Toward Efficient Data-Free Unlearning
Chenhao Zhang, Shaofei Shen, Weitong Chen, Miao Xu
Abstract
Machine unlearning without access to real data distribution is challenging. The existing method based on data-free distillation achieved unlearning by filtering out synthetic samples containing forgetting information but struggled to distill the retaining-related knowledge efficiently. In this work, we analyze that such a problem is due to over-filtering, which reduces the synthesized retaining-related information. We propose a novel method, Inhibited Synthetic PostFilter (ISPF), to tackle this challenge from two perspectives: First, the Inhibited Synthetic, by reducing the synthesized forgetting information; Second, the PostFilter, by fully utilizing the retaining-related information in synthesized samples. Experimental results demonstrate that the proposed ISPF effectively tackles the challenge and outperforms existing methods.
BibTeX
@article{Zhang_Shen_Chen_Xu_2025, title={Toward Efficient Data-Free Unlearning}, volume={39}, url={https://ojs.aaai.org/index.php/AAAI/article/view/34393}, DOI={10.1609/aaai.v39i21.34393}, abstractNote={Machine unlearning without access to real data distribution is challenging. The existing method based on data-free distillation achieved unlearning by filtering out synthetic samples containing forgetting information but struggled to distill the retaining-related knowledge efficiently. In this work, we analyze that such a problem is due to over-filtering, which reduces the synthesized retaining-related information. We propose a novel method, Inhibited Synthetic PostFilter (ISPF), to tackle this challenge from two perspectives: First, the Inhibited Synthetic, by reducing the synthesized forgetting information; Second, the PostFilter, by fully utilizing the retaining-related information in synthesized samples. Experimental results demonstrate that the proposed ISPF effectively tackles the challenge and outperforms existing methods.}, number={21}, journal={Proceedings of the AAAI Conference on Artificial Intelligence}, author={Zhang, Chenhao and Shen, Shaofei and Chen, Weitong and Xu, Miao}, year={2025}, month={Apr.}, pages={22372-22379} }