Built with and by Teycir Ben Soltane•
How to Use•FAQ•GitHub•arXiv.org•
Share:
ArXivCSExplorer
☆☆Bookmarks🏆RSSHow to UseFAQ
Back to Paper
cs.LGcs.CR

Local ID: 2605.11170v2

AI Summary: gemma4:e4b

Unlearning with Asymmetric Sources: Improved Unlearning-Utility Trade-off with Public Data

By Ahmed Mehdi Inane, Vincent Quirion, Gintare Karolina Dziugaite, Ioannis Mitliagkas

Revision History Timeline

v15/11/2026
5/11/2026

No submitter comment provided.

v26/2/2026
6/2/2026

No submitter comment provided.

★ Version indexed in Explorer

Comparing v1 vs v2

Green = Added • Red = Removed

Title Comparison

Unlearning with Asymmetric Sources: Improved Unlearning-Utility Trade-off with Public Data

Authors Comparison

Removed:Gintare Karolina Dzugaite
Added:Gintare Karolina Dziugaite
Unchanged:Ahmed Mehdi Inane, Vincent Quirion, Ioannis Mitliagkas

v1 Comment

No comment for this version.

v2 Comment

No comment for this version.

Abstract Word Diff

Noise-based certified machine unlearning currently faces a hard ceiling: the noise magnitude required to certify unlearning typically destroys model utility, particularly for large-scale deletion requests. While leveraging public data is a standard technique in differential privacy to relax this tension, its role in unlearning remains unexplored. We address this gap by introducing Asymmetric Langevin Unlearning (ALU), a framework that uses public data to mitigate privacy costs. We prove that public data injection suppresses the unlearning cost by a factor of $O(1/n_{\mathrm{pub}}^2)$, guaranteeing a strict computational advantage over retraining. This establishes a new control mechanism: practitioners can mitigate the need for high noise-and the associated utility loss-by increasing the volume of public data. Crucially, we analyze the realistic setting of distribution mismatch, explicitly characterizing how shifts between public and private sources impact utility. We show that ALU enables mass unlearning of constant dataset fractions -- a regime where standard symmetric methods become impractical -- while maintaining high utility. Empirical evaluations using variational Rényi divergence and membership inference attacks confirm that ALU effectively thwarts privacy attacks while preserving utility under reasonable distribution shifts.
View Full Version History on arXiv