Publication:
Enhancing Continual Noisy Label Learning with�Uncertainty-Based Sample Selection and�Feature Enhancement

Simple item page
dc.citedby1
dc.contributor.authorGuo G.en_US
dc.contributor.authorWei Z.en_US
dc.contributor.authorCheng J.en_US
dc.contributor.authorid58805753200en_US
dc.contributor.authorid58805777500en_US
dc.contributor.authorid22833734200en_US
dc.date.accessioned2025-03-03T07:48:18Z
dc.date.available2025-03-03T07:48:18Z
dc.date.issued2024
dc.description.abstractThe task of continual learning is to design algorithms that can address the problem of catastrophic forgetting. However, in the real world, there are noisy labels due to inaccurate human annotations and other factors, which seem to exacerbate catastrophic forgetting. To tackle both catastrophic forgetting and noise issues, we propose an innovative framework. Our framework leverages sample uncertainty to purify the data stream and selects representative samples for replay, effectively alleviating catastrophic forgetting. Additionally, we adopt a semi-supervised approach for fine-tuning to ensure the involvement of all available samples. Simultaneously, we incorporate contrastive learning and entropy minimization to mitigate noise memorization in the model. We validate the effectiveness of our proposed method through extensive experiments on two benchmark datasets, CIFAR-10 and CIFAR-100. For CIFAR-10, we achieve a performance gain of 2% under 20% noise conditions. ? The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2024.en_US
dc.description.natureFinalen_US
dc.identifier.doi10.1007/978-981-99-8543-2_40
dc.identifier.epage510
dc.identifier.scopus2-s2.0-85181983496
dc.identifier.spage498
dc.identifier.urihttps://www.scopus.com/inward/record.uri?eid=2-s2.0-85181983496&doi=10.1007%2f978-981-99-8543-2_40&partnerID=40&md5=f79b7a4392845c0d29b3d4190ac18737
dc.identifier.urihttps://irepository.uniten.edu.my/handle/123456789/37178
dc.identifier.volume14432 LNCS
dc.pagecount12
dc.publisherSpringer Science and Business Media Deutschland GmbHen_US
dc.sourceScopus
dc.sourcetitleLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
dc.subjectCatastrophic forgetting
dc.subjectContinual learning
dc.subjectFeature enhancement
dc.subjectNoisy data
dc.subjectNoisy labels
dc.subjectReal-world
dc.subjectReplay
dc.subjectSample features
dc.subjectSamples selection
dc.subjectUncertainty
dc.subjectEntropy
dc.titleEnhancing Continual Noisy Label Learning with�Uncertainty-Based Sample Selection and�Feature Enhancementen_US
dc.typeConference paperen_US
dspace.entity.typePublication
Files
Collections
SCOPUS