Stochastic Inference for Scalable Probabilistic Modeling of Binary Matrices

Jose Miguel Hernandez-Lobato, Neil Houlsby, Zoubin Ghahramani
Proceedings of the 31st International Conference on Machine Learning, PMLR 32(2):379-387, 2014.

Abstract

Fully observed large binary matrices appear in a wide variety of contexts. To model them, probabilistic matrix factorization (PMF) methods are an attractive solution. However, current batch algorithms for PMF can be inefficient because they need to analyze the entire data matrix before producing any parameter updates. We derive an efficient stochastic inference algorithm for PMF models of fully observed binary matrices. Our method exhibits faster convergence rates than more expensive batch approaches and has better predictive performance than scalable alternatives. The proposed method includes new data subsampling strategies which produce large gains over standard uniform subsampling. We also address the task of automatically selecting the size of the minibatches of data used by our method. For this, we derive an algorithm that adjusts this hyper-parameter online.

Cite this Paper


BibTeX
@InProceedings{pmlr-v32-hernandez-lobatoa14, title = {Stochastic Inference for Scalable Probabilistic Modeling of Binary Matrices}, author = {Hernandez-Lobato, Jose Miguel and Houlsby, Neil and Ghahramani, Zoubin}, booktitle = {Proceedings of the 31st International Conference on Machine Learning}, pages = {379--387}, year = {2014}, editor = {Xing, Eric P. and Jebara, Tony}, volume = {32}, number = {2}, series = {Proceedings of Machine Learning Research}, address = {Bejing, China}, month = {22--24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v32/hernandez-lobatoa14.pdf}, url = {https://proceedings.mlr.press/v32/hernandez-lobatoa14.html}, abstract = {Fully observed large binary matrices appear in a wide variety of contexts. To model them, probabilistic matrix factorization (PMF) methods are an attractive solution. However, current batch algorithms for PMF can be inefficient because they need to analyze the entire data matrix before producing any parameter updates. We derive an efficient stochastic inference algorithm for PMF models of fully observed binary matrices. Our method exhibits faster convergence rates than more expensive batch approaches and has better predictive performance than scalable alternatives. The proposed method includes new data subsampling strategies which produce large gains over standard uniform subsampling. We also address the task of automatically selecting the size of the minibatches of data used by our method. For this, we derive an algorithm that adjusts this hyper-parameter online.} }
Endnote
%0 Conference Paper %T Stochastic Inference for Scalable Probabilistic Modeling of Binary Matrices %A Jose Miguel Hernandez-Lobato %A Neil Houlsby %A Zoubin Ghahramani %B Proceedings of the 31st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2014 %E Eric P. Xing %E Tony Jebara %F pmlr-v32-hernandez-lobatoa14 %I PMLR %P 379--387 %U https://proceedings.mlr.press/v32/hernandez-lobatoa14.html %V 32 %N 2 %X Fully observed large binary matrices appear in a wide variety of contexts. To model them, probabilistic matrix factorization (PMF) methods are an attractive solution. However, current batch algorithms for PMF can be inefficient because they need to analyze the entire data matrix before producing any parameter updates. We derive an efficient stochastic inference algorithm for PMF models of fully observed binary matrices. Our method exhibits faster convergence rates than more expensive batch approaches and has better predictive performance than scalable alternatives. The proposed method includes new data subsampling strategies which produce large gains over standard uniform subsampling. We also address the task of automatically selecting the size of the minibatches of data used by our method. For this, we derive an algorithm that adjusts this hyper-parameter online.
RIS
TY - CPAPER TI - Stochastic Inference for Scalable Probabilistic Modeling of Binary Matrices AU - Jose Miguel Hernandez-Lobato AU - Neil Houlsby AU - Zoubin Ghahramani BT - Proceedings of the 31st International Conference on Machine Learning DA - 2014/06/18 ED - Eric P. Xing ED - Tony Jebara ID - pmlr-v32-hernandez-lobatoa14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 32 IS - 2 SP - 379 EP - 387 L1 - http://proceedings.mlr.press/v32/hernandez-lobatoa14.pdf UR - https://proceedings.mlr.press/v32/hernandez-lobatoa14.html AB - Fully observed large binary matrices appear in a wide variety of contexts. To model them, probabilistic matrix factorization (PMF) methods are an attractive solution. However, current batch algorithms for PMF can be inefficient because they need to analyze the entire data matrix before producing any parameter updates. We derive an efficient stochastic inference algorithm for PMF models of fully observed binary matrices. Our method exhibits faster convergence rates than more expensive batch approaches and has better predictive performance than scalable alternatives. The proposed method includes new data subsampling strategies which produce large gains over standard uniform subsampling. We also address the task of automatically selecting the size of the minibatches of data used by our method. For this, we derive an algorithm that adjusts this hyper-parameter online. ER -
APA
Hernandez-Lobato, J.M., Houlsby, N. & Ghahramani, Z.. (2014). Stochastic Inference for Scalable Probabilistic Modeling of Binary Matrices. Proceedings of the 31st International Conference on Machine Learning, in Proceedings of Machine Learning Research 32(2):379-387 Available from https://proceedings.mlr.press/v32/hernandez-lobatoa14.html.

Related Material