Open-Set Likelihood Maximization for Few-Shot Learning - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Open-Set Likelihood Maximization for Few-Shot Learning

Résumé

We tackle the Few-Shot Open-Set Recognition (FSOSR) problem, i.e. classifying instances among a set of classes for which we only have a few labeled samples, while simultaneously detecting instances that do not belong to any known class. We explore the popular transductive setting, which leverages the unlabelled query instances at inference. Motivated by the observation that existing transductive methods perform poorly in open-set scenarios, we propose a generalization of the maximum likelihood principle, in which latent scores down-weighing the influence of potential outliers are introduced alongside the usual parametric model. Our formulation embeds supervision constraints from the support set and additional penalties discouraging overconfident predictions on the query set. We proceed with a blockcoordinate descent, with the latent scores and parametric model co-optimized alternately, thereby benefiting from each other. We call our resulting formulation Open-Set Likelihood Optimization (OSLO). OSLO is interpretable and fully modular; it can be applied on top of any pre-trained model seamlessly. Through extensive experiments, we show that our method surpasses existing inductive and transductive methods on both aspects of open-set recognition, namely inlier classification and outlier detection. Code is available at https://github.com/ebennequin/fewshot-open-set.
Fichier principal
Vignette du fichier
Boudiaf_Open-Set_Likelihood_Maximization_for_Few-Shot_Learning_CVPR_2023_paper.pdf (1.44 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04327201 , version 1 (06-12-2023)

Identifiants

  • HAL Id : hal-04327201 , version 1

Citer

Malik Boudiaf, Etienne Bennequin, Myriam Tami, Antoine Toubhans, Pablo Piantanida, et al.. Open-Set Likelihood Maximization for Few-Shot Learning. CVPR, 2023, Vancouver, Canada. ⟨hal-04327201⟩
14 Consultations
4 Téléchargements

Partager

Gmail Facebook X LinkedIn More