Login
Talks & Seminars
Learning to bridge Schema Heterogeneity
Jayant Madhavan, Univ of Washington
Date & Time: April 3, 2003 11:30
Venue: CSE Seminar hall
Abstract:
Data stored in different data sources have different schemas.The relationships between the data in different sources are described by expressions (called semantic mappings) between their schemas. These mappings play a key role in facilitating data sharing in any application involving multiple data sources. The first phase in determining the set of expressions that specify how data from one source maps to another is to find the pairs of concepts in the different schemas that are in some way related to each other. This phase is called Schema Matching, and is the focus of this talk. In this talk, I will describe a progression of schema matching techniques that I have been (and still am) involved in as part of the database and machine learning groups at the University of Washington. The key feature of our techniques are the use of multiple sources of evidence in an extensible framework and the extensive use of machine learning. The focus has been on developing semi-automated approaches that learn from past experience (previous matching tasks) and progressively improve their matching ability. The different projects I will describe are joint work with Phil Bernstein, AnHai Doan, Pedro Domingos, Alon Halevy, and Pradeep Shenoy. Relevant literature: 1. Learning to Match Ontologies on the Semantic Web, Doan,Madhavan,Domingos, Halevy, WWW'2002. 2.www.cs.washington.edu/homes/jayant
Speaker Profile:
Jayant Madhavan Univ of Washington
List of Talks

Webmail

Username:
Password:
Faculty CSE IT
Forgot Password
    [+] Sitemap     Feedback