MODELING LATENT VARIABLE UNCERTAINTY FOR LOSS-BASED LEARNING

Abstract : We consider the problem of parameter estimation using weakly supervised datasets, where a training sample consists of the input and a partially specified annotation, which we refer to as the output. The missing information in the annotation is modeled using latent variables. Previous methods overburden a single distribution with two separate tasks: (i) modeling the uncertainty in the latent variables during training; and (ii) making accurate predictions for the output and the latent variables during testing. We propose a novel framework that separates the demands of the two tasks using two distributions: (i) a conditional distribution to model the uncertainty of the latent variables for a given input-output pair; and (ii) a delta distribution to predict the output and the latent variables for a given input. During learning, we encourage agreement between the two distributions by minimizing a loss-based dissimilarity coefficient. Our approach generalizes latent SVM in two important ways: (i) it models the uncertainty over latent variables instead of relying on a pointwise estimate; and (ii) it allows the use of loss functions that depend on latent variables, which greatly increases its applicability. We demonstrate the efficacy of our approach on two challenging problems---object detection and action detection---using publicly available datasets.
Type de document :
Communication dans un congrès
ICML, Jun 2012, Edinburgh, United Kingdom. 2012
Liste complète des métadonnées

Littérature citée [14 références]  Voir  Masquer  Télécharger

https://hal.inria.fr/hal-00773605
Contributeur : M. Pawan Kumar <>
Soumis le : lundi 14 janvier 2013 - 13:39:59
Dernière modification le : jeudi 29 mars 2018 - 13:36:02
Document(s) archivé(s) le : lundi 15 avril 2013 - 04:00:53

Fichier

KPK-ICML2012.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-00773605, version 1

Collections

Citation

M. Pawan Kumar, Ben Packer, Daphne Koller. MODELING LATENT VARIABLE UNCERTAINTY FOR LOSS-BASED LEARNING. ICML, Jun 2012, Edinburgh, United Kingdom. 2012. 〈hal-00773605〉

Partager

Métriques

Consultations de la notice

227

Téléchargements de fichiers

154