On Structured Prediction Theory with Calibrated Convex Surrogate Losses

Anton Osokin 1, 2 Francis Bach 3 Simon Lacoste-Julien 4, 5
1 WILLOW - Models of visual object recognition and scene understanding
Inria de Paris, DI-ENS - Département d'informatique de l'École normale supérieure
3 SIERRA - Statistical Machine Learning and Parsimony
DI-ENS - Département d'informatique de l'École normale supérieure, CNRS - Centre National de la Recherche Scientifique, Inria de Paris
Abstract : We provide novel theoretical insights on structured prediction in the context of efficient convex surrogate loss minimization with consistency guarantees. For any task loss, we construct a convex surrogate that can be optimized via stochastic gradient descent and we prove tight bounds on the so-called “calibration function” relating the excess surrogate risk to the actual risk. In contrast to prior related work, we carefully monitor the effect of the exponential number of classes in the learning guarantees as well as on the optimization complexity. As an interesting consequence, we formalize the intuition that some task losses make learning harder than others, and that the classical 0-1 loss is ill-suited for structured prediction.
Type de document :
Communication dans un congrès
The Thirty-first Annual Conference on Neural Information Processing Systems (NIPS), Dec 2017, Long Beach, United States. 2017
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01611691
Contributeur : Anton Osokin <>
Soumis le : vendredi 6 octobre 2017 - 11:54:23
Dernière modification le : jeudi 26 avril 2018 - 10:29:05

Lien texte intégral

Identifiants

  • HAL Id : hal-01611691, version 1
  • ARXIV : 1703.02403

Collections

Citation

Anton Osokin, Francis Bach, Simon Lacoste-Julien. On Structured Prediction Theory with Calibrated Convex Surrogate Losses. The Thirty-first Annual Conference on Neural Information Processing Systems (NIPS), Dec 2017, Long Beach, United States. 2017. 〈hal-01611691〉

Partager

Métriques

Consultations de la notice

209