Speaker: Daphne Liu (joint work with Jim Delgrande, Torsten Schaub,
Date and Time: Thursday June 23, 2005 @ 1:30pm
Place: ASB 9705
COBA: A consistency-based approach for belief change
The problem of revision arises when, for example, an intelligent agent has to modify its belief(s) because it has acquired more recent or preferred information. The ability to revise one's knowledge is essential for sound reasoning. While an agent wants to incorporate the new information into its beliefs, it also wishes to retain as much of its existing knowledge as consistently possible and maintain a consistent set of beliefs.
We propose a general, consistency-based framework for expressing belief change, focusing on revision and contraction. This flexible framework is easily extensible to other belief change operations such as the merging of knowledge bases. The revision and contraction operators also have good formal properties while being amenable to implementation. Starting from a high-level algorithmic scheme, we develop a system for expressing and computing belief change. We describe its current implementation, as well as experimental results.