Machine unlearning algorithms, designed for selective removal of training data from models, have emerged as a promising approach to growing privacy concerns. In this work, we expose a critical yet underexplored vulnerability in the deployment of unlearning systems: the assumption that the data requested for removal is always part of the original training set. We present a threat model where an attacker can degrade model accuracy by submitting adversarial unlearning requests for data not present in the training set. We propose white-box and black-box attack algorithms and evaluate them through a case study on image classification tasks using the CIFAR-10 and ImageNet datasets, targeting a family of widely used unlearning methods. Our results show extremely poor test accuracy following the attack: 3.6% on CIFAR-10 and 0.4% on ImageNet for white-box attacks, and 8.5% on CIFAR-10 and 1.3% on ImageNet for black-box attacks. Additionally, we evaluate various verification mechanisms to detect the legitimacy of unlearning requests and reveal the challenges in verification, as most of the mechanisms fail to detect stealthy attacks without severely impairing their ability to process valid requests. These findings underscore the urgent need for research on more robust request verification methods and unlearning protocols, should the deployment of machine unlearning systems become more prevalent in the future.
Zero-shot Class Unlearning via Layer-wise Relevance Analysis and Neuronal Path Perturbation
School of Information Engineering, Zhongnan University of Economics and Law(信息工程学院,中南财经政法大学);China University of Geosciences, Wuhan(中国地质大学(武汉));City University of Macau(澳门城市大学)
Machine Unlearning with Minimal Gradient Dependence for High Unlearning Ratios
School of Information, Renmin University of China(中国人民大学信息学院);School of Computer Science and Big Data, Minjiang University(闽江学院计算机与大数据学院);RMIT University(皇家墨尔本理工大学)
Towards Lifecycle Unlearning Commitment Management: Measuring Sample-level Unlearning Completeness
Provable Responsible AI and Data Analytics Lab(可证负责任人工智能与数据 analytics 实验室);King Abdullah University of Science and Technology(卡塔尔国王 Abdullah 科学与技术大学);National University of Singapore(新加坡国家大学);Johns Hopkins University(约翰霍普金斯大学)
本文提出 IAM 框架,通过插值模型在样本上的泛化—拟合差距度量样本级遗忘完整性,可低成本扩展至 LLM,并揭示近似遗忘中过度与不足遗忘风险。
TruVRF: Towards Triple-Granularity Verification on Machine Unlearning
School of Computer Science and Technology, Zhejiang University(浙江大学计算机科学与技术学院);School of Computer Science and Engineering, Nanjing University of Science and Technology(南京理工大学计算机科学与工程学院)
Towards Irreversible Machine Unlearning for Diffusion Models
Department of Electrical and Computer Engineering, College of Design and Engineering, National University of Singapore(新加坡国立大学电子与计算机工程系,设计与工程学院);Betterdata, Singapore(新加坡Betterdata公司);Department of Computer Science, University of Sheffield(谢菲尔德大学计算机科学系)
Textual Unlearning Gives a False Sense of Unlearning
The State Key Laboratory of Blockchain and Data Security, Zhejiang University, P. R. China(区块链与数据安全国家重点实验室,浙江大学,中华人民共和国);School of Cyber Science and Technology, Zhejiang University, P. R. China(网络安全与技术学院,浙江大学,中华人民共和国);ETH Zurich, Switzerland(苏黎世联邦理工学院,瑞士)
BlindU: Blind Machine Unlearning without Revealing Erasing Data
School of Computer Science, University of Technology Sydney(计算机科学学院,技术悉尼大学);School of Cyber Science and Engineering, Southeast University(网络科学与工程学院,东南大学)
Towards Robust Knowledge Unlearning: An Adversarial Framework for Assessing and Improving Unlearning Robustness in Large Language Models
Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所);School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院)
Poisoning Attacks and Defenses to Federated Unlearning
ShanghaiTech University(上海科技大学);Xidian University(西安电子科技大学);North Carolina State University(北卡罗来纳州立大学);University of North Texas(北德克萨斯大学);University of Louisville(路易斯维尔大学)