Learning Fast-Inference Bayesian Networks

Abstract

We propose new methods for learning Bayesian networks (BNs) that reliably support fast inference. We utilize maximum state space size as a more fine-grained measure for the BN's reasoning complexity than the standard treewidth measure, thereby accommodating the possibility that variables range over domains of different sizes. Our methods combine heuristic BN structure learning algorithms with the recently introduced MaxSAT-powered local improvement method (Peruvemba Ramaswamy and Szeider, AAAI'21). Our experiments show that our new learning methods produce BNs that support significantly faster exact probabilistic inference than BNs learned with treewidth bounds.

PDF NeurIPS OpenReview Code Semantic Scholar

Cite

Text

Ramaswamy and Szeider. "Learning Fast-Inference Bayesian Networks." Neural Information Processing Systems, 2021.

Markdown

[Ramaswamy and Szeider. "Learning Fast-Inference Bayesian Networks." Neural Information Processing Systems, 2021.](https://mlanthology.org/neurips/2021/ramaswamy2021neurips-learning/)

BibTeX

@inproceedings{ramaswamy2021neurips-learning,
  title     = {{Learning Fast-Inference Bayesian Networks}},
  author    = {Ramaswamy, Vaidyanathan Peruvemba and Szeider, Stefan},
  booktitle = {Neural Information Processing Systems},
  year      = {2021},
  url       = {https://mlanthology.org/neurips/2021/ramaswamy2021neurips-learning/}
}