AAAI Publications, Workshops at the Thirtieth AAAI Conference on Artificial Intelligence

Font Size: 
Lazy Arithmetic Circuits
Seyed Mehran Kazemi, David Poole

Last modified: 2016-03-29

Abstract


Compiling a Bayesian network into a secondary structure, such as a junction tree or arithmetic circuit allows for offline computations before observations arrive, and quick inference for the marginal of all variables. However, query-based algorithms, such as variable elimination and recursive conditioning, that compute the posterior marginal of few variables given some observations, allow pruning of irrelevant variables, which can reduce the size of the problem. Madsen and Jensen show how lazy evaluation of junction trees can allow both compilation and pruning. In this paper, we adapt the lazy evaluation to arithmetic circuits, allowing the best of both worlds: pruning due to observations and query variables as well as compilation while exploiting local structure and determinism.

Keywords


Probabilistic Graphical Models; Exact Probabilistic Inference; Knowledge Compilation; Arithmetic Circuits; Lazy Evaluation; Network Pruning

Full Text: PDF