Communication Dans Un Congrès Année : 2025

UNEM: UNrolled Generalized EM for Transductive Few-Shot Learning

Résumé

Transductive few-shot learning has recently triggered wide attention in computer vision. Yet, current methods introduce key hyper-parameters, which control the prediction statistics of the test batches, such as the level of class balance, affecting performances significantly. Such hyper-parameters are empirically grid-searched over validation data, and their configurations may vary substantially with the target dataset and pre-training model, making such empirical searches both sub-optimal and computationally intractable. In this work, we advocate and introduce the unrolling paradigm, also referred to as "learning to optimize", in the context of few-shot learning, thereby learning efficiently and effectively a set of optimized hyperparameters. Specifically, we unroll a generalization of the ubiquitous Expectation-Maximization (EM) optimizer into a neural network architecture, mapping each of its iterates to a layer and learning a set of key hyper-parameters over validation data. Our unrolling approach covers various statistical feature distributions and pre-training paradigms, including recent foundational vision-language models and standard vision-only classifiers. We report comprehensive experiments, which cover a breadth of fine-grained downstream image classification tasks, showing significant gains brought by the proposed unrolled EM algorithm over iterative variants. The achieved improvements reach up to 10% and 7.5% on vision-only and vision-language benchmarks, respectively. The source code and learned parameters are available at https://github.com/ZhouLong0/ UNEM-Transductive.

Fichier principal
Vignette du fichier
Transductive_FSL_unrolling.pdf (4.69 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05092880 , version 1 (02-06-2025)

Licence

Identifiants

  • HAL Id : hal-05092880 , version 1

Citer

Long Zhou, Fereshteh Shakeri, Aymen Sadraoui, Mounir Kaaniche, Jean-Christophe Pesquet, et al.. UNEM: UNrolled Generalized EM for Transductive Few-Shot Learning. CVPR 2025 - Computer Vision and Pattern Recognition Conference, Jun 2025, Nashville (Tenessee), United States. ⟨hal-05092880⟩
127 Consultations
164 Téléchargements

Partager

  • More