Auditing of Unlearning Algorithms

• URL: http://arxiv.org/abs/2607.05898v1
• Date: Tue, 07 Jul 2026 06:51:55 GMT
• Title: Auditing of Unlearning Algorithms
• Abstract summary: We propose a practical auditor that computes data-dependent lower bounds on the unlearning parameter $varepsilon$ using membership inference attacks.<n>We find a sharp separation: algorithms with rigorous guarantees, such as model clipping and rewind-to-delete, achieve very small $varepsilon$ bounds that do not falsify their unlearning guarantees.<n>Our auditor provides a practical tool for empirically falsifying unlearning claims through a hypothesis-testing framework, and we validate it on CIFAR-100 and Shakespeare text.
• Score: 28.787875302334058
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Evaluating whether unlearning algorithms truly remove training data influence remains an open challenge. We propose a practical auditor that computes data-dependent lower bounds on the unlearning parameter $\varepsilon$ using membership inference attacks. Evaluating multiple unlearning algorithms, we find a sharp separation: algorithms with rigorous guarantees, such as model clipping and rewind-to-delete, achieve very small $\varepsilon$ bounds that do not falsify their unlearning guarantees, whereas empirical methods such as Hessian-based unlearning, interleaved ascent-descent, ascent on the forget set, and fine-tuning on the retain set exhibit large bounds, indicating poor unlearning. Our auditor provides a practical tool for empirically falsifying unlearning claims through a hypothesis-testing framework, and we validate it on CIFAR-100 and Shakespeare text.

Related papers

• Auditing Machine Unlearning: A Systematic Research on Whether Models Truly Forget [25.03537230361785]
  We propose the first practical and general-purpose auditing framework for machine unlearning.<n>Our framework addresses the key limitations of existing methods by eliminating the need for retraining-from-scratch baselines.<n>We show that our framework is robust against fake unlearning attempts and generalizes well to large language models.
  arXiv  Detail & Related papers  (2026-06-15T01:56:09Z)
• Unlearning That Lasts: Utility-Preserving, Robust, and Almost Irreversible Forgetting in LLMs [31.768387661474904]
  Unlearning in large language models (LLMs) involves precisely removing specific information from a pre-trained model.<n>This is crucial to ensure safety of LLMs by deleting private data or harmful knowledge acquired during pre-training.<n>We introduce JensUn, where we leverage the Jensen-Shannon Divergence as the training objective for both forget and retain sets.<n>In extensive experiments, JensUn achieves better forget-utility trade-off than competing methods, and even demonstrates strong resilience to benign relearning.
  arXiv  Detail & Related papers  (2025-09-02T20:38:53Z)
• Rewind-to-Delete: Certified Machine Unlearning for Nonconvex Functions [22.183624306817563]
  Machine unlearning algorithms aim to efficiently data from a model without it from scratch.<n> certified machine unlearning is a strong theoretical guarantee based on differential generalization.
  arXiv  Detail & Related papers  (2024-09-15T15:58:08Z)
• Provable Optimization for Adversarial Fair Self-supervised Contrastive Learning [49.417414031031264]
  This paper studies learning fair encoders in a self-supervised learning setting. All data are unlabeled and only a small portion of them are annotated with sensitive attributes.
  arXiv  Detail & Related papers  (2024-06-09T08:11:12Z)
• Noisy Correspondence Learning with Self-Reinforcing Errors Mitigation [63.180725016463974]
  Cross-modal retrieval relies on well-matched large-scale datasets that are laborious in practice. We introduce a novel noisy correspondence learning framework, namely textbfSelf-textbfReinforcing textbfErrors textbfMitigation (SREM)
  arXiv  Detail & Related papers  (2023-12-27T09:03:43Z)
• Correcting Underrepresentation and Intersectional Bias for Classification [49.1574468325115]
  We consider the problem of learning from data corrupted by underrepresentation bias. We show that with a small amount of unbiased data, we can efficiently estimate the group-wise drop-out rates. We show that our algorithm permits efficient learning for model classes of finite VC dimension.
  arXiv  Detail & Related papers  (2023-06-19T18:25:44Z)
• Adaptive Negative Evidential Deep Learning for Open-set Semi-supervised Learning [69.81438976273866]
  Open-set semi-supervised learning (Open-set SSL) considers a more practical scenario, where unlabeled data and test data contain new categories (outliers) not observed in labeled data (inliers) We introduce evidential deep learning (EDL) as an outlier detector to quantify different types of uncertainty, and design different uncertainty metrics for self-training and inference. We propose a novel adaptive negative optimization strategy, making EDL more tailored to the unlabeled dataset containing both inliers and outliers.
  arXiv  Detail & Related papers  (2023-03-21T09:07:15Z)
• Forget Unlearning: Towards True Data-Deletion in Machine Learning [18.656957502454592]
  We show that unlearning is not equivalent to data deletion and does not guarantee the "right to be forgotten" We propose an accurate, computationally efficient, and secure data-deletion machine learning algorithm in the online setting.
  arXiv  Detail & Related papers  (2022-10-17T10:06:11Z)
• Algorithms that Approximate Data Removal: New Results and Limitations [2.6905021039717987]
  We study the problem of deleting user data from machine learning models trained using empirical risk minimization. We develop an online unlearning algorithm that is both computationally and memory efficient.
  arXiv  Detail & Related papers  (2022-09-25T17:20:33Z)
• Low-Regret Active learning [64.36270166907788]
  We develop an online learning algorithm for identifying unlabeled data points that are most informative for training. At the core of our work is an efficient algorithm for sleeping experts that is tailored to achieve low regret on predictable (easy) instances.
  arXiv  Detail & Related papers  (2021-04-06T22:53:45Z)
• Provably Robust Metric Learning [98.50580215125142]
  We show that existing metric learning algorithms can result in metrics that are less robust than the Euclidean distance. We propose a novel metric learning algorithm to find a Mahalanobis distance that is robust against adversarial perturbations. Experimental results show that the proposed metric learning algorithm improves both certified robust errors and empirical robust errors.
  arXiv  Detail & Related papers  (2020-06-12T09:17:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.