mlsanityData to 7 Oct 2026

Unveiling and Mitigating Memorization in Text-to-image Diffusion Models through Cross Attention

arXiv 2403.11052 ↗First posted 17 Mar 2024cs.CV cs.CRcited by 57

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Recent advancements in text-to-image diffusion models have demonstrated their remarkable capability to generate high-quality images from textual prompts. However, increasing research indicates that these models memorize and replicate images from their training data, raising tremendous concerns about potential copyright infringement and privacy risks. In our study, we provide a novel perspective to understand this memorization phenomenon by examining its relationship with cross-attention mechanisms. We reveal that during memorization, the cross-attention tends to focus disproportionately on the embeddings of specific tokens. The diffusion model is overfitted to these token embeddings, memorizing corresponding training images. To elucidate this phenomenon, we further identify and discuss various intrinsic findings of cross-attention that contribute to memorization. Building on these insights, we introduce an innovative approach to detect and mitigate memorization in diffusion models. The advantage of our proposed method is that it will not compromise the speed of either the training or the inference processes in these models while preserving the quality of generated images. Our code is available at https://github.com/renjie3/MemAttn .

Key facts. Cited by 57 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 33 of its 39 references that are in the map. First posted on arXiv 17 Mar 2024.

Builds on

33 of 39 references.

Closest ancestors · 7

Same direction, newest first

  1. 2023Understanding (Un)Intended Memorization in Text-to-Image Generative Models6 Dec
  2. 2023Memory Triggers: Unveiling Memorization in Text-To-Image Generative Models through Word-Level Duplication6 Dec
  3. 2023PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis30 Sep
  4. 2023Understanding and Mitigating Copying in Diffusion Models31 May
  5. 2023On the De-duplication of LAION-2B17 Mar
  6. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  7. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 26

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2023GPT-4 Technical Reportcited by 18,132
  3. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,539
  4. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,322
  5. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  6. 2022DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generationcited by 3,286
  7. 2022Prompt-to-Prompt Image Editing with Cross Attention Controlcited by 2,409
  8. 2020Extracting Training Data from Large Language Modelscited by 2,144
  9. 2023Diffusion Model Alignment Using Direct Preference Optimizationcited by 767
  10. 2022Quantifying Memorization Across Neural Language Modelscited by 755
  11. 2023Extracting Training Data from Diffusion Modelscited by 723
  12. 2023Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Modelscited by 696
14 more
  1. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Modelscited by 418
  2. 2018Stacked Cross Attention for Image-Text Matchingcited by 414
  3. 2023ELITE: Encoding Visual Concepts into Textual Embeddings for Customized Text-to-Image Generationcited by 414
  4. 2023Video-P2P: Video Editing with Cross-attention Controlcited by 312
  5. 2019Analyzing the Structure of Attention in a Transformer Language Modelcited by 241
  6. 2022What the DAAM: Interpreting Stable Diffusion Using Cross Attentioncited by 233
  7. 2022A Self-Supervised Descriptor for Image Copy Detectioncited by 175
  8. 2023Stabilizing Transformer Training by Preventing Attention Entropy Collapsecited by 154
  9. 2022The Privacy Onion Effect: Memorization is Relativecited by 128
  10. 2023Ambient Diffusion: Learning Clean Distributions from Corrupted Datacited by 108
  11. 2023Directed Diffusion: Direct Control of Object Placement through Attention Guidancecited by 69
  12. 2023A Reproducible Extraction of Training Images from Diffusion Modelscited by 60
  13. 2022Memorization in NLP Fine-tuning Methodscited by 43
  14. 2020Text-to-Image Generation Grounded by Fine-Grained User Attentioncited by 36

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 17; 2025: 22; 2026: 18.

57 papers.

Sort:

By area:

  1. 2024STEREO: A Two-Stage Framework for Adversarially Robust Concept Erasing from Text-to-Image Diffusion Modelscited by 39
  2. 2024Finding NeMo: Localizing Neurons Responsible For Memorization in Diffusion Modelscited by 36
  3. 2025On the Edge of Memorization in Diffusion Modelscited by 22
  4. 2024Classifier-Free Guidance inside the Attraction Basin May Cause Memorizationcited by 21
  5. 2024Understanding and Mitigating Memorization in Generative Models via Sharpness of Probability Landscapescited by 20
  6. 2024Exploring Local Memorization in Diffusion Models via Bright Ending Attentioncited by 19
  7. 2025Does Generation Require Memorization? Creative Diffusion Models using Ambient Diffusioncited by 17
  8. 2024Memorized Images in Diffusion Models share a Subspace that can be Located and Deletedcited by 13
  9. 2024Replication in Visual Diffusion Models: A Survey and Outlookcited by 11
  10. 2025Generalization of Diffusion Models Arises with a Balanced Representation Spacecited by 10
  11. 2025Demystifying Diffusion Policies: Action Memorization and Simple Lookup Table Alternativescited by 9
  12. 2024SIDE: Surrogate Conditional Data Extraction from Diffusion Modelscited by 7
38 more
  1. 2025Adjusting Initial Noise to Mitigate Memorization in Text-to-Image Diffusion Modelscited by 7
  2. 2024An Inversion-based Measure of Memorization for Diffusion Modelscited by 6
  3. 2024Trustworthy Text-to-Image Diffusion Models: A Timely and Focused Surveycited by 6
  4. 2025How Diffusion Models Memorizecited by 6
  5. 2024MemBench: Memorized Image Trigger Prompt Dataset for Diffusion Modelscited by 5
  6. 2024Investigating Memorization in Video Diffusion Modelscited by 5
  7. 2024CopyrightShield: Enhancing Diffusion Model Security against Copyright Infringement Attackscited by 5
  8. 2025Enhancing Privacy-Utility Trade-offs to Mitigate Memorization in Diffusion Modelscited by 5
  9. 2025Diffusion models under low-noise regimecited by 5
  10. 2025Gradient Variance Reveals Failure Modes in Flow-Based Generative Modelscited by 5
  11. 2026Nucleus-Image: Sparse MoE for Image Generationcited by 5
  12. 2025LCMem: A Universal Model for Robust Image Memorization Detectioncited by 4
  13. 2024MemControl: Mitigating Memorization in Diffusion Models via Automated Parameter Selectioncited by 3
  14. 2024Dual-Model Defense: Safeguarding Diffusion Models from Membership Inference Attacks through Disjoint Data Splittingcited by 3
  15. 2025SolidMark: Evaluating Image Memorization in Generative Modelscited by 3
  16. 2025Responsible Diffusion: A Comprehensive Survey on Safety, Ethics, and Trust in Diffusion Modelscited by 3
  17. 2025The Devil is in the Prompts: De-Identification Traces Enhance Memorization Risks in Synthetic Chest X-Ray Generationcited by 2
  18. 2025Demystifying Foreground-Background Memorization in Diffusion Modelscited by 2
  19. 2026Detecting and Mitigating Memorization in Diffusion Models through Anisotropy of the Log-Probabilitycited by 2
  20. 2026Diffusion Models Memorize in Training -- and Generalize in Inferencecited by 2
  21. 2024Leveraging Model Guidance to Extract Training Data from Personalized Diffusion Modelscited by 1
  22. 2025Sparse Fine-Tuning of Transformers for Generative Taskscited by 1
  23. 2025LLM in the Middle: A Systematic Review of Threats and Mitigations to Real-World LLM-based Systemscited by 1
  24. 2025Dominant vs. Dominated: Concept-Level Generative Collapse in Diffusion Modelscited by 1
  25. 2026The Illusion of Forgetting: Attack Unlearned Diffusion via Initial Latent Variable Optimizationcited by 1
  26. 2026Steering Away from Memorization: Reachability-Constrained Reinforcement Learning for Text-to-Image Diffusioncited by 1
  27. 2024EnTruth: Enhancing the Traceability of Unauthorized Dataset Usage in Text-to-image Diffusion Models with Minimal and Robust Alterationscited by 0
  28. 2025Redistribute Ensemble Training for Mitigating Memorization in Diffusion Modelscited by 0
  29. 2025FPAN: Mitigating Replication in Diffusion Models through the Fine-Grained Probabilistic Addition of Noise to Token Embeddingscited by 0
  30. 2025The Persistence of Cultural Memory: Investigating Multimodal Iconicity in Diffusion Modelscited by 0
  31. 2025CAPTAIN: Semantic Feature Injection for Memorization Mitigation in Text-to-Image Diffusion Modelscited by 0
  32. 2025Beyond Memorization: Selective Learning for Copyright-Safe Diffusion Model Trainingcited by 0
  33. 2026Noise as a Probe: Membership Inference Attacks on Diffusion Models Leveraging Initial Noisecited by 0
  34. 2026Consistency-Preserving Concept Erasure via Unsafe-Safe Pairing and Directional Fisher-weighted Adaptationcited by 0
  35. 2026Benchmarking Unlearning for Vision Transformerscited by 0
  36. 2026You Don't Need All That Attention: Surgical Memorization Mitigation in Text-to-Image Diffusion Modelscited by 0
  37. 2026Memorization In Stable Diffusion Is Unexpectedly Driven by CLIP Embeddingscited by 0
  38. 2026Filtering Memorization from Parameter-Space in Diffusion Modelscited by 0

50 most cited of 57 All 57 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024Detecting, Explaining, and Mitigating Memorization in Diffusion Models
  2. 2025Finding DoRI: Discovery of Retained Images in Diffusion Models
  3. 2024Enhancing Semantic Fidelity in Text-to-Image Synthesis: Attention Regulation in Diffusion Models
  4. 2026Mitigating Memorization in Text-to-Image Diffusion via Region-Aware Prompt Augmentation and Multimodal Copy Detection
  5. 2024Attention Overlap Is Responsible for The Entity Missing Problem in Text-to-image Diffusion Models!
  6. 2023CONFORM: Contrast is All You Need For High-Fidelity Text-to-Image Diffusion Models
  7. 2023A-STAR: Test-time Attention Segregation and Retention for Text-to-image Synthesis
  8. 2026Concept Unlearning via Cross-Attention Activation Projection for Diffusion Models