mlsanityData to 7 Oct 2026

SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in Both Image Classification and Generation

arXiv 2310.12508 ↗First posted 19 Oct 2023cs.LG cs.AIcited by 376

On the map: Diffusion models → Robustness, privacy and adversarial aspects

also in 1 more
Abstract (arXiv)

With evolving data regulations, machine unlearning (MU) has become an important tool for fostering trust and safety in today's AI models. However, existing MU methods focusing on data and/or weight perspectives often suffer limitations in unlearning accuracy, stability, and cross-domain applicability. To address these challenges, we introduce the concept of 'weight saliency' for MU, drawing parallels with input saliency in model explanation. This innovation directs MU's attention toward specific model weights rather than the entire model, improving effectiveness and efficiency. The resultant method that we call saliency unlearning (SalUn) narrows the performance gap with 'exact' unlearning (model retraining from scratch after removing the forgetting data points). To the best of our knowledge, SalUn is the first principled MU approach that can effectively erase the influence of forgetting data, classes, or concepts in both image classification and generation tasks. As highlighted below, For example, SalUn yields a stability advantage in high-variance random data forgetting, e.g., with a 0.2% gap compared to exact unlearning on the CIFAR-10 dataset. Moreover, in preventing conditional diffusion models from generating harmful images, SalUn achieves nearly 100% unlearning accuracy, outperforming current state-of-the-art baselines like Erased Stable Diffusion and Forget-Me-Not. Codes are available at https://github.com/OPTML-Group/Unlearn-Saliency. (WARNING: This paper contains model outputs that may be offensive in nature.)

Key facts. Cited by 376 papers in the map (data through 7 Oct 2026). Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models. Builds on 63 of its 75 references that are in the map. First posted on arXiv 19 Oct 2023.

Builds on

63 of 75 references.

Closest ancestors · 4

Same direction, newest first

  1. 2023To Generate or Not? Safety-Driven Unlearned Diffusion Models Are Still Easy To Generate Unsafe Images ... For Now18 Oct
  2. 2023Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models30 Mar
  3. 2023Erasing Concepts from Diffusion Models13 Mar
  4. 2022Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Models9 Nov

Other references · 59

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  3. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  4. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  5. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,690
  6. 2022Classifier-Free Diffusion Guidancecited by 5,717
  7. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  8. 2017Axiomatic Attribution for Deep Networkscited by 3,350
  9. 2022Locating and Editing Factual Associations in GPTcited by 2,609
  10. 2015Learning Deep Features for Discriminative Localizationcited by 2,430
  11. 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,348
  12. 2015Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Codingcited by 2,291
38 more
  1. 2013Visualizing and Understanding Convolutional Networkscited by 2,278
  2. 2018The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networkscited by 2,250
  3. 2017Understanding Black-box Predictions via Influence Functionscited by 1,714
  4. 2021LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairscited by 1,544
  5. 2018Sanity Checks for Saliency Mapscited by 1,149
  6. 2017SmoothGrad: removing noise by adding noisecited by 1,097
  7. 2014Striving for Simplicity: The All Convolutional Netcited by 983
  8. 2017Grad-CAM++: Improved Visual Explanations for Deep Convolutional Networkscited by 936
  9. 2021Membership Inference Attacks From First Principlescited by 839
  10. 2018RISE: Randomized Input Sampling for Explanation of Black-box Modelscited by 825
  11. 2021Scaling Vision with Sparse Mixture of Expertscited by 740
  12. 2019Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networkscited by 599
  13. 2021Knowledge Neurons in Pretrained Transformerscited by 590
  14. 2021Editing Factual Knowledge in Language Modelscited by 548
  15. 2019Certified Data Removal from Machine Learning Modelscited by 512
  16. 2019Making AI Forget You: Data Deletion in Machine Learningcited by 447
  17. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Modelscited by 418
  18. 2021Remember What You Want to Forget: Algorithms for Machine Unlearningcited by 360
  19. 2021Multimodal datasets: misogyny, pornography, and malignant stereotypescited by 326
  20. 2023Data Selection for Language Models via Importance Resamplingcited by 322
  21. 2023Studying Large Language Model Generalization with Influence Functionscited by 311
  22. 2020Amnesiac Machine Learningcited by 305
  23. 2020Descent-to-Delete: Gradient-Based Methods for Machine Unlearningcited by 303
  24. 2021Unrolling SGD: Understanding Factors Influencing Machine Unlearningcited by 274
  25. 2023TRAK: Attributing Model Behavior at Scalecited by 273
  26. 2020Approximate Data Deletion from Machine Learning Modelscited by 250
  27. 2021Machine Unlearning of Features and Labelscited by 223
  28. 2022Datamodels: Predicting Predictions from Training Datacited by 186
  29. 2020fastai: A Layered API for Deep Learningcited by 183
  30. 2023Selective Amnesia: A Continual Learning Approach to Forgetting in Deep Generative Modelscited by 175
  31. 2021On the Necessity of Auditable Algorithmic Definitions for Machine Unlearningcited by 164
  32. 2023Can Sensitive Information Be Deleted From LLMs? Objectives for Defending Against Extraction Attackscited by 156
  33. 2023Boundary Unlearningcited by 132
  34. 2021Graph Unlearningcited by 129
  35. 2021Machine Unlearning via Algorithmic Stabilitycited by 128
  36. 2022The Right to be Forgotten in Federated Learning: An Efficient Realization with Rapid Retrainingcited by 109
  37. 2021Federated Unlearning via Class-Discriminative Pruningcited by 108
  38. 2021Leveraging Sparse Linear Layers for Debuggable Deep Networkscited by 82

50 most cited of 59 All 59 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 2; 2024: 78; 2025: 141; 2026: 155.

376 papers.

Sort:

By area:

  1. 2024MUSE: Machine Unlearning Six-Way Evaluation for Language Modelscited by 251
  2. 2024Rethinking Machine Unlearning for Large Language Modelscited by 247
  3. 2024Defensive Unlearning with Adversarial Training for Robust Concept Erasure in Diffusion Modelscited by 164
  4. 2024Large Language Model Unlearning via Embedding-Corrupted Promptscited by 112
  5. 2024SOUL: Unlocking the Power of Second-Order Optimization for LLM Unlearningcited by 104
  6. 2024What makes unlearning hard and what to do about itcited by 89
  7. 2024Catastrophic Failure of LLM Unlearning via Quantizationcited by 76
  8. 2025Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyondcited by 76
  9. 2025SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoderscited by 75
  10. 2024LLM Unlearning via Loss Adjustment with Only Forget Datacited by 71
  11. 2025Set You Straight: Auto-Steering Denoising Trajectories to Sidestep Unwanted Conceptscited by 61
  12. 2024Erasing Undesirable Influence in Diffusion Modelscited by 52
38 more
  1. 2024ConceptPrune: Concept Editing in Diffusion Models via Skilled Neuron Pruningcited by 51
  2. 2024Scissorhands: Scrub Data Influence via Connection Sensitivity in Networkscited by 50
  3. 2024Machine Unlearning in Generative AI: A Surveycited by 49
  4. 2024Are we making progress in unlearning? Findings from the first NeurIPS unlearning competitioncited by 48
  5. 2024UnlearnCanvas: Stylized Image Dataset for Enhanced Machine Unlearning Evaluation in Diffusion Modelscited by 47
  6. 2024UnUnlearning: Unlearning is not sufficient for content regulation in advanced generative AIcited by 46
  7. 2024Challenging Forgets: Unveiling the Worst-Case Forget Sets in Machine Unlearningcited by 44
  8. 2024Single Image Unlearning: Efficient Machine Unlearning in Multimodal Large Language Modelscited by 44
  9. 2024Unified Gradient-Based Machine Unlearning with Remain Geometry Enhancementcited by 43
  10. 2025Exploring Criteria of Loss Reweighting to Enhance LLM Unlearningcited by 39
  11. 2024Machine Unlearning Fails to Remove Data Poisoning Attackscited by 35
  12. 2024Revisiting Who's Harry Potter: Towards Targeted Unlearning from a Causal Intervention Perspectivecited by 33
  13. 2024Jailbreak Attacks and Defenses against Multimodal Generative Models: A Surveycited by 31
  14. 2024Erasing Concepts from Text-to-Image Diffusion Models with Few-shot Unlearningcited by 29
  15. 2025MMUnlearner: Reformulating Multimodal Machine Unlearning in the Era of Multimodal Large Language Modelscited by 29
  16. 2025Decoupled Distillation to Erase: A General Unlearning Method for Any Class-centric Taskscited by 28
  17. 2024On Effects of Steering Latent Representation for Large Language Model Unlearningcited by 27
  18. 2025Localized Concept Erasure for Text-to-Image Diffusion Models Using Training-Free Gated Low-Rank Adaptationcited by 27
  19. 2024To Forget or Not? Towards Practical Knowledge Unlearning for Large Language Modelscited by 25
  20. 2024WAGLE: Strategic Weight Attribution for Effective and Modular Unlearning in Large Language Modelscited by 25
  21. 2024TraSCE: Trajectory Steering for Concept Erasurecited by 25
  22. 2024Towards Robust Knowledge Unlearning: An Adversarial Framework for Assessing and Improving Unlearning Robustness in Large Language Modelscited by 24
  23. 2025CURE: Concept Unlearning via Orthogonal Representation Editing in Diffusion Modelscited by 24
  24. 2025BLUR: A Bi-Level Optimization Approach for LLM Unlearningcited by 23
  25. 2024Avoiding Copyright Infringement via Large Language Model Unlearningcited by 22
  26. 2024Learning to Refuse: Towards Mitigating Privacy Risks in LLMscited by 22
  27. 2024Score Forgetting Distillation: A Swift, Data-Free Method for Machine Unlearning in Diffusion Modelscited by 22
  28. 2023MultiDelete for Multimodal Machine Unlearningcited by 20
  29. 2024Machine Unlearning via Null Space Calibrationcited by 19
  30. 2024Learning to Unlearn for Robust Machine Unlearningcited by 19
  31. 2024Holistic Unlearning Benchmark: A Multi-Faceted Evaluation for Text-to-Image Diffusion Model Unlearningcited by 19
  32. 2025Learning to Unlearn while Retaining: Combating Gradient Conflicts in Machine Unlearningcited by 19
  33. 2025Data Unlearning in Diffusion Modelscited by 18
  34. 2024Zero-Shot Class Unlearning in CLIP with Synthetic Samplescited by 17
  35. 2024MUNBa: Machine Unlearning via Nash Bargainingcited by 17
  36. 2025Tool Unlearning for Tool-Augmented LLMscited by 17
  37. 2025GUARD: Generation-time LLM Unlearning via Adaptive Restriction and Detectioncited by 17
  38. 2024MU-Bench: A Multitask Multimodal Benchmark for Machine Unlearningcited by 16

50 most cited of 376 All 376 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Why Some Models Resist Unlearning: A Linear Stability Perspective
  2. 2026ICED: Concept-level Machine Unlearning via Interpretable Concept Decomposition
  3. 2025UNO: Unlearning via Orthogonalization in Generative models
  4. 2026Before Forgetting, Learn to Remember: Revisiting Foundational Learning Failures in LVLM Unlearning Benchmarks
  5. 2026Suppression Is Not Forgetting: Residual Recoverability in Visual Concept Unlearning for VLMs
  6. 2026CATA: Continual Machine Unlearning via Conflict-Averse Task Arithmetic
  7. 2026Stochastic Meta-Unlearning: Bridging Language Backbone and Multimodal Unlearning
  8. 2023Learning Saliency From Fixations