Shadow Unlearning: A Neuro-Semantic Approach to Fidelity-Preserving Faceless Forgetting in LLMs

TL;DR

This paper introduces Shadow Unlearning, a privacy-preserving method for selectively removing training data influence from LLMs without exposing PII, achieving high efficiency and maintaining model utility.

Contribution

It proposes a novel neuro-semantic framework called NSPU for privacy-preserving unlearning that is more efficient and effective than existing methods.

Findings

NSPU outperforms standard unlearning methods in effectiveness.

NSPU preserves model utility while removing sensitive data.

The approach is at least 10 times more computationally efficient.

Abstract

Machine unlearning aims to selectively remove the influence of specific training samples to satisfy privacy regulations such as the GDPR's 'Right to be Forgotten'. However, many existing methods require access to the data being removed, exposing it to membership inference attacks and potential misuse of Personally Identifiable Information (PII). We address this critical challenge by proposing Shadow Unlearning, a novel paradigm of approximate unlearning, that performs machine unlearning on anonymized forget data without exposing PII. We further propose a novel privacy-preserving framework, Neuro-Semantic Projector Unlearning (NSPU) to achieve Shadow unlearning. To evaluate our method, we compile Multi-domain Fictitious Unlearning (MuFU) forget set across five diverse domains and introduce an evaluation stack to quantify the trade-off between knowledge retention and unlearning effectiveness. Experimental results on various LLMs show that NSPU achieves superior unlearning performance, preserves model utility, and enhances user privacy. Additionally, the proposed approach is at least 10 times more computationally efficient than standard unlearning approaches. Our findings foster a new direction for privacy-aware machine unlearning that balances data protection and model fidelity.