Consensus Message Passing for Layered Graphical Models

Varun Jampani, S. M. Ali Eslami, Daniel Tarlow, Pushmeet Kohli, John Winn
; Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics, PMLR 38:425-433, 2015.

Abstract

Generative models provide a powerful framework for probabilistic reasoning. However, in many domains their use has been hampered by the practical difficulties of inference. This is particularly the case in computer vision, where models of the imaging process tend to be large, loopy and layered. For this reason bottom-up conditional models have traditionally dominated in such domains. We find that widely-used, general-purpose message passing inference algorithms such as Expectation Propagation (EP) and Variational Message Passing (VMP) fail on the simplest of vision models. With these models in mind, we introduce a modification to message passing that learns to exploit their layered structure by passing ’consensus’ messages that guide inference towards good solutions. Experiments on a variety of problems show that the proposed technique leads to significantly more accurate inference results, not only when compared to standard EP and VMP, but also when compared to competitive bottom-up conditional models.

Cite this Paper


BibTeX
@InProceedings{pmlr-v38-jampani15, title = {{Consensus Message Passing for Layered Graphical Models}}, author = {Varun Jampani and S. M. Ali Eslami and Daniel Tarlow and Pushmeet Kohli and John Winn}, booktitle = {Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics}, pages = {425--433}, year = {2015}, editor = {Guy Lebanon and S. V. N. Vishwanathan}, volume = {38}, series = {Proceedings of Machine Learning Research}, address = {San Diego, California, USA}, month = {09--12 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v38/jampani15.pdf}, url = {http://proceedings.mlr.press/v38/jampani15.html}, abstract = {Generative models provide a powerful framework for probabilistic reasoning. However, in many domains their use has been hampered by the practical difficulties of inference. This is particularly the case in computer vision, where models of the imaging process tend to be large, loopy and layered. For this reason bottom-up conditional models have traditionally dominated in such domains. We find that widely-used, general-purpose message passing inference algorithms such as Expectation Propagation (EP) and Variational Message Passing (VMP) fail on the simplest of vision models. With these models in mind, we introduce a modification to message passing that learns to exploit their layered structure by passing ’consensus’ messages that guide inference towards good solutions. Experiments on a variety of problems show that the proposed technique leads to significantly more accurate inference results, not only when compared to standard EP and VMP, but also when compared to competitive bottom-up conditional models.} }
Endnote
%0 Conference Paper %T Consensus Message Passing for Layered Graphical Models %A Varun Jampani %A S. M. Ali Eslami %A Daniel Tarlow %A Pushmeet Kohli %A John Winn %B Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2015 %E Guy Lebanon %E S. V. N. Vishwanathan %F pmlr-v38-jampani15 %I PMLR %J Proceedings of Machine Learning Research %P 425--433 %U http://proceedings.mlr.press %V 38 %W PMLR %X Generative models provide a powerful framework for probabilistic reasoning. However, in many domains their use has been hampered by the practical difficulties of inference. This is particularly the case in computer vision, where models of the imaging process tend to be large, loopy and layered. For this reason bottom-up conditional models have traditionally dominated in such domains. We find that widely-used, general-purpose message passing inference algorithms such as Expectation Propagation (EP) and Variational Message Passing (VMP) fail on the simplest of vision models. With these models in mind, we introduce a modification to message passing that learns to exploit their layered structure by passing ’consensus’ messages that guide inference towards good solutions. Experiments on a variety of problems show that the proposed technique leads to significantly more accurate inference results, not only when compared to standard EP and VMP, but also when compared to competitive bottom-up conditional models.
RIS
TY - CPAPER TI - Consensus Message Passing for Layered Graphical Models AU - Varun Jampani AU - S. M. Ali Eslami AU - Daniel Tarlow AU - Pushmeet Kohli AU - John Winn BT - Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics PY - 2015/02/21 DA - 2015/02/21 ED - Guy Lebanon ED - S. V. N. Vishwanathan ID - pmlr-v38-jampani15 PB - PMLR SP - 425 DP - PMLR EP - 433 L1 - http://proceedings.mlr.press/v38/jampani15.pdf UR - http://proceedings.mlr.press/v38/jampani15.html AB - Generative models provide a powerful framework for probabilistic reasoning. However, in many domains their use has been hampered by the practical difficulties of inference. This is particularly the case in computer vision, where models of the imaging process tend to be large, loopy and layered. For this reason bottom-up conditional models have traditionally dominated in such domains. We find that widely-used, general-purpose message passing inference algorithms such as Expectation Propagation (EP) and Variational Message Passing (VMP) fail on the simplest of vision models. With these models in mind, we introduce a modification to message passing that learns to exploit their layered structure by passing ’consensus’ messages that guide inference towards good solutions. Experiments on a variety of problems show that the proposed technique leads to significantly more accurate inference results, not only when compared to standard EP and VMP, but also when compared to competitive bottom-up conditional models. ER -
APA
Jampani, V., Eslami, S.M.A., Tarlow, D., Kohli, P. & Winn, J.. (2015). Consensus Message Passing for Layered Graphical Models. Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics, in PMLR 38:425-433

Related Material