mlsanityData to 14 Oct 2026

Not Every Time and Frequency Need to Be Forgotten in Diffusion Unlearning

arXiv 2510.17917 ↗First posted 20 Oct 2025cs.LG cs.AI cs.CRcited by 2

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

also in 1 more
Abstract (arXiv)

Data unlearning aims to remove the influence of specific training samples from a trained model. In fine-tuning methods, data unlearning relies primarily on loss maximization over forget samples, which often leads to quality degradation or incomplete forgetting. Existing methods perform unlearning uniformly across diffusion stages, ignoring diffusion dynamics from noise to data. Our systematic study of diffusion phases shows that forgetting in diffusion models is uneven across time and frequency, with theoretical justification of distributive distortion and forgetting-utility trade-off. By selectively forgetting time and frequency in diffusion models, we achieve both higher unlearning success rates and improved generation quality across diverse settings, including both conditional and unconditional scenarios. We also introduce an improved SSCD metric that measures dissimilarity using a normalized perturbation distance. Together, we provide practical insights for understanding and improving data unlearning in diffusion models.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models. Builds on 47 of its 56 references that are in the map. First posted on arXiv 20 Oct 2025.

Builds on

47 of 56 references.

Closest ancestors · 12

Same direction, newest first

  1. 2025Data Unlearning in Diffusion Models2 Mar
  2. 2025Training-Free Safe Denoisers for Safe Use of Diffusion Models11 Feb
  3. 2024Meta-Unlearning on Diffusion Models: Preventing Relearning Unlearned Concepts16 Oct
  4. 2024SAFREE: Training-Free and Adaptive Guard for Safe Text-to-Image And Video Generation16 Oct
  5. 2024STEREO: A Two-Stage Framework for Adversarially Robust Concept Erasing from Text-to-Image Diffusion Models29 Aug
  6. 2024Membership Inference on Text-to-Image Diffusion Models via Conditional Likelihood Discrepancy23 May
  7. 2024Erasing Undesirable Influence in Diffusion Models11 Jan
  8. 2023SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in Both Image Classification and Generation19 Oct
  9. 2023Ring-A-Bell! How Reliable are Concept Removal Methods for Diffusion Models?16 Oct
  10. 2023An Efficient Membership Inference Attack for the Diffusion Model by Proximal Initialization26 May
  11. 2023Erasing Concepts from Diffusion Models13 Mar
  12. 2023Are Diffusion Models Vulnerable to Membership Inference Attacks?2 Feb

Other references · 35

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  3. 2020Denoising Diffusion Implicit Modelscited by 9,781
  4. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  5. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  6. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  7. 2022Scalable Diffusion Models with Transformerscited by 6,054
  8. 2022Generative Adversarial Networkscited by 5,627
  9. 2014Deep Learning Face Attributes in the Wildcited by 3,693
  10. 2019Machine Unlearningcited by 1,061
  11. 2022Exploring CLIP for Assessing the Look and Feel of Imagescited by 991
  12. 2023Diffusion Model Alignment Using Direct Preference Optimizationcited by 790
23 more
  1. 2019Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networkscited by 618
  2. 2021Deep Learning on a Data Diet: Finding Important Examples Early in Trainingcited by 533
  3. 2024Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearningcited by 504
  4. 2022Perception Prioritized Training of Diffusion Modelscited by 303
  5. 2025A General Framework for Inference-time Scaling and Steering of Diffusion Modelscited by 224
  6. 2022A Self-Supervised Descriptor for Image Copy Detectioncited by 176
  7. 2024Dynamical Regimes of Diffusion Modelscited by 142
  8. 2024Aligning Diffusion Models by Optimizing Human Utilitycited by 101
  9. 2024Detecting, Explaining, and Mitigating Memorization in Diffusion Modelscited by 95
  10. 2024A Phase Transition in Diffusion Models Reveals the Hierarchical Nature of Datacited by 93
  11. 2023Spontaneous Symmetry Breaking in Generative Diffusion Modelscited by 80
  12. 2024Direct Unlearning Optimization for Robust and Safe Text-to-Image Modelscited by 50
  13. 2024Critical windows: non-asymptotic theory for feature emergence in diffusion modelscited by 47
  14. 2023The Journey, Not the Destination: How Data Guides Diffusion Modelscited by 43
  15. 2024Dynamic Negative Guidance of Diffusion Modelscited by 28
  16. 2024Reconstruction Attacks on Machine Unlearning: Simple Models are Vulnerablecited by 25
  17. 2023Denoising Diffusion Step-aware Modelscited by 24
  18. 2025Diffusion-NPO: Negative Preference Optimization for Better Preference Aligned Generation of Diffusion Modelscited by 23
  19. 2024Shielded Diffusion: Generating Novel and Diverse Images using Sparse Repellencycited by 22
  20. 2024Classifier-Free Guidance inside the Attraction Basin May Cause Memorizationcited by 21
  21. 2024Understanding and Mitigating Memorization in Generative Models via Sharpness of Probability Landscapescited by 21
  22. 2025LLM Unlearning Reveals a Stronger-Than-Expected Coreset Effect in Current Benchmarkscited by 12
  23. 2025Blink of an eye: a simple theory for feature localization in generative modelscited by 11

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026Unlearning in Diffusion Models: A Unified Framework with KL Divergence and Likelihood Constraintscited by 2
  2. 2026Mitigating Sequential Reappearance in Diffusion Data-Point Unlearningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Continual Data Unlearning in Diffusion Models via Transition-based Regularization
  2. 2025ReTrack: Data Unlearning in Diffusion Models through Redirecting the Denoising Trajectory
  3. 2026Enabling Preference-driven Unlearning in Few-step Distilled Text-to-Image Diffusion Models
  4. 2023The Emergence of Reproducibility and Generalizability in Diffusion Models
  5. 2023Exploring Continual Learning of Diffusion Models
  6. 2026Understanding diffusion models requires rethinking (again) generalization
  7. 2026Weeding Out Bad Seeds: Initial-Noise-Robust Unlearning for Text-to-Image Diffusion Models
  8. 2024Unlearning or Concealment? A Critical Analysis and Evaluation Metrics for Unlearning in Diffusion Models