Learning with Instance-Dependent Noisy Labels by Anchor Hallucination and Hard Sample Label Correction

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Huang, Po-Hsuan, Lin, Chia-Ching, Hsu, Chih-Fan, Chang, Ming-Ching, Chen, Wei-Chao
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914864610410496
author Huang, Po-Hsuan
Lin, Chia-Ching
Hsu, Chih-Fan
Chang, Ming-Ching
Chen, Wei-Chao
author_facet Huang, Po-Hsuan
Lin, Chia-Ching
Hsu, Chih-Fan
Chang, Ming-Ching
Chen, Wei-Chao
contents Learning from noisy-labeled data is crucial for real-world applications. Traditional Noisy-Label Learning (NLL) methods categorize training data into clean and noisy sets based on the loss distribution of training samples. However, they often neglect that clean samples, especially those with intricate visual patterns, may also yield substantial losses. This oversight is particularly significant in datasets with Instance-Dependent Noise (IDN), where mislabeling probabilities correlate with visual appearance. Our approach explicitly distinguishes between clean vs.noisy and easy vs. hard samples. We identify training samples with small losses, assuming they have simple patterns and correct labels. Utilizing these easy samples, we hallucinate multiple anchors to select hard samples for label correction. Corrected hard samples, along with the easy samples, are used as labeled data in subsequent semi-supervised training. Experiments on synthetic and real-world IDN datasets demonstrate the superior performance of our method over other state-of-the-art NLL methods.
format Preprint
id arxiv_https___arxiv_org_abs_2407_07331
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Learning with Instance-Dependent Noisy Labels by Anchor Hallucination and Hard Sample Label Correction
Huang, Po-Hsuan
Lin, Chia-Ching
Hsu, Chih-Fan
Chang, Ming-Ching
Chen, Wei-Chao
Computer Vision and Pattern Recognition
Artificial Intelligence
Learning from noisy-labeled data is crucial for real-world applications. Traditional Noisy-Label Learning (NLL) methods categorize training data into clean and noisy sets based on the loss distribution of training samples. However, they often neglect that clean samples, especially those with intricate visual patterns, may also yield substantial losses. This oversight is particularly significant in datasets with Instance-Dependent Noise (IDN), where mislabeling probabilities correlate with visual appearance. Our approach explicitly distinguishes between clean vs.noisy and easy vs. hard samples. We identify training samples with small losses, assuming they have simple patterns and correct labels. Utilizing these easy samples, we hallucinate multiple anchors to select hard samples for label correction. Corrected hard samples, along with the easy samples, are used as labeled data in subsequent semi-supervised training. Experiments on synthetic and real-world IDN datasets demonstrate the superior performance of our method over other state-of-the-art NLL methods.
title Learning with Instance-Dependent Noisy Labels by Anchor Hallucination and Hard Sample Label Correction
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2407.07331