Machine Unlearning

Follow this topic Ask about this topic

Machine unlearning asks how to remove a person’s data, a class, or a learned concept from a model after training, often to support privacy, safety, or changing data rights. Approaches range from exact or certified deletion methods, such as data partitioning and noisy or influence-based updates, to approximate parameter edits, representation changes, and inference-time steering; some systems are designed from the outset to make deletion easier. Researchers evaluate methods against retrained models and test forgetting, retained utility, privacy attacks, recovery, and repeated deletion across language, vision, diffusion, graph, and multimodal systems. A central tension is that suppressed outputs do not necessarily mean information is erased, while stronger removal can damage related knowledge; scalable methods with reliable guarantees and low collateral harm remain a challenge.

Latest papers

351 papers56 papers in the last four weeks, up 70% on the four weeks before. 0.3% of all new papers.

Exploring and Bridging Knowledge Holes in Unlearned Multimodal Large Language Models
Aug 3, 2026Junxiang You, Junkai Chen, Yuhao He +3Multimodal Large Language Models MLLM Unlearning
University of Chinese Academy of Sciences · Institute of Automation, Chinese Academy of Sciences · 3Cloudspace Technology
QR-Erase: Efficient Subspace-Based Machine Unlearning with Layer Localization
Aug 2, 2026Tyler Lizzo, Larry HeckMachine Unlearning LLM Unlearning
Georgia Institute of Technology
Toward Fine-Grained Forgetting:Attribute Unlearning for Multimodal Large Language Models
Aug 2, 2026Junkai Lin, Junkai Chen, Siqi Hou +6VLM Unlearning Privacy Leakage in Language Models
Institute of Automation, Chinese Academy of Sciences · The Chinese University of Hong Kong · The Australian National University +3
Generated Images Are Easier to Forget: A Machine Unlearning Perspective for Synthetic Image Detection
Aug 1, 2026Jun Nie, Yonggang Zhang, Tongliang Liu +3Vision Foundation Models AI-Generated Image Detection
University of Science and Technology of China · Hong Kong Baptist University · The Hong Kong University of Science and Technology +1
Similarity-Aware Machine Unlearning
Jul 31, 2026Madhavan Citalamangalam Kumaran, Midhun Parakkal Unni, Vicky Kouni +1Machine Unlearning Machine Unlearning Evaluation
Department of Computer Science, University of Manchester, UK · Centre for Machine Intelligence, University of Sheffield, UK · School of Computer Science, University of Sheffield, Sheffield, UK +1
Beyond Binary Rewards: A Comparative Study of Reward Design for Reinforcement Unlearning
Jul 30, 2026Efstratios Zaradoukas, Davide Gabrielli, Bardh Prenkaj +1Reinforcement Learning with Verifiable Rewards Machine Unlearning
Technical University of Munich, Germany · Munich Center for Machine Learning, Germany · Sapienza University of Rome, Italy
Understanding Machine Unlearning Through the Lens of Mode Connectivity
Jul 27, 2026Jiali Cheng, Hadi AmiriLoss Landscape Mode Connectivity
University of Massachusetts Lowell
DECAF: De-Clustering for Adaptive Representational Unlearning
Jul 27, 2026Anjie Le, Can Peng, Hongcheng Guo +1Machine Unlearning
Institute of Biomedical Engineering, University of Oxford, UK · Fudan University.
Gradient Concentration, Not Weight Saliency, Explains Representation-Level Class Unlearning
Jul 23, 2026Billel Habbati, Alessio Merlo, Luca Verderame +1Gradient-Based Attribution Machine Unlearning
DIBRIS – Department of Informatics, Bioengineering, Robotics and Systems Engineering, University of Genova, Via Dodecaneso 35, 16146 Genova, Italy · Centre for Defense Higher Studies (CASD), Palazzo Salviati, Piazza della Rovere 83, 00165 Roma, Italy
Unlearning Under Imbalance: Benchmarking Fairness in Multimodal LLM Unlearning
Jul 23, 2026Lorenzo Orsingher, Thomas De Min, Massimiliano Mancini +2VLM Unlearning Algorithmic Fairness
University of Trento, Trento, Italy · Fondazione Bruno Kessler, Trento, Italy
TOUR: A Trajectory-Level Unlearning Benchmark for Offline Reinforcement Learning
Jul 23, 2026Chaofan Pan, Lingfei Ren, Xiangyu Jiang +6RL Benchmarks Membership Inference Attacks
1Southwestern University of Finance and Economics · 2Shanxi University
On Optimization Complexity of Second-Order Certified Unlearning
Jul 22, 2026Nikita Doikov, Anastasia KoloskovaSecond-Order Optimization Machine Unlearning
Cornell University, Ithaca, USA · University of Zurich, Zurich, Switzerland
Unlearning as Distribution Restoration: A Controlled Counterfactual Study, a Validated Selective Screen, and the Limits of Oracle-Free Certification
Jul 21, 2026Sen Yang, Yuen-Hei YeungMachine Unlearning Machine Unlearning Evaluation
Stern School of Business New York University · Courant Institute of Mathematical Sciences New York University
Stochastic Meta-Unlearning: Bridging Language Backbone and Multimodal Unlearning
Jul 21, 2026Zijie Liu, Jinhao Duan, Gaowen Liu +2VLM Unlearning Vision-Language Models
UNC at Chapel Hill · Cisco Research · Michigan State University
One Modality to Forget Them All: Enhancing Cross-Modal Unlearning in Vision-Language Models
Jul 17, 2026Sudharshan Balaji, Yili Ren, Guangjing Wang +2VLM Unlearning VLM Robustness
University of South Florida · USA · University of Massachusetts Lowell
Inference-Time Concept Suppression and Video-Centric Evaluation for Text-to-Video Models
Jul 15, 2026Wenxuan Chen, Wenjie FengVideo Diffusion Models T2V Generation
University of Science and Technology of China
Forgetful Attention: A Trainable Support-Vector Memory with Certified Selection and Exact Unlearning
Jul 13, 2026Vishwajith RameshMemory-Augmented Neural Networks Attention Mechanisms
Vy Labs, Inc.
Signal-Guided Optimization for Machine Unlearning
Jul 13, 2026Xujia Li, Dan Li, Jian Lou +1Privacy-Preserving ML Machine Unlearning
University of Science and Technology of China · Sun Yat-sen University
Forget Narrowly, Retain Broadly: Unlearning as an Asymmetric Generalization Problem
Jul 10, 2026Amit Peleg, Naman Deep Singh, Naama Pearl +2Factual Knowledge in Language Models Machine Unlearning
Tübingen AI Center, University of Tübingen
AutoAnchor: Stable Diffusion Unlearning Using Cross-Attention as a Manifold Surrogate
Jul 9, 2026Siyuan Wen, Jiahao Zeng, Ningning DingDiffusion Model Unlearning Machine Unlearning
Hong Kong University of Science and Technology (Guangzhou)
Multimodal Unlearning Across Vision, Language, Video, and Audio: Survey of Methods, Datasets, and Benchmarks
Jul 8, 2026Nobin Sarwar, Shubhashis Roy Dipta, Zheyuan Liu +1VLM Unlearning MLLM Unlearning
University of Maryland, Baltimore County · University of Notre Dame · UNC Chapel Hill
POPS: Recovering Unlearned Multi-Modality Knowledge in MLLMs with Prompt-Optimized Parameter Shaking
Jul 7, 2026Zhangheng LI, Jianing Zhu, Junyuan Hong +4Multimodal Large Language Models Privacy Leakage in Language Models
University of Texas at Austin · DEVCOM Army Research Laboratory
Auditing of Unlearning Algorithms
Jul 7, 2026Sahasrajit Sarmasarkar, Anastasia Koloskova, Sanmi KoyejoMachine Unlearning Machine Unlearning Evaluation
Stanford University · University of Zurich
Association Restoration Test: Revealing Restorable Shortcuts after Unlearning
Jul 7, 2026Amy Lu, Changxiu JiSpurious Correlation Robustness Shortcut Learning
Department of Computer Science, Stanford University
MPSelectTune: Prompt-type Selection for Fine-tuning improves Concept Unlearning in LLMs
Jul 4, 2026Shubhadip Nag, Srinjoy Das, Agniva Saha +5LLM Fine-Tuning Machine Unlearning
*IIT Kharagpur · †MPI-SWS · ‡HPE Labs
MUTE: Return-Preserving Communication Unlearning for Efficient Multi-Agent Coordination
Jul 3, 2026Rui Zuo, Qinwei Huang, Mingyang Li +3Multi-Agent System Optimization Multi-Agent Coordination
Syracuse University · Air Force Research Laboratory
PPE-Bench: A Benchmark for Evaluating MLLM Unlearning under Private-Public Entanglement
Jul 3, 2026Xianren Zhang, Delvin Ce Zhang, Dongwon Lee +1Privacy Leakage in Language Models MLLM Unlearning
The Pennsylvania State University · University of Sheffield
LACUNA: A Testbed for Evaluating Localization Precision for LLM Unlearning
Jul 2, 2026Matteo Boglioni, Thibault Rousset, Siva Reddy +2Privacy Leakage in Language Models Machine Unlearning
Mila – Quebec Artificial Intelligence Institute · McGill University · McGill-NLP/LACUNA
DriftScope: Measuring The Hidden Effects of Diffusion Model Adaptation
Jun 30, 2026Héctor Laria, Yiping Han, Julian D. Santamaria +4Diffusion Model Adaptation Diffusion Model Unlearning
Computer Vision Center, Barcelona, Spain · Universitat Autonoma de Barcelona, Barcelona, Spain · Program of Computer Science, City University of Hong Kong (Dongguan), China +1
Look But Don't Touch with Sparse Autoencoders for Unlearning in Diffusion Models
Jun 30, 2026Enrico Cassano, Riccardo Renzulli, Rayyan Ahmed +2Sparse Autoencoders Diffusion Model Unlearning
University of Turin, Computer Science Department, Italy · LTCI, Télécom Paris, Institut Polytechnique de Paris, France

Aug 3, 2026J/K move · Enter open · S save

Exploring and Bridging Knowledge Holes in Unlearned Multimodal Large Language Models

Junxiang You, Junkai Chen, Yuhao He, Ruiqi Liu+2

University of Chinese Academy of Sciences · Institute of Automation, Chinese Academy of Sciences · 3Cloudspace Technology

Multimodal Large Language Models MLLM Unlearning Machine Unlearning LLM Unlearning Machine Unlearning Evaluation MLLM Safety

Open paper AskPDF arXiv

Machine unlearning offers a promising approach to remove unsafe content from Multimodal Large Language Models (MLLMs), yet ensuring the precision of unlearning remains a persistent challenge. One reason is that current MLLM unlearning evaluation paradigms suffer from a critical blind spot: they assess model utility through benchmarks whose representations are distant from the forget set, failing to capture knowledge holes---severe degradation on benign adjacent inputs. To probe knowledge holes in unlearned MLLMs, we construct a benchmark that captures unintended degradation on benign inputs sharing generic patterns with the forget set, and confirm through controlled experiments that they are a systematic consequence of commonly used approaches. Furthermore, to bridge this gap, we propose Selective Protection with Anchored Regularization, which protects generic patterns via anchored activation filtering while reinforcing them through entity-abstracted enhancement. Our experiments on SafeEraser demonstrate that SPAR recovers over 98% of vanilla response quality compared to below 50% for standard baselines---while achieving 0.00% attack success rate and competitive model utility. These results underscore the necessity of more fine-grained evaluation for trustworthy MLLM unlearning.