Learning with Noisy Labels by Adaptive Gradient-Based Outlier Removal

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sedova, Anastasiia, Zellinger, Lena, Roth, Benjamin
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929197041057792
author Sedova, Anastasiia
Zellinger, Lena
Roth, Benjamin
author_facet Sedova, Anastasiia
Zellinger, Lena
Roth, Benjamin
contents An accurate and substantial dataset is essential for training a reliable and well-performing model. However, even manually annotated datasets contain label errors, not to mention automatically labeled ones. Previous methods for label denoising have primarily focused on detecting outliers and their permanent removal - a process that is likely to over- or underfilter the dataset. In this work, we propose AGRA: a new method for learning with noisy labels by using Adaptive GRAdient-based outlier removal. Instead of cleaning the dataset prior to model training, the dataset is dynamically adjusted during the training process. By comparing the aggregated gradient of a batch of samples and an individual example gradient, our method dynamically decides whether a corresponding example is helpful for the model at this point or is counter-productive and should be left out for the current update. Extensive evaluation on several datasets demonstrates AGRA's effectiveness, while a comprehensive results analysis supports our initial hypothesis: permanent hard outlier removal is not always what model benefits the most from.
format Preprint
id arxiv_https___arxiv_org_abs_2306_04502
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Learning with Noisy Labels by Adaptive Gradient-Based Outlier Removal
Sedova, Anastasiia
Zellinger, Lena
Roth, Benjamin
Machine Learning
Artificial Intelligence
An accurate and substantial dataset is essential for training a reliable and well-performing model. However, even manually annotated datasets contain label errors, not to mention automatically labeled ones. Previous methods for label denoising have primarily focused on detecting outliers and their permanent removal - a process that is likely to over- or underfilter the dataset. In this work, we propose AGRA: a new method for learning with noisy labels by using Adaptive GRAdient-based outlier removal. Instead of cleaning the dataset prior to model training, the dataset is dynamically adjusted during the training process. By comparing the aggregated gradient of a batch of samples and an individual example gradient, our method dynamically decides whether a corresponding example is helpful for the model at this point or is counter-productive and should be left out for the current update. Extensive evaluation on several datasets demonstrates AGRA's effectiveness, while a comprehensive results analysis supports our initial hypothesis: permanent hard outlier removal is not always what model benefits the most from.
title Learning with Noisy Labels by Adaptive Gradient-Based Outlier Removal
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2306.04502