Statistical and Computational Tradeoffs in Stochastic Composite Likelihood

Joshua Dillon, Guy Lebanon
Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics, PMLR 5:129-136, 2009.

Abstract

Maximum likelihood estimators are often of limited practical use due to the intensive computation they require. We propose a family of alternative estimators that maximize a stochastic variation of the composite likelihood function. We prove the consistency of the estimators, provide formulas for their asymptotic variance and computational complexity, and discuss experimental results in the context of Boltzmann machines and conditional random fields. The theoretical and experimental studies demonstrate the effectiveness of the estimators in achieving a predefined balance between computational complexity and statistical accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v5-dillon09a, title = {Statistical and Computational Tradeoffs in Stochastic Composite Likelihood}, author = {Dillon, Joshua and Lebanon, Guy}, booktitle = {Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics}, pages = {129--136}, year = {2009}, editor = {van Dyk, David and Welling, Max}, volume = {5}, series = {Proceedings of Machine Learning Research}, address = {Hilton Clearwater Beach Resort, Clearwater Beach, Florida USA}, month = {16--18 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v5/dillon09a/dillon09a.pdf}, url = {https://proceedings.mlr.press/v5/dillon09a.html}, abstract = {Maximum likelihood estimators are often of limited practical use due to the intensive computation they require. We propose a family of alternative estimators that maximize a stochastic variation of the composite likelihood function. We prove the consistency of the estimators, provide formulas for their asymptotic variance and computational complexity, and discuss experimental results in the context of Boltzmann machines and conditional random fields. The theoretical and experimental studies demonstrate the effectiveness of the estimators in achieving a predefined balance between computational complexity and statistical accuracy.} }
Endnote
%0 Conference Paper %T Statistical and Computational Tradeoffs in Stochastic Composite Likelihood %A Joshua Dillon %A Guy Lebanon %B Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2009 %E David van Dyk %E Max Welling %F pmlr-v5-dillon09a %I PMLR %P 129--136 %U https://proceedings.mlr.press/v5/dillon09a.html %V 5 %X Maximum likelihood estimators are often of limited practical use due to the intensive computation they require. We propose a family of alternative estimators that maximize a stochastic variation of the composite likelihood function. We prove the consistency of the estimators, provide formulas for their asymptotic variance and computational complexity, and discuss experimental results in the context of Boltzmann machines and conditional random fields. The theoretical and experimental studies demonstrate the effectiveness of the estimators in achieving a predefined balance between computational complexity and statistical accuracy.
RIS
TY - CPAPER TI - Statistical and Computational Tradeoffs in Stochastic Composite Likelihood AU - Joshua Dillon AU - Guy Lebanon BT - Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics DA - 2009/04/15 ED - David van Dyk ED - Max Welling ID - pmlr-v5-dillon09a PB - PMLR DP - Proceedings of Machine Learning Research VL - 5 SP - 129 EP - 136 L1 - http://proceedings.mlr.press/v5/dillon09a/dillon09a.pdf UR - https://proceedings.mlr.press/v5/dillon09a.html AB - Maximum likelihood estimators are often of limited practical use due to the intensive computation they require. We propose a family of alternative estimators that maximize a stochastic variation of the composite likelihood function. We prove the consistency of the estimators, provide formulas for their asymptotic variance and computational complexity, and discuss experimental results in the context of Boltzmann machines and conditional random fields. The theoretical and experimental studies demonstrate the effectiveness of the estimators in achieving a predefined balance between computational complexity and statistical accuracy. ER -
APA
Dillon, J. & Lebanon, G.. (2009). Statistical and Computational Tradeoffs in Stochastic Composite Likelihood. Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 5:129-136 Available from https://proceedings.mlr.press/v5/dillon09a.html.

Related Material