ABAE: Auxiliary Balanced AutoEncoder for class-imbalanced semi-supervised learning

Semi-supervised learning has achieved extraordinary success in prevalent image-classification benchmarks. However, a class-balanced distribution that differs notably from real-world data distribution is required. In general, models trained under class-imbalanced semi-supervised learning conditions a...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:Pattern recognition letters Ročník 182; s. 118 - 124
Hlavní autori: Tang, Qianying, Wei, Xiang, Su, Qi, Zhang, Shunli
Médium: Journal Article
Jazyk:English
Vydavateľské údaje: Elsevier B.V 01.06.2024
Predmet:
Autoencoder
Class-imbalanced learning
Label propagation
Semi-supervised learning
Class-imbalanced learning
Label propagation
Semi-supervised learning
Autoencoder
ISSN:0167-8655, 1872-7344
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:Semi-supervised learning has achieved extraordinary success in prevalent image-classification benchmarks. However, a class-balanced distribution that differs notably from real-world data distribution is required. In general, models trained under class-imbalanced semi-supervised learning conditions are severely biased towards the majority classes. To address this issue, we propose a novel framework called ABAE by implanting an Auxiliary Balanced AutoEncoder branch into existing semi-supervised learning algorithms. Considering that adaptive feature augmentation for different classes can alleviate confirmation bias, we devise a class-aware reconstruction loss to train the AutoEncoder module. To further smooth the output, we adopt a graph-based label propagation scheme at the end of the AutoEncoder. Extensive experiments on CIFAR-10/100-LT, SVHN-LT and Small ImageNet-127 demonstrate the effectiveness of ABAE. •Propose an Auxiliary Balanced AutoEncoder branch(ABAE) with a class-aware reconstruction loss for CISSL.•Utilize the adaptive feature augmentation for different classes to alleviate the confirmation bias.•Adopt a graph-based label propagation scheme to smooth the output of the ABAE.•Demonstrate the effectiveness of the ABAE through experiments on various imbalanced datasets.
ISSN:0167-8655
1872-7344
DOI:10.1016/j.patrec.2024.04.004