Compressing Bayes Net CPTs with Persistent Leaky Causes

Yang Xiang, Qian Jiang
Proceedings of the Eighth International Conference on Probabilistic Graphical Models, PMLR 52:535-546, 2016.

Abstract

Non-Impeding Noisy-AND (NIN-AND) Trees (NATs) offer a highly expressive compressed casual model for significantly reducing space and inference time of Bayesian Nets (BNs). A causal model often includes a leaky cause for all causes not explicitly named. A leaky cause may be persistent or not. A conditional probability table (CPT) in a BN often behaves as if there is a persistent leaky cause (PLC). We discuss limitations for not modeling PLC explicitly during compression. We also reveal challenges if PLC is explicitly modeled. We extend an earlier solution that is limited to binary NAT models and is incomplete, to a solution that is applicable to multi-valued NAT models and is complete. We demonstrate the effectiveness of the solution experimentally for compressing general BN CPTs with PLCs.

Cite this Paper


BibTeX
@InProceedings{pmlr-v52-xiang16, title = {Compressing {B}ayes Net {CPT}s with Persistent Leaky Causes}, author = {Xiang, Yang and Jiang, Qian}, booktitle = {Proceedings of the Eighth International Conference on Probabilistic Graphical Models}, pages = {535--546}, year = {2016}, editor = {Antonucci, Alessandro and Corani, Giorgio and Campos}, Cassio Polpo}, volume = {52}, series = {Proceedings of Machine Learning Research}, address = {Lugano, Switzerland}, month = {06--09 Sep}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v52/xiang16.pdf}, url = {https://proceedings.mlr.press/v52/xiang16.html}, abstract = {Non-Impeding Noisy-AND (NIN-AND) Trees (NATs) offer a highly expressive compressed casual model for significantly reducing space and inference time of Bayesian Nets (BNs). A causal model often includes a leaky cause for all causes not explicitly named. A leaky cause may be persistent or not. A conditional probability table (CPT) in a BN often behaves as if there is a persistent leaky cause (PLC). We discuss limitations for not modeling PLC explicitly during compression. We also reveal challenges if PLC is explicitly modeled. We extend an earlier solution that is limited to binary NAT models and is incomplete, to a solution that is applicable to multi-valued NAT models and is complete. We demonstrate the effectiveness of the solution experimentally for compressing general BN CPTs with PLCs.} }
Endnote
%0 Conference Paper %T Compressing Bayes Net CPTs with Persistent Leaky Causes %A Yang Xiang %A Qian Jiang %B Proceedings of the Eighth International Conference on Probabilistic Graphical Models %C Proceedings of Machine Learning Research %D 2016 %E Alessandro Antonucci %E Giorgio Corani %E Cassio Polpo Campos} %F pmlr-v52-xiang16 %I PMLR %P 535--546 %U https://proceedings.mlr.press/v52/xiang16.html %V 52 %X Non-Impeding Noisy-AND (NIN-AND) Trees (NATs) offer a highly expressive compressed casual model for significantly reducing space and inference time of Bayesian Nets (BNs). A causal model often includes a leaky cause for all causes not explicitly named. A leaky cause may be persistent or not. A conditional probability table (CPT) in a BN often behaves as if there is a persistent leaky cause (PLC). We discuss limitations for not modeling PLC explicitly during compression. We also reveal challenges if PLC is explicitly modeled. We extend an earlier solution that is limited to binary NAT models and is incomplete, to a solution that is applicable to multi-valued NAT models and is complete. We demonstrate the effectiveness of the solution experimentally for compressing general BN CPTs with PLCs.
RIS
TY - CPAPER TI - Compressing Bayes Net CPTs with Persistent Leaky Causes AU - Yang Xiang AU - Qian Jiang BT - Proceedings of the Eighth International Conference on Probabilistic Graphical Models DA - 2016/08/15 ED - Alessandro Antonucci ED - Giorgio Corani ED - Cassio Polpo Campos} ID - pmlr-v52-xiang16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 52 SP - 535 EP - 546 L1 - http://proceedings.mlr.press/v52/xiang16.pdf UR - https://proceedings.mlr.press/v52/xiang16.html AB - Non-Impeding Noisy-AND (NIN-AND) Trees (NATs) offer a highly expressive compressed casual model for significantly reducing space and inference time of Bayesian Nets (BNs). A causal model often includes a leaky cause for all causes not explicitly named. A leaky cause may be persistent or not. A conditional probability table (CPT) in a BN often behaves as if there is a persistent leaky cause (PLC). We discuss limitations for not modeling PLC explicitly during compression. We also reveal challenges if PLC is explicitly modeled. We extend an earlier solution that is limited to binary NAT models and is incomplete, to a solution that is applicable to multi-valued NAT models and is complete. We demonstrate the effectiveness of the solution experimentally for compressing general BN CPTs with PLCs. ER -
APA
Xiang, Y. & Jiang, Q.. (2016). Compressing Bayes Net CPTs with Persistent Leaky Causes. Proceedings of the Eighth International Conference on Probabilistic Graphical Models, in Proceedings of Machine Learning Research 52:535-546 Available from https://proceedings.mlr.press/v52/xiang16.html.

Related Material