Memory-efficient inference in dynamic graphical models using multiple cores

Galen Andrew, Jeff Bilmes
Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, PMLR 22:47-53, 2012.

Abstract

We introduce the archipelagos algorithm for memory-efficient multi-core inference in dynamic graphical models. By making use of several processors running in parallel, the archipelagos algorithm uses exponentially less memory compared to basic forward-backward message passing algorithms (O(log T) compared to O(T) on sequences of length T) and, under often-satisfied assumptions on the relative speed of passing forward and backward messages, runs no slower. We also describe a simple variant of the algorithm that achieves a factor of two speedup over forward-backward on a single core. Experiments with our implementation of archipelagos for the computation of posterior marginal probabilities in an HMM validate the space/time complexity analysis: using four cores, the required memory on our test problem was reduced from 8 GB to 319 KB (a factor of 25000) relative to forward-backward, but completed in essentially the same time. The archipelagos algorithm applies to any dynamic graphical model, including dynamic Bayesian networks, conditional random fields, and hidden conditional random fields.

Cite this Paper


BibTeX
@InProceedings{pmlr-v22-andrew12, title = {Memory-efficient inference in dynamic graphical models using multiple cores}, author = {Andrew, Galen and Bilmes, Jeff}, booktitle = {Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics}, pages = {47--53}, year = {2012}, editor = {Lawrence, Neil D. and Girolami, Mark}, volume = {22}, series = {Proceedings of Machine Learning Research}, address = {La Palma, Canary Islands}, month = {21--23 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v22/andrew12/andrew12.pdf}, url = {https://proceedings.mlr.press/v22/andrew12.html}, abstract = {We introduce the archipelagos algorithm for memory-efficient multi-core inference in dynamic graphical models. By making use of several processors running in parallel, the archipelagos algorithm uses exponentially less memory compared to basic forward-backward message passing algorithms (O(log T) compared to O(T) on sequences of length T) and, under often-satisfied assumptions on the relative speed of passing forward and backward messages, runs no slower. We also describe a simple variant of the algorithm that achieves a factor of two speedup over forward-backward on a single core. Experiments with our implementation of archipelagos for the computation of posterior marginal probabilities in an HMM validate the space/time complexity analysis: using four cores, the required memory on our test problem was reduced from 8 GB to 319 KB (a factor of 25000) relative to forward-backward, but completed in essentially the same time. The archipelagos algorithm applies to any dynamic graphical model, including dynamic Bayesian networks, conditional random fields, and hidden conditional random fields.} }
Endnote
%0 Conference Paper %T Memory-efficient inference in dynamic graphical models using multiple cores %A Galen Andrew %A Jeff Bilmes %B Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2012 %E Neil D. Lawrence %E Mark Girolami %F pmlr-v22-andrew12 %I PMLR %P 47--53 %U https://proceedings.mlr.press/v22/andrew12.html %V 22 %X We introduce the archipelagos algorithm for memory-efficient multi-core inference in dynamic graphical models. By making use of several processors running in parallel, the archipelagos algorithm uses exponentially less memory compared to basic forward-backward message passing algorithms (O(log T) compared to O(T) on sequences of length T) and, under often-satisfied assumptions on the relative speed of passing forward and backward messages, runs no slower. We also describe a simple variant of the algorithm that achieves a factor of two speedup over forward-backward on a single core. Experiments with our implementation of archipelagos for the computation of posterior marginal probabilities in an HMM validate the space/time complexity analysis: using four cores, the required memory on our test problem was reduced from 8 GB to 319 KB (a factor of 25000) relative to forward-backward, but completed in essentially the same time. The archipelagos algorithm applies to any dynamic graphical model, including dynamic Bayesian networks, conditional random fields, and hidden conditional random fields.
RIS
TY - CPAPER TI - Memory-efficient inference in dynamic graphical models using multiple cores AU - Galen Andrew AU - Jeff Bilmes BT - Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics DA - 2012/03/21 ED - Neil D. Lawrence ED - Mark Girolami ID - pmlr-v22-andrew12 PB - PMLR DP - Proceedings of Machine Learning Research VL - 22 SP - 47 EP - 53 L1 - http://proceedings.mlr.press/v22/andrew12/andrew12.pdf UR - https://proceedings.mlr.press/v22/andrew12.html AB - We introduce the archipelagos algorithm for memory-efficient multi-core inference in dynamic graphical models. By making use of several processors running in parallel, the archipelagos algorithm uses exponentially less memory compared to basic forward-backward message passing algorithms (O(log T) compared to O(T) on sequences of length T) and, under often-satisfied assumptions on the relative speed of passing forward and backward messages, runs no slower. We also describe a simple variant of the algorithm that achieves a factor of two speedup over forward-backward on a single core. Experiments with our implementation of archipelagos for the computation of posterior marginal probabilities in an HMM validate the space/time complexity analysis: using four cores, the required memory on our test problem was reduced from 8 GB to 319 KB (a factor of 25000) relative to forward-backward, but completed in essentially the same time. The archipelagos algorithm applies to any dynamic graphical model, including dynamic Bayesian networks, conditional random fields, and hidden conditional random fields. ER -
APA
Andrew, G. & Bilmes, J.. (2012). Memory-efficient inference in dynamic graphical models using multiple cores. Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 22:47-53 Available from https://proceedings.mlr.press/v22/andrew12.html.

Related Material