Forgetting Similar Samples: Can Machine Unlearning Do it Better?

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Heng, Zhu, Tianqing, Ye, Dayong, Zhang, Lefeng, Wang, Le, Zhou, Wanlei
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912817265770496
author Xu, Heng
Zhu, Tianqing
Ye, Dayong
Zhang, Lefeng
Wang, Le
Zhou, Wanlei
author_facet Xu, Heng
Zhu, Tianqing
Ye, Dayong
Zhang, Lefeng
Wang, Le
Zhou, Wanlei
contents Machine unlearning, a process enabling pre-trained models to remove the influence of specific training samples, has attracted significant attention in recent years. Although extensive research has focused on developing efficient machine unlearning strategies, we argue that these methods mainly aim at removing samples rather than removing samples' influence on the model, thus overlooking the fundamental definition of machine unlearning. In this paper, we first conduct a comprehensive study to evaluate the effectiveness of existing unlearning schemes when the training dataset includes many samples similar to those targeted for unlearning. Specifically, we evaluate: Do existing unlearning methods truly adhere to the original definition of machine unlearning and effectively eliminate all influence of target samples when similar samples are present in the training dataset? Our extensive experiments, conducted on four carefully constructed datasets with thorough analysis, reveal a notable gap between the expected and actual performance of most existing unlearning methods for image and language models, even for the retraining-from-scratch baseline. Additionally, we also explore potential solutions to enhance current unlearning approaches.
format Preprint
id arxiv_https___arxiv_org_abs_2601_06938
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Forgetting Similar Samples: Can Machine Unlearning Do it Better?
Xu, Heng
Zhu, Tianqing
Ye, Dayong
Zhang, Lefeng
Wang, Le
Zhou, Wanlei
Machine Learning
Cryptography and Security
Machine unlearning, a process enabling pre-trained models to remove the influence of specific training samples, has attracted significant attention in recent years. Although extensive research has focused on developing efficient machine unlearning strategies, we argue that these methods mainly aim at removing samples rather than removing samples' influence on the model, thus overlooking the fundamental definition of machine unlearning. In this paper, we first conduct a comprehensive study to evaluate the effectiveness of existing unlearning schemes when the training dataset includes many samples similar to those targeted for unlearning. Specifically, we evaluate: Do existing unlearning methods truly adhere to the original definition of machine unlearning and effectively eliminate all influence of target samples when similar samples are present in the training dataset? Our extensive experiments, conducted on four carefully constructed datasets with thorough analysis, reveal a notable gap between the expected and actual performance of most existing unlearning methods for image and language models, even for the retraining-from-scratch baseline. Additionally, we also explore potential solutions to enhance current unlearning approaches.
title Forgetting Similar Samples: Can Machine Unlearning Do it Better?
topic Machine Learning
Cryptography and Security
url https://arxiv.org/abs/2601.06938