mlsanityData to 7 Oct 2026

Understanding and Mitigating Copying in Diffusion Models

arXiv 2305.20086 ↗First posted 31 May 2023cs.LG cs.CR cs.CVcited by 205

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Images generated by diffusion models like Stable Diffusion are increasingly widespread. Recent works and even lawsuits have shown that these models are prone to replicating their training data, unbeknownst to the user. In this paper, we first analyze this memorization problem in text-to-image diffusion models. While it is widely believed that duplicated images in the training set are responsible for content replication at inference time, we observe that the text conditioning of the model plays a similarly important role. In fact, we see in our experiments that data replication often does not happen for unconditional models, while it is common in the text-conditional case. Motivated by our findings, we then propose several techniques for reducing data replication at both training and inference time by randomizing and augmenting image captions in the training set.

Key facts. Cited by 205 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 21 of its 30 references that are in the map. First posted on arXiv 31 May 2023.

Builds on

21 of 30 references.

Closest ancestors · 3

Same direction, newest first

  1. 2023On the De-duplication of LAION-2B17 Mar
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  3. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 18

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  3. 2022BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generationcited by 4,310
  4. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  5. 2021CLIPScore: A Reference-free Evaluation Metric for Image Captioningcited by 2,388
  6. 2022Quantifying Memorization Across Neural Language Modelscited by 755
  7. 2023Extracting Training Data from Diffusion Modelscited by 723
  8. 2022Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Modelscited by 501
  9. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Modelscited by 418
  10. 2022DiffusionDB: A Large-scale Prompt Gallery Dataset for Text-to-Image Generative Modelscited by 376
  11. 2023Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discoverycited by 329
  12. 2023Ablating Concepts in Text-to-Image Diffusion Modelscited by 312
6 more
  1. 2022Red-Teaming the Stable Diffusion Safety Filtercited by 257
  2. 2022A Self-Supervised Descriptor for Image Copy Detectioncited by 175
  3. 2022Measuring Forgetting of Memorized Training Examplescited by 118
  4. 2023On Provable Copyright Protection for Generative Modelscited by 97
  5. 2022Do Language Models Plagiarize?cited by 72
  6. 2021This Person (Probably) Exists. Identity Membership Attacks Against GAN Generated Facescited by 25

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 27; 2024: 68; 2025: 70; 2026: 40.

205 papers.

Sort:

By area:

  1. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,094
  2. 2023Unified Concept Editing in Diffusion Modelscited by 377
  3. 2023One-Dimensional Adapter to Rule Them All: Concepts, Diffusion Models and Erasing Applicationscited by 164
  4. 2023CADS: Unleashing the Diversity of Diffusion Models through Condition-Annealed Samplingcited by 124
  5. 2023On Memorization in Diffusion Modelscited by 122
  6. 2024Measuring Style Similarity in Diffusion Modelscited by 106
  7. 2024Detecting, Explaining, and Mitigating Memorization in Diffusion Modelscited by 94
  8. 2023When Foundation Model Meets Federated Learning: Motivations, Challenges, and Future Directionscited by 83
  9. 2023Circumventing Concept Erasure Methods For Text-to-Image Generative Modelscited by 81
  10. 2025Why Diffusion Models Don't Memorize: The Role of Implicit Dynamical Regularization in Trainingcited by 78
  11. 2023On the Design Fundamentals of Diffusion Models: A Surveycited by 59
  12. 2024Unveiling and Mitigating Memorization in Text-to-image Diffusion Models through Cross Attentioncited by 57
38 more
  1. 2024R.A.C.E.: Robust Adversarial Concept Erasure for Secure Text-to-Image Diffusion Modelcited by 56
  2. 2024Breaking the Curse of Dimensionality: Diffusion Models Efficiently Learn Low-Dimensional Distributionscited by 53
  3. 2024Erasing Undesirable Influence in Diffusion Modelscited by 52
  4. 2024Towards Memorization-Free Diffusion Modelscited by 49
  5. 2025On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticitycited by 48
  6. 2024Consistent Diffusion Meets Tweedie: Training Exact Ambient Diffusion Models with Noisy Datacited by 44
  7. 2024A Geometric Framework for Understanding Memorization in Generative Modelscited by 40
  8. 2024STEREO: A Two-Stage Framework for Adversarially Robust Concept Erasing from Text-to-Image Diffusion Modelscited by 39
  9. 2024Understanding Generalizability of Diffusion Models Requires Rethinking the Hidden Gaussian Structurecited by 38
  10. 2024A Good Score Does not Lead to A Good Generative Modelcited by 36
  11. 2024Finding NeMo: Localizing Neurons Responsible For Memorization in Diffusion Modelscited by 36
  12. 2024Copyright Protection in Generative AI: A Technical Perspectivecited by 35
  13. 2024The Stronger the Diffusion Model, the Easier the Backdoor: Data Poisoning to Induce Copyright Breaches Without Adjusting Finetuning Pipelinecited by 33
  14. 2024Losing dimensions: Geometric memorization in generative diffusioncited by 28
  15. 2024Diffusion Soup: Model Merging for Text-to-Image Diffusion Modelscited by 25
  16. 2025Bigger Isn't Always Memorizing: Early Stopping Overparameterized Diffusion Modelscited by 24
  17. 2023Federated Generative Learning with Foundation Modelscited by 23
  18. 2025Generalization through variance: how noise shapes inductive biases in diffusion modelscited by 23
  19. 2024CPR: Retrieval Augmented Generation for Copyright Protectioncited by 22
  20. 2024Automatic Jailbreaking of the Text-to-Image Generative AI Systemscited by 22
  21. 2023The Emergence of Reproducibility and Generalizability in Diffusion Modelscited by 21
  22. 2024Classifier-Free Guidance inside the Attraction Basin May Cause Memorizationcited by 21
  23. 2023On Copyright Risks of Text-to-Image Diffusion Modelscited by 20
  24. 2024A Survey on Diffusion Models for Recommender Systemscited by 20
  25. 2024Understanding and Mitigating Memorization in Generative Models via Sharpness of Probability Landscapescited by 20
  26. 2025Denoising Score Matching with Random Features: Insights on Diffusion Models from Precise Learning Curvescited by 20
  27. 2023Investigating Data Memorization in 3D Latent Diffusion Models for Medical Image Synthesiscited by 19
  28. 2024Exploring Local Memorization in Diffusion Models via Bright Ending Attentioncited by 19
  29. 2025Training-Free Safe Denoisers for Safe Use of Diffusion Modelscited by 19
  30. 2024Hidden in the Noise: Two-Stage Robust Watermarking for Imagescited by 18
  31. 2025Data Unlearning in Diffusion Modelscited by 18
  32. 2025Ambient Diffusion Omni: Training Good Models with Bad Datacited by 18
  33. 2025A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspectivecited by 18
  34. 2025Does Generation Require Memorization? Creative Diffusion Models using Ambient Diffusioncited by 17
  35. 2023Variational Bayesian Imaging with an Efficient Surrogate Score-based Priorcited by 16
  36. 2025Taking a Big Step: Large Learning Rates in Denoising Score Matching Prevent Memorizationcited by 16
  37. 2023EDMSound: Spectrogram Based Diffusion Models for Efficient and High-Quality Audio Synthesiscited by 15
  38. 2024A Survey on Safe Multi-Modal Learning Systemcited by 15

50 most cited of 205 All 205 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Finding DoRI: Discovery of Retained Images in Diffusion Models
  2. 2024Not Every Image is Worth a Thousand Words: Quantifying Originality in Stable Diffusion
  3. 2025Locality-Aware Continual Unlearning for Diffusion Models
  4. 2025Reusing Computation in Text-to-Image Diffusion for Efficient Generation of Image Sets
  5. 2026i1: A Simple and Fully Open Recipe for Strong Text-to-Image Models
  6. 2025Preserve and Personalize: Personalized Text-to-Image Diffusion Models without Distributional Drift
  7. 2025On the Edge of Memorization in Diffusion Models
  8. 2024Membership Inference on Text-to-Image Diffusion Models via Conditional Likelihood Discrepancy