MPU: Towards Secure and Privacy-Preserving Knowledge Unlearning for Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918501441077248 |
|---|---|
| author | Wang, Tiantong Yan, Xinyu Wu, Tiantong Hao, Yurong Xie, Pengjun Lim, Wei Yang Bryan |
| author_facet | Wang, Tiantong Yan, Xinyu Wu, Tiantong Hao, Yurong Xie, Pengjun Lim, Wei Yang Bryan |
| contents | Machine unlearning for large language models often faces a privacy dilemma in which strict constraints prohibit sharing either the server's parameters or the client's forget set. To address this dual non-disclosure constraint, we propose MPU, an algorithm-agnostic privacy-preserving Multiple Perturbed Copies Unlearning framework that primarily introduces two server-side modules: Pre-Process for randomized copy generation and Post-Process for update aggregation. In Pre-Process, the server distributes multiple perturbed and reparameterized model instances, allowing the client to execute unlearning locally on its private forget set without accessing the server's exact original parameters. After local unlearning, the server performs Post-Process by inverting the reparameterization and aggregating updates with a harmonic denoising procedure to alleviate the impact of perturbation. Experiments with seven unlearning algorithms show that MPU achieves comparable unlearning performance to noise-free baselines, with most algorithms' average degradation well below 1% up to 10% noise, and can even outperform the noise-free baseline for some algorithms under 1% noise. Code is available at https://github.com/Tristan0318/MPU. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_23798 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MPU: Towards Secure and Privacy-Preserving Knowledge Unlearning for Large Language Models Wang, Tiantong Yan, Xinyu Wu, Tiantong Hao, Yurong Xie, Pengjun Lim, Wei Yang Bryan Machine Learning Artificial Intelligence Cryptography and Security Distributed, Parallel, and Cluster Computing Machine unlearning for large language models often faces a privacy dilemma in which strict constraints prohibit sharing either the server's parameters or the client's forget set. To address this dual non-disclosure constraint, we propose MPU, an algorithm-agnostic privacy-preserving Multiple Perturbed Copies Unlearning framework that primarily introduces two server-side modules: Pre-Process for randomized copy generation and Post-Process for update aggregation. In Pre-Process, the server distributes multiple perturbed and reparameterized model instances, allowing the client to execute unlearning locally on its private forget set without accessing the server's exact original parameters. After local unlearning, the server performs Post-Process by inverting the reparameterization and aggregating updates with a harmonic denoising procedure to alleviate the impact of perturbation. Experiments with seven unlearning algorithms show that MPU achieves comparable unlearning performance to noise-free baselines, with most algorithms' average degradation well below 1% up to 10% noise, and can even outperform the noise-free baseline for some algorithms under 1% noise. Code is available at https://github.com/Tristan0318/MPU. |
| title | MPU: Towards Secure and Privacy-Preserving Knowledge Unlearning for Large Language Models |
| topic | Machine Learning Artificial Intelligence Cryptography and Security Distributed, Parallel, and Cluster Computing |
| url | https://arxiv.org/abs/2602.23798 |