Unlocking the Power of Open Set : A New Perspective for Open-Set Noisy Label Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wan, Wenhai, Wang, Xinrui, Xie, Ming-Kun, Li, Shao-Yuan, Huang, Sheng-Jun, Chen, Songcan
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911782321258496
author Wan, Wenhai
Wang, Xinrui
Xie, Ming-Kun
Li, Shao-Yuan
Huang, Sheng-Jun
Chen, Songcan
author_facet Wan, Wenhai
Wang, Xinrui
Xie, Ming-Kun
Li, Shao-Yuan
Huang, Sheng-Jun
Chen, Songcan
contents Learning from noisy data has attracted much attention, where most methods focus on closed-set label noise. However, a more common scenario in the real world is the presence of both open-set and closed-set noise. Existing methods typically identify and handle these two types of label noise separately by designing a specific strategy for each type. However, in many real-world scenarios, it would be challenging to identify open-set examples, especially when the dataset has been severely corrupted. Unlike the previous works, we explore how models behave when faced with open-set examples, and find that \emph{a part of open-set examples gradually get integrated into certain known classes}, which is beneficial for the separation among known classes. Motivated by the phenomenon, we propose a novel two-step contrastive learning method CECL (Class Expansion Contrastive Learning) which aims to deal with both types of label noise by exploiting the useful information of open-set examples. Specifically, we incorporate some open-set examples into closed-set classes to enhance performance while treating others as delimiters to improve representative ability. Extensive experiments on synthetic and real-world datasets with diverse label noise demonstrate the effectiveness of CECL.
format Preprint
id arxiv_https___arxiv_org_abs_2305_04203
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Unlocking the Power of Open Set : A New Perspective for Open-Set Noisy Label Learning
Wan, Wenhai
Wang, Xinrui
Xie, Ming-Kun
Li, Shao-Yuan
Huang, Sheng-Jun
Chen, Songcan
Machine Learning
Computer Vision and Pattern Recognition
Learning from noisy data has attracted much attention, where most methods focus on closed-set label noise. However, a more common scenario in the real world is the presence of both open-set and closed-set noise. Existing methods typically identify and handle these two types of label noise separately by designing a specific strategy for each type. However, in many real-world scenarios, it would be challenging to identify open-set examples, especially when the dataset has been severely corrupted. Unlike the previous works, we explore how models behave when faced with open-set examples, and find that \emph{a part of open-set examples gradually get integrated into certain known classes}, which is beneficial for the separation among known classes. Motivated by the phenomenon, we propose a novel two-step contrastive learning method CECL (Class Expansion Contrastive Learning) which aims to deal with both types of label noise by exploiting the useful information of open-set examples. Specifically, we incorporate some open-set examples into closed-set classes to enhance performance while treating others as delimiters to improve representative ability. Extensive experiments on synthetic and real-world datasets with diverse label noise demonstrate the effectiveness of CECL.
title Unlocking the Power of Open Set : A New Perspective for Open-Set Noisy Label Learning
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2305.04203