A Bayesian metareasoner for algorithm selection for real-time Bayesian network inference problems
Haipeng Guo · National Conference on Artificial Intelligence · 2002
Bayesian network (BN) inference has long been seen as a very important and hard problem in AI. Both exact and approximate BN inference are NP-hard [Co90, Sh94]. To date researchers have developed many different kinds of exact and approximate BN inference algorithms. Each of these has different properties and works better for different classes of inference problems. Given a BN inference problem instance, it is usually hard but important to decide in advance which algorithm among a set of choices is the most appropriate. This problem is known as the algorithm selection problem [Ri76]. The goal of this research is to design and implement a meta-level reasoning system that acts as a “BN inference expert” and is able to quickly select the most appropriate algorithm for any given Bayesian network inference problem, and then predict the run time performance.