Structured Output Learning with Abstention: Application to Accurate Opinion Prediction

Alexandre Garcia, Chloé Clavel, Slim Essid, Florence d’Alche-Buc
; Proceedings of the 35th International Conference on Machine Learning, PMLR 80:1695-1703, 2018.

Abstract

Motivated by Supervised Opinion Analysis, we propose a novel framework devoted to Structured Output Learning with Abstention (SOLA). The structure prediction model is able to abstain from predicting some labels in the structured output at a cost chosen by the user in a flexible way. For that purpose, we decompose the problem into the learning of a pair of predictors, one devoted to structured abstention and the other, to structured output prediction. To compare fully labeled training data with predictions potentially containing abstentions, we define a wide class of asymmetric abstention-aware losses. Learning is achieved by surrogate regression in an appropriate feature space while prediction with abstention is performed by solving a new pre-image problem. Thus, SOLA extends recent ideas about Structured Output Prediction via surrogate problems and calibration theory and enjoys statistical guarantees on the resulting excess risk. Instantiated on a hierarchical abstention-aware loss, SOLA is shown to be relevant for fine-grained opinion mining and gives state-of-the-art results on this task. Moreover, the abstention-aware representations can be used to competitively predict user-review ratings based on a sentence-level opinion predictor.

Cite this Paper


BibTeX
@InProceedings{pmlr-v80-garcia18a, title = {Structured Output Learning with Abstention: Application to Accurate Opinion Prediction}, author = {Garcia, Alexandre and Clavel, Chlo{\'e} and Essid, Slim and d'Alche-Buc, Florence}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {1695--1703}, year = {2018}, editor = {Jennifer Dy and Andreas Krause}, volume = {80}, series = {Proceedings of Machine Learning Research}, address = {Stockholmsmässan, Stockholm Sweden}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/garcia18a/garcia18a.pdf}, url = {http://proceedings.mlr.press/v80/garcia18a.html}, abstract = {Motivated by Supervised Opinion Analysis, we propose a novel framework devoted to Structured Output Learning with Abstention (SOLA). The structure prediction model is able to abstain from predicting some labels in the structured output at a cost chosen by the user in a flexible way. For that purpose, we decompose the problem into the learning of a pair of predictors, one devoted to structured abstention and the other, to structured output prediction. To compare fully labeled training data with predictions potentially containing abstentions, we define a wide class of asymmetric abstention-aware losses. Learning is achieved by surrogate regression in an appropriate feature space while prediction with abstention is performed by solving a new pre-image problem. Thus, SOLA extends recent ideas about Structured Output Prediction via surrogate problems and calibration theory and enjoys statistical guarantees on the resulting excess risk. Instantiated on a hierarchical abstention-aware loss, SOLA is shown to be relevant for fine-grained opinion mining and gives state-of-the-art results on this task. Moreover, the abstention-aware representations can be used to competitively predict user-review ratings based on a sentence-level opinion predictor.} }
Endnote
%0 Conference Paper %T Structured Output Learning with Abstention: Application to Accurate Opinion Prediction %A Alexandre Garcia %A Chloé Clavel %A Slim Essid %A Florence d’Alche-Buc %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-garcia18a %I PMLR %J Proceedings of Machine Learning Research %P 1695--1703 %U http://proceedings.mlr.press %V 80 %W PMLR %X Motivated by Supervised Opinion Analysis, we propose a novel framework devoted to Structured Output Learning with Abstention (SOLA). The structure prediction model is able to abstain from predicting some labels in the structured output at a cost chosen by the user in a flexible way. For that purpose, we decompose the problem into the learning of a pair of predictors, one devoted to structured abstention and the other, to structured output prediction. To compare fully labeled training data with predictions potentially containing abstentions, we define a wide class of asymmetric abstention-aware losses. Learning is achieved by surrogate regression in an appropriate feature space while prediction with abstention is performed by solving a new pre-image problem. Thus, SOLA extends recent ideas about Structured Output Prediction via surrogate problems and calibration theory and enjoys statistical guarantees on the resulting excess risk. Instantiated on a hierarchical abstention-aware loss, SOLA is shown to be relevant for fine-grained opinion mining and gives state-of-the-art results on this task. Moreover, the abstention-aware representations can be used to competitively predict user-review ratings based on a sentence-level opinion predictor.
APA
Garcia, A., Clavel, C., Essid, S. & d’Alche-Buc, F.. (2018). Structured Output Learning with Abstention: Application to Accurate Opinion Prediction. Proceedings of the 35th International Conference on Machine Learning, in PMLR 80:1695-1703

Related Material