Dual Decomposition for Joint Discrete-Continuous Optimization

Christopher Zach
Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, PMLR 31:632-640, 2013.

Abstract

We analyse convex formulations for combined discrete-continuous MAP inference using the dual decomposition method. As a consquence we can provide a more intuitive derivation for the resulting convex relaxation than presented in the literature. Further, we show how to strengthen the relaxation by reparametrizing the potentials, hence convex relaxations for discrete-continuous inference does not share an important feature of LP relaxations for discrete labeling problems: incorporating unary potentials into higher order ones affects the quality of the relaxation. We argue that the convex model for discrete-continuous inference is very general and can be used as alternative for alternation-based methods often employed for such joint inference tasks.

Cite this Paper


BibTeX
@InProceedings{pmlr-v31-zach13a, title = {Dual Decomposition for Joint Discrete-Continuous Optimization}, author = {Zach, Christopher}, booktitle = {Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics}, pages = {632--640}, 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/zach13a.pdf}, url = {https://proceedings.mlr.press/v31/zach13a.html}, abstract = {We analyse convex formulations for combined discrete-continuous MAP inference using the dual decomposition method. As a consquence we can provide a more intuitive derivation for the resulting convex relaxation than presented in the literature. Further, we show how to strengthen the relaxation by reparametrizing the potentials, hence convex relaxations for discrete-continuous inference does not share an important feature of LP relaxations for discrete labeling problems: incorporating unary potentials into higher order ones affects the quality of the relaxation. We argue that the convex model for discrete-continuous inference is very general and can be used as alternative for alternation-based methods often employed for such joint inference tasks. } }
Endnote
%0 Conference Paper %T Dual Decomposition for Joint Discrete-Continuous Optimization %A Christopher Zach %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-zach13a %I PMLR %P 632--640 %U https://proceedings.mlr.press/v31/zach13a.html %V 31 %X We analyse convex formulations for combined discrete-continuous MAP inference using the dual decomposition method. As a consquence we can provide a more intuitive derivation for the resulting convex relaxation than presented in the literature. Further, we show how to strengthen the relaxation by reparametrizing the potentials, hence convex relaxations for discrete-continuous inference does not share an important feature of LP relaxations for discrete labeling problems: incorporating unary potentials into higher order ones affects the quality of the relaxation. We argue that the convex model for discrete-continuous inference is very general and can be used as alternative for alternation-based methods often employed for such joint inference tasks.
RIS
TY - CPAPER TI - Dual Decomposition for Joint Discrete-Continuous Optimization AU - Christopher Zach 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-zach13a PB - PMLR DP - Proceedings of Machine Learning Research VL - 31 SP - 632 EP - 640 L1 - http://proceedings.mlr.press/v31/zach13a.pdf UR - https://proceedings.mlr.press/v31/zach13a.html AB - We analyse convex formulations for combined discrete-continuous MAP inference using the dual decomposition method. As a consquence we can provide a more intuitive derivation for the resulting convex relaxation than presented in the literature. Further, we show how to strengthen the relaxation by reparametrizing the potentials, hence convex relaxations for discrete-continuous inference does not share an important feature of LP relaxations for discrete labeling problems: incorporating unary potentials into higher order ones affects the quality of the relaxation. We argue that the convex model for discrete-continuous inference is very general and can be used as alternative for alternation-based methods often employed for such joint inference tasks. ER -
APA
Zach, C.. (2013). Dual Decomposition for Joint Discrete-Continuous Optimization. Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 31:632-640 Available from https://proceedings.mlr.press/v31/zach13a.html.

Related Material