Algorithms for Irrelevance-Based Partial MAPs
This work addresses the need for efficient explanation methods in belief networks, but it appears incremental as it builds on existing MAP algorithms.
The paper tackled the problem of computing irrelevance-based partial MAPs for domain-independent explanation in belief networks by examining two definitions and proving key properties, resulting in a modified best-first algorithm for effective computation.
Irrelevance-based partial MAPs are useful constructs for domain-independent explanation using belief networks. We look at two definitions for such partial MAPs, and prove important properties that are useful in designing algorithms for computing them effectively. We make use of these properties in modifying our standard MAP best-first algorithm, so as to handle irrelevance-based partial MAPs.