Hinge-loss Markov Random Fields: Convex Inference for Structured Prediction

Stephen Bach, Bert Huang, Ben London, Lise Getoor
Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence, PMLR R11:272-281, 2013.

Abstract

Graphical models for structured domains are powerful tools, but the computational com- plexities of combinatorial prediction spaces can force restrictions on models, or re- quire approximate inference in order to be tractable. Instead of working in a combina- torial space, we use hinge-loss Markov ran- dom fields (HL-MRFs), an expressive class of graphical models with log-concave density functions over continuous variables, which can represent confidences in discrete predic- tions. This paper demonstrates that HL- MRFs are general tools for fast and accu- rate structured prediction. We introduce the first inference algorithm that is both scalable and applicable to the full class of HL-MRFs, and show how to train HL-MRFs with several learning algorithms. Our experiments show that HL-MRFs match or surpass the predic- tive performance of state-of-the-art methods, including discrete models, in four application domains.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR11-bach13a, title = {Hinge-loss {M}arkov Random Fields: Convex Inference for Structured Prediction}, author = {Bach, Stephen and Huang, Bert and London, Ben and Getoor, Lise}, booktitle = {Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence}, pages = {272--281}, year = {2013}, editor = {Nicholson, Ann and Smyth, Padhraic}, volume = {R11}, series = {Proceedings of Machine Learning Research}, month = {12--14 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/r11/main/assets/bach13a/bach13a.pdf}, url = {https://proceedings.mlr.press/r11/bach13a.html}, abstract = {Graphical models for structured domains are powerful tools, but the computational com- plexities of combinatorial prediction spaces can force restrictions on models, or re- quire approximate inference in order to be tractable. Instead of working in a combina- torial space, we use hinge-loss Markov ran- dom fields (HL-MRFs), an expressive class of graphical models with log-concave density functions over continuous variables, which can represent confidences in discrete predic- tions. This paper demonstrates that HL- MRFs are general tools for fast and accu- rate structured prediction. We introduce the first inference algorithm that is both scalable and applicable to the full class of HL-MRFs, and show how to train HL-MRFs with several learning algorithms. Our experiments show that HL-MRFs match or surpass the predic- tive performance of state-of-the-art methods, including discrete models, in four application domains.}, note = {Reissued by PMLR on 04 October 2026.} }
Endnote
%0 Conference Paper %T Hinge-loss Markov Random Fields: Convex Inference for Structured Prediction %A Stephen Bach %A Bert Huang %A Ben London %A Lise Getoor %B Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence %C Proceedings of Machine Learning Research %D 2013 %E Ann Nicholson %E Padhraic Smyth %F pmlr-vR11-bach13a %I PMLR %P 272--281 %U https://proceedings.mlr.press/r11/bach13a.html %V R11 %X Graphical models for structured domains are powerful tools, but the computational com- plexities of combinatorial prediction spaces can force restrictions on models, or re- quire approximate inference in order to be tractable. Instead of working in a combina- torial space, we use hinge-loss Markov ran- dom fields (HL-MRFs), an expressive class of graphical models with log-concave density functions over continuous variables, which can represent confidences in discrete predic- tions. This paper demonstrates that HL- MRFs are general tools for fast and accu- rate structured prediction. We introduce the first inference algorithm that is both scalable and applicable to the full class of HL-MRFs, and show how to train HL-MRFs with several learning algorithms. Our experiments show that HL-MRFs match or surpass the predic- tive performance of state-of-the-art methods, including discrete models, in four application domains. %Z Reissued by PMLR on 04 October 2026.
APA
Bach, S., Huang, B., London, B. & Getoor, L.. (2013). Hinge-loss Markov Random Fields: Convex Inference for Structured Prediction. Proceedings of the 29th Conference on Uncertainty in Artificial Intelligence, in Proceedings of Machine Learning Research R11:272-281 Available from https://proceedings.mlr.press/r11/bach13a.html. Reissued by PMLR on 04 October 2026.

Related Material