Active learning for imbalanced datasets - Archive ouverte HAL Accéder directement au contenu
Proceedings/Recueil Des Communications Année : 2020

Active learning for imbalanced datasets

Apprentissage actif avec prise en compte du déséquilibre

Résumé

Active learning increases the effectiveness of labeling when only subsets of unlabeled datasets can be processed manually. To our knowledge, existing algorithms are designed under the assumption that datasets are balanced. However, many real-life datasets are actually imbalanced and we propose two adaptations of active learning to tackle imbalance. First, we modify acquisition functions to select samples by taking advantage of a deep model pretrained on a source domain. Second, we introduce a balancing step in the acquisition process to reduce the imbalance of the labeled subset. Evaluation is done with four imbalanced datasets using existing active learning methods and their modifications introduced here. Results show that our adaptations are useful as long as knowledge from the source domain is transferable to target domains.
Fichier principal
Vignette du fichier
AL_WACV__final__Camera_ready_.pdf (292.86 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04342293 , version 1 (13-12-2023)

Identifiants

Citer

Umang Aggarwal, Adrian Popescu, Céline Hudelot. Active learning for imbalanced datasets. WACV 2020 - IEEE Winter Conference on Applications of Computer Vision, IEEE, pp.1417-1426, 2020, Conference proceedings: 2020 IEEE Winter Conference on Applications of Computer Vision (WACV), 978-1-7281-6553-0. ⟨10.1109/WACV45572.2020.9093475⟩. ⟨hal-04342293⟩
17 Consultations
20 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More