Spectral Regularization for Max-Margin Sequence Tagging

Ariadna Quattoni, Borja Balle, Xavier Carreras, Amir Globerson
Proceedings of the 31st International Conference on Machine Learning, PMLR 32(2):1710-1718, 2014.

Abstract

We frame max-margin learning of latent variable structured prediction models as a convex optimization problem, making use of scoring functions computed by input-output observable operator models. This learning problem can be expressed as an optimization involving a low-rank Hankel matrix that represents the input-output operator model. The direct outcome of our work is a new spectral regularization method for max-margin structured prediction. Our experiments confirm that our proposed regularization framework leads to an effective way of controlling the capacity of structured prediction models.

Cite this Paper

BibTeX
@InProceedings{pmlr-v32-quattoni14, title = {Spectral Regularization for Max-Margin Sequence Tagging}, author = {Quattoni, Ariadna and Balle, Borja and Carreras, Xavier and Globerson, Amir}, booktitle = {Proceedings of the 31st International Conference on Machine Learning}, pages = {1710--1718}, year = {2014}, editor = {Xing, Eric P. and Jebara, Tony}, volume = {32}, number = {2}, series = {Proceedings of Machine Learning Research}, address = {Bejing, China}, month = {22--24 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v32/quattoni14.pdf}, url = {https://proceedings.mlr.press/v32/quattoni14.html}, abstract = {We frame max-margin learning of latent variable structured prediction models as a convex optimization problem, making use of scoring functions computed by input-output observable operator models. This learning problem can be expressed as an optimization involving a low-rank Hankel matrix that represents the input-output operator model. The direct outcome of our work is a new spectral regularization method for max-margin structured prediction. Our experiments confirm that our proposed regularization framework leads to an effective way of controlling the capacity of structured prediction models.} }
Endnote
%0 Conference Paper %T Spectral Regularization for Max-Margin Sequence Tagging %A Ariadna Quattoni %A Borja Balle %A Xavier Carreras %A Amir Globerson %B Proceedings of the 31st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2014 %E Eric P. Xing %E Tony Jebara %F pmlr-v32-quattoni14 %I PMLR %P 1710--1718 %U https://proceedings.mlr.press/v32/quattoni14.html %V 32 %N 2 %X We frame max-margin learning of latent variable structured prediction models as a convex optimization problem, making use of scoring functions computed by input-output observable operator models. This learning problem can be expressed as an optimization involving a low-rank Hankel matrix that represents the input-output operator model. The direct outcome of our work is a new spectral regularization method for max-margin structured prediction. Our experiments confirm that our proposed regularization framework leads to an effective way of controlling the capacity of structured prediction models.
RIS
TY - CPAPER TI - Spectral Regularization for Max-Margin Sequence Tagging AU - Ariadna Quattoni AU - Borja Balle AU - Xavier Carreras AU - Amir Globerson BT - Proceedings of the 31st International Conference on Machine Learning DA - 2014/06/18 ED - Eric P. Xing ED - Tony Jebara ID - pmlr-v32-quattoni14 PB - PMLR DP - Proceedings of Machine Learning Research VL - 32 IS - 2 SP - 1710 EP - 1718 L1 - http://proceedings.mlr.press/v32/quattoni14.pdf UR - https://proceedings.mlr.press/v32/quattoni14.html AB - We frame max-margin learning of latent variable structured prediction models as a convex optimization problem, making use of scoring functions computed by input-output observable operator models. This learning problem can be expressed as an optimization involving a low-rank Hankel matrix that represents the input-output operator model. The direct outcome of our work is a new spectral regularization method for max-margin structured prediction. Our experiments confirm that our proposed regularization framework leads to an effective way of controlling the capacity of structured prediction models. ER -
APA
Quattoni, A., Balle, B., Carreras, X. & Globerson, A.. (2014). Spectral Regularization for Max-Margin Sequence Tagging. Proceedings of the 31st International Conference on Machine Learning, in Proceedings of Machine Learning Research 32(2):1710-1718 Available from https://proceedings.mlr.press/v32/quattoni14.html.

Related Material