Knowledge Swapping via Learning and Unlearning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xing, Mingyu, Cheng, Lechao, Tang, Shengeng, Wang, Yaxiong, Zhong, Zhun, Wang, Meng
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916616860598272
author Xing, Mingyu
Cheng, Lechao
Tang, Shengeng
Wang, Yaxiong
Zhong, Zhun
Wang, Meng
author_facet Xing, Mingyu
Cheng, Lechao
Tang, Shengeng
Wang, Yaxiong
Zhong, Zhun
Wang, Meng
contents We introduce \textbf{Knowledge Swapping}, a novel task designed to selectively regulate knowledge of a pretrained model by enabling the forgetting of user\-specified information, retaining essential knowledge, and acquiring new knowledge simultaneously. By delving into the analysis of knock-on feature hierarchy, we find that incremental learning typically progresses from low\-level representations to higher\-level semantics, whereas forgetting tends to occur in the opposite direction\-starting from high-level semantics and moving down to low-level features. Building upon this, we propose to benchmark the knowledge swapping task with the strategy of \textit{Learning Before Forgetting}. Comprehensive experiments on various tasks like image classification, object detection, and semantic segmentation validate the effectiveness of the proposed strategy. The source code is available at \href{https://github.com/xingmingyu123456/KnowledgeSwapping}{https://github.com/xingmingyu123456/KnowledgeSwapping}.
format Preprint
id arxiv_https___arxiv_org_abs_2502_08075
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Knowledge Swapping via Learning and Unlearning
Xing, Mingyu
Cheng, Lechao
Tang, Shengeng
Wang, Yaxiong
Zhong, Zhun
Wang, Meng
Computer Vision and Pattern Recognition
We introduce \textbf{Knowledge Swapping}, a novel task designed to selectively regulate knowledge of a pretrained model by enabling the forgetting of user\-specified information, retaining essential knowledge, and acquiring new knowledge simultaneously. By delving into the analysis of knock-on feature hierarchy, we find that incremental learning typically progresses from low\-level representations to higher\-level semantics, whereas forgetting tends to occur in the opposite direction\-starting from high-level semantics and moving down to low-level features. Building upon this, we propose to benchmark the knowledge swapping task with the strategy of \textit{Learning Before Forgetting}. Comprehensive experiments on various tasks like image classification, object detection, and semantic segmentation validate the effectiveness of the proposed strategy. The source code is available at \href{https://github.com/xingmingyu123456/KnowledgeSwapping}{https://github.com/xingmingyu123456/KnowledgeSwapping}.
title Knowledge Swapping via Learning and Unlearning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.08075