Bethe Bounds and Approximating the Global Optimum

Adrian Weller, Tony Jebara
Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, PMLR 31:618-631, 2013.

Abstract

Inference in general Markov random fields (MRFs) is NP-hard, though identifying the maximum a posteriori (MAP) configuration of pairwise MRFs with submodular cost functions is efficiently solvable using graph cuts. Marginal inference, however, even for this restricted class, is #P-hard. Restricting to binary pairwise models, we prove new formulations of derivatives of the Bethe free energy, provide bounds on the derivatives and bracket the locations of stationary points. Several results apply whether the model is associative or not. Applying these to discretized pseudo-marginals in the associative case, we present a polynomial time approximation scheme for global optimization of the Bethe free energy provided the maximum degree ∆=O(\log n), where n is the number of variables. Runtime is guaranteed O(ε^-3/2 n^6 Σ^3/4 Ω^3/2), where Σ=O(∆/n) is the fraction of possible edges present and Ωis a function of MRF parameters. We examine use of the algorithm in practice, demonstrating runtime that is typically much faster, and discuss several extensions.

Cite this Paper


BibTeX
@InProceedings{pmlr-v31-weller13a, title = {Bethe Bounds and Approximating the Global Optimum}, author = {Weller, Adrian and Jebara, Tony}, booktitle = {Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics}, pages = {618--631}, year = {2013}, editor = {Carvalho, Carlos M. and Ravikumar, Pradeep}, volume = {31}, series = {Proceedings of Machine Learning Research}, address = {Scottsdale, Arizona, USA}, month = {29 Apr--01 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v31/weller13a.pdf}, url = {http://proceedings.mlr.press/v31/weller13a.html}, abstract = {Inference in general Markov random fields (MRFs) is NP-hard, though identifying the maximum a posteriori (MAP) configuration of pairwise MRFs with submodular cost functions is efficiently solvable using graph cuts. Marginal inference, however, even for this restricted class, is #P-hard. Restricting to binary pairwise models, we prove new formulations of derivatives of the Bethe free energy, provide bounds on the derivatives and bracket the locations of stationary points. Several results apply whether the model is associative or not. Applying these to discretized pseudo-marginals in the associative case, we present a polynomial time approximation scheme for global optimization of the Bethe free energy provided the maximum degree ∆=O(\log n), where n is the number of variables. Runtime is guaranteed O(ε^-3/2 n^6 Σ^3/4 Ω^3/2), where Σ=O(∆/n) is the fraction of possible edges present and Ωis a function of MRF parameters. We examine use of the algorithm in practice, demonstrating runtime that is typically much faster, and discuss several extensions.} }
Endnote
%0 Conference Paper %T Bethe Bounds and Approximating the Global Optimum %A Adrian Weller %A Tony Jebara %B Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2013 %E Carlos M. Carvalho %E Pradeep Ravikumar %F pmlr-v31-weller13a %I PMLR %P 618--631 %U http://proceedings.mlr.press/v31/weller13a.html %V 31 %X Inference in general Markov random fields (MRFs) is NP-hard, though identifying the maximum a posteriori (MAP) configuration of pairwise MRFs with submodular cost functions is efficiently solvable using graph cuts. Marginal inference, however, even for this restricted class, is #P-hard. Restricting to binary pairwise models, we prove new formulations of derivatives of the Bethe free energy, provide bounds on the derivatives and bracket the locations of stationary points. Several results apply whether the model is associative or not. Applying these to discretized pseudo-marginals in the associative case, we present a polynomial time approximation scheme for global optimization of the Bethe free energy provided the maximum degree ∆=O(\log n), where n is the number of variables. Runtime is guaranteed O(ε^-3/2 n^6 Σ^3/4 Ω^3/2), where Σ=O(∆/n) is the fraction of possible edges present and Ωis a function of MRF parameters. We examine use of the algorithm in practice, demonstrating runtime that is typically much faster, and discuss several extensions.
RIS
TY - CPAPER TI - Bethe Bounds and Approximating the Global Optimum AU - Adrian Weller AU - Tony Jebara BT - Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics DA - 2013/04/29 ED - Carlos M. Carvalho ED - Pradeep Ravikumar ID - pmlr-v31-weller13a PB - PMLR DP - Proceedings of Machine Learning Research VL - 31 SP - 618 EP - 631 L1 - http://proceedings.mlr.press/v31/weller13a.pdf UR - http://proceedings.mlr.press/v31/weller13a.html AB - Inference in general Markov random fields (MRFs) is NP-hard, though identifying the maximum a posteriori (MAP) configuration of pairwise MRFs with submodular cost functions is efficiently solvable using graph cuts. Marginal inference, however, even for this restricted class, is #P-hard. Restricting to binary pairwise models, we prove new formulations of derivatives of the Bethe free energy, provide bounds on the derivatives and bracket the locations of stationary points. Several results apply whether the model is associative or not. Applying these to discretized pseudo-marginals in the associative case, we present a polynomial time approximation scheme for global optimization of the Bethe free energy provided the maximum degree ∆=O(\log n), where n is the number of variables. Runtime is guaranteed O(ε^-3/2 n^6 Σ^3/4 Ω^3/2), where Σ=O(∆/n) is the fraction of possible edges present and Ωis a function of MRF parameters. We examine use of the algorithm in practice, demonstrating runtime that is typically much faster, and discuss several extensions. ER -
APA
Weller, A. & Jebara, T.. (2013). Bethe Bounds and Approximating the Global Optimum. Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 31:618-631 Available from http://proceedings.mlr.press/v31/weller13a.html.

Related Material