Abstract
Non-Impeding Noisy-AND (NIN-AND) Tree (NAT) models offer a highly expressive approximate representation for significantly reducing the space of Bayesian Nets (BNs). They can also significantly improve efficiency of BN inference, as shown for binary NAT models. To enable these advantages for general BNs, advancements on three technical challenges are made in this work. We overcome the limitation of well-defined Pairwise Causal Interaction (PCI) bits and present a flexible PCI pattern extraction from general target Conditional Probability Tables (CPTs). We extend parameter estimation for binary NAT models to constrained gradient descent for compressing target CPTs into multi-valued NAT models. The effectiveness of the compression is demonstrated experimentally. A novel framework is also developed for PCI pattern extraction when persistent leaky causes exist.
Keywords
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Jensen, F.V., Lauritzen, S.L., Olesen, K.G.: Bayesian updating in causal probabilistic networks by local computations. Comput. Stat. Q. 4, 269–282 (1990)
Madsen, A.L., D’Ambrosio, B.: A factorized representation of independence of causal influence, lazy propagation. Int. J. Uncertainty Fuzziness Knowl.-Based Syst. 8(2), 151–166 (2000)
Madsen, A.L., Jensen, F.V.: Lazy propagation: A junction tree inference algorithm based on lazy evaluation. Artif. Intell. 113(1–2), 203–245 (1999)
Pearl, J.: Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kaufmann, Burlington (1988)
Savicky, P., Vomlel, J.: Exploiting tensor rank-one decomposition in probabilistic inference. Kybernetika 43(5), 747–764 (2007)
Takikawa, M., D’Ambrosio, B.: Multiplicative factorization of noisy-max. In: Proceedings of 15th Conference Uncertainty in Artificial Intelligence, pp. 622–630 (1999)
Yang, X.: Acquisition and computation issues with NIN-AND tree models. In: Myllymaki, P., Roos, T., Jaakkola, T., (eds.) Proceedings of 5th European Workshop on Probabilistic Graphical Models, Finland, pp. 281–289 (2010)
Yang, X.: Bayesian network inference with NIN-AND tree models. In: Cano, A., Gomez-Olmedo, M., Nielsen, T.D. (eds.) Proceedings 6th European Workshop on Probabilistic Graphical Models, Granadapp, pp. 363–370 (2012)
Xiang, Y.: Non-impeding noisy-and tree causal models over multi-valued variables. Int. J. Approx. Reason. 53(7), 988–1002 (2012)
Xiang, Y., Li, Y., Zhu, Z.J.: Towards effective elicitation of NIN-AND tree causal models. In: Godo, L., Pugliese, A. (eds.) SUM 2009. LNCS, vol. 5785, pp. 282–296. Springer, Heidelberg (2009)
Xiang, Y., Liu, Q.: Compression of Bayesian networks with NIN-AND tree modeling. In: van der Gaag, L.C., Feelders, A.J. (eds.) PGM 2014. LNCS, vol. 8754, pp. 551–566. Springer, Heidelberg (2014)
Xiang, Y., Truong, M.: Acquisition of causal models for local distributions in Bayesian networks. IEEE Trans. Cybern. 44(9), 1591–1604 (2014)
Zagorecki, A., Druzdzel, M.J.: Knowledge engineering for Bayesian networks: How common are noisy-MAX distributions in practice? IEEE Trans. Syst. Man Cybern. Syst. 43(1), 186–195 (2013)
Zhang, N., Poole, D.: Exploiting causal independence in Bayesian network inference. J. Artif. Intell. Res. 5, 301–328 (1996)
Acknowledgement
Financial support from NSERC Discovery Grant is acknowledged. We thank anonymous reviewers. We apologize for not moving explanations of figures and tables from text to captions, as it does not appear feasible to us.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this paper
Cite this paper
Xiang, Y., Jiang, Q. (2016). Compression of General Bayesian Net CPTs. In: Khoury, R., Drummond, C. (eds) Advances in Artificial Intelligence. Canadian AI 2016. Lecture Notes in Computer Science(), vol 9673. Springer, Cham. https://doi.org/10.1007/978-3-319-34111-8_35
Download citation
DOI: https://doi.org/10.1007/978-3-319-34111-8_35
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-34110-1
Online ISBN: 978-3-319-34111-8
eBook Packages: Computer ScienceComputer Science (R0)