mlsanityData to 7 Oct 2026

BlindU: Blind Machine Unlearning without Revealing Erasing Data

arXiv 2601.07214 ↗First posted 12 Jan 2026cs.CR cs.AI cs.CVcited by 0

On the map: LLM safety and security → Federated unlearning and poisoning

Abstract (arXiv)

Machine unlearning enables data holders to remove the contribution of their specified samples from trained models to protect their privacy. However, it is paradoxical that most unlearning methods require the unlearning requesters to firstly upload their data to the server as a prerequisite for unlearning. These methods are infeasible in many privacy-preserving scenarios where servers are prohibited from accessing users' data, such as federated learning (FL). In this paper, we explore how to implement unlearning under the condition of not uncovering the erasing data to the server. We propose Blind Unlearning (BlindU), which carries out unlearning using compressed representations instead of original inputs. BlindU only involves the server and the unlearning user: the user locally generates privacy-preserving representations, and the server performs unlearning solely on these representations and their labels. For the FL model training, we employ the information bottleneck (IB) mechanism. The encoder of the IB-based FL model learns representations that distort maximum task-irrelevant information from inputs, allowing FL users to generate compressed representations locally. For effective unlearning using compressed representation, BlindU integrates two dedicated unlearning modules tailored explicitly for IB-based models and uses a multiple gradient descent algorithm to balance forgetting and utility retaining. While IB compression already provides protection for task-irrelevant information of inputs, to further enhance the privacy protection, we introduce a noise-free differential privacy (DP) masking method to deal with the raw erasing data before compressing. Theoretical analysis and extensive experimental results illustrate the superiority of BlindU in privacy protection and unlearning effectiveness compared with the best existing privacy-preserving unlearning benchmarks.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Federated unlearning and poisoning, one of 12 directions in LLM safety and security. Builds on 37 of its 65 references that are in the map. First posted on arXiv 12 Jan 2026.

Builds on

37 of 65 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022The Right to be Forgotten in Federated Learning: An Efficient Realization with Rapid Retraining14 Mar

Other references · 36

  1. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,339
  2. 2019Machine Unlearningcited by 1,020
  3. 2019Deep Leakage from Gradientscited by 1,004
  4. 2016Deep Variational Information Bottleneckcited by 915
  5. 2018Multi-Task Learning as Multi-Objective Optimizationcited by 809
  6. 2020Inverting Gradients -- How easy is it to break privacy in federated learning?cited by 707
  7. 2018MINE: Mutual Information Neural Estimationcited by 680
  8. 2019Certified Data Removal from Machine Learning Modelscited by 512
  9. 2018Exploiting Unintended Feature Leakage in Collaborative Learningcited by 509
  10. 2016f-GAN: Training Generative Neural Samplers using Variational Divergence Minimizationcited by 504
  11. 2024Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearningcited by 484
  12. 2021Remember What You Want to Forget: Algorithms for Machine Unlearningcited by 360
24 more
  1. 2021See through Gradients: Image Batch Recovery via GradInversioncited by 320
  2. 2023Towards Unbounded Machine Unlearningcited by 261
  3. 2020When Machine Unlearning Jeopardizes Privacycited by 181
  4. 2016Information Dropout: Learning Optimal Representations Through Noisy Computationcited by 131
  5. 2021When the Curious Abandon Honesty: Federated Learning Is Not Privatecited by 126
  6. 2019Updates-Leak: Data Set Inference and Reconstruction Attacks in Online Learningcited by 117
  7. 2020Variational Bayesian Unlearningcited by 114
  8. 2021Decentralized Federated Averagingcited by 114
  9. 2022Personalized Federated Learning via Variational Bayesian Inferencecited by 68
  10. 2025Rethinking LLM Unlearning Objectives: A Gradient Perspective and Go Beyondcited by 59
  11. 2019Local Differential Privacy for Deep Learningcited by 46
  12. 2024Learn What You Want to Unlearn: Unlearning Inversion Attacks against Machine Unlearningcited by 43
  13. 2021A Bayesian Federated Learning Framework with Online Laplace Approximationcited by 41
  14. 2020Federated Model Distillation with Noise-Free Differential Privacycited by 40
  15. 2019Explaining a black-box using Deep Variational Information Bottleneck Approachcited by 35
  16. 2021Recognizing Predictive Substructures with Subgraph Information Bottleneckcited by 28
  17. 2024Towards Effective Evaluations and Comparisons for LLM Unlearning Methodscited by 27
  18. 2022Membership Inference via Backdooringcited by 24
  19. 2022Knowledge Removal in Sampling-based Bayesian Inferencecited by 20
  20. 2025LLM Unlearning with LLM Beliefscited by 17
  21. 2023ERASER: Machine Unlearning in MLaaS via an Inference Serving-Aware Approachcited by 15
  22. 2019CryptoNN: Training Neural Networks over Encrypted Datacited by 14
  23. 2024SPEAR:Exact Gradient Inversion of Batches in Federated Learningcited by 14
  24. 2024Decoupling the Class Label and the Target Concept in Machine Unlearningcited by 12

Built on by

Not cited yet.