Cost-Based Budget Active Learning for Deep Learning - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Cost-Based Budget Active Learning for Deep Learning

Résumé

Majorly classical Active Learning (AL) approach usually uses statistical theory such as entropy and margin to measure instance utility, however it fails to capture the data distribution information contained in the unlabeled data. This can eventually cause the classifier to select outlier instances to label. Meanwhile, the loss associated with misla-beling an instance in a typical classification task is much higher than the loss associated with the opposite error. To address these challenges, we propose a Cost-Based Bugdet Active Learning (CBAL) which considers the classification uncertainty as well as instance diversity in a population constrained by a budget. A principled approach based on the min-max is considered to minimize both the labeling and decision cost of the selected instances, this ensures a near-optimal results with significantly less computational effort. Extensive experimental results show that the proposed approach outperforms several state-of-the-art active learning approaches.
Fichier principal
Vignette du fichier
BAL_DL_C_LD_final2.pdf (256.1 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02983761 , version 1 (30-10-2020)

Identifiants

  • HAL Id : hal-02983761 , version 1

Citer

Patrick K Gikunda, Nicolas Jouandeau. Cost-Based Budget Active Learning for Deep Learning. 9th European Starting AI Researchers’ Symposium (STAIRS), Aug 2020, Saint-Jacques-de-Compostelle, Spain. ⟨hal-02983761⟩
36 Consultations
61 Téléchargements

Partager

Gmail Facebook X LinkedIn More