: Wed, 11 Dec 2002 14:03:53 +0100 DAIMI SEMINAR René DePont Christensen, Thomas Rask Thomsen Dreamgate REAL LIFE SEARCH ENGINE DEVELOPMENT The speakers have been involved in the development of search- and retrieval software, under real life demands. Working to achieve a good balance between recall and relevance whilst providing scalability and speed, it has been extremely important to incorporate the latest scientific research. Link structure has been exploited for both crawling and ranking using various Kleinberg/PageRank mutants. An extremely expressive query language has been invented to encode Naive Bayesian Classifiers, this enables the efficient generation of statistical document profiles, which support a very intuitive search approach. The PageRank and Kleinberg algorithms and their concepts are sketched, and the approach to mutating them is described contrasted with real world limitations. Naive Bayesian Classifiers and the query language approach to search engine construction is explained. Theoretical prerequisites are minimal, the talk focuses on the practical approach to building search software that is both scientifically rigorous and real-time efficient as well as user-friendly. Time: Thursday 12th December 15.15 - 17.00 Place: Ny Munkegade, Aud. D4 Hosts: Gerth S. Brodal and Rolf Fagerberg -------- René DePont Christensen holds a Ph.D degree in mathematics and an M.S. degree in mathematics and computer science from Odense University. Thomas Rask Thomsen holds a M.S. degree in computer science and mathematics from Aalborg University. They both are co-founders of the software company Dreamgate, which specializes in software for indexing and searching unstructured textual data, including documents on the web. This talk is a guest lecture in the course "Algorithms for Web Indexing and Searching", but doubles as a DAIMI seminar. Everybody is welcome. -- Rolf Fagerberg e-mail: rolf@daimi.au.dk Department of Computer Science web: www.daimi.au.dk/~rolf University of Aarhus phone: +45 89 42 34 70 Ny Munkegade, Bldg. 540 local phone: 3470 DK-8000 Aarhus C fax: +45 89 42 32 55 Denmark home phone: +45 86 28 10 04