mlsanityData to 7 Oct 2026

Towards Safe Self-Distillation of Internet-Scale Text-to-Image Diffusion Models

arXiv 2307.05977 ↗First posted 12 Jul 2023cs.CV cs.AI cs.LGcited by 41

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

Abstract (arXiv)

Large-scale image generation models, with impressive quality made possible by the vast amount of data available on the Internet, raise social concerns that these models may generate harmful or copyrighted content. The biases and harmfulness arise throughout the entire training process and are hard to completely remove, which have become significant hurdles to the safe deployment of these models. In this paper, we propose a method called SDD to prevent problematic content generation in text-to-image diffusion models. We self-distill the diffusion model to guide the noise estimate conditioned on the target removal concept to match the unconditional one. Compared to the previous methods, our method eliminates a much greater proportion of harmful content from the generated images without degrading the overall image quality. Furthermore, our method allows the removal of multiple concepts at once, whereas previous works are limited to removing a single concept at a time.

Key facts. Cited by 41 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 29 of its 39 references that are in the map. First posted on arXiv 12 Jul 2023.

Builds on

29 of 39 references.

Closest ancestors · 4

Same direction, newest first

  1. 2023Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models30 Mar
  2. 2023Ablating Concepts in Text-to-Image Diffusion Models23 Mar
  3. 2023Erasing Concepts from Diffusion Models13 Mar
  4. 2022Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion Models9 Nov

Other references · 25

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Language Models are Few-Shot Learnerscited by 32,199
  3. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  4. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  5. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  6. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  7. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  8. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  9. 2022Classifier-Free Diffusion Guidancecited by 5,717
  10. 2016Overcoming catastrophic forgetting in neural networkscited by 4,718
  11. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,712
  12. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
13 more
  1. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  2. 2020Exploring Simple Siamese Representation Learningcited by 2,853
  3. 2022Prompt-to-Prompt Image Editing with Cross Attention Controlcited by 2,409
  4. 2021CLIPScore: A Reference-free Evaluation Metric for Image Captioningcited by 2,388
  5. 2022Multi-Concept Customization of Text-to-Image Diffusioncited by 1,099
  6. 2022Pseudo Numerical Methods for Diffusion Models on Manifoldscited by 695
  7. 2019Be Your Own Teacher: Improve the Performance of Convolutional Neural Networks via Self Distillationcited by 491
  8. 2021Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLPcited by 336
  9. 2022Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scalecited by 262
  10. 2023Stable Bias: Analyzing Societal Representations in Diffusion Modelscited by 225
  11. 2022Can Machines Help Us Answering Question 16 in Datasheets, and In Turn Reflecting on Inappropriate Content?cited by 150
  12. 2022A Prompt Array Keeps the Bias Away: Debiasing Vision-Language Models with Adversarial Learningcited by 104
  13. 2022Exploring the Limits of Domain-Adaptive Training for Detoxifying Large-Scale Language Modelscited by 67

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 3; 2024: 22; 2025: 10; 2026: 6.

41 papers.

Sort:

By area:

  1. 2023Unified Concept Editing in Diffusion Modelscited by 377
  2. 2024MACE: Mass Concept Erasure in Diffusion Modelscited by 298
  3. 2023Concept Sliders: LoRA Adaptors for Precise Control in Diffusion Modelscited by 148
  4. 2025Set You Straight: Auto-Steering Denoising Trajectories to Sidestep Unwanted Conceptscited by 61
  5. 2024R.A.C.E.: Robust Adversarial Concept Erasure for Secure Text-to-Image Diffusion Modelcited by 56
  6. 2024Margin-aware Preference Optimization for Aligning Diffusion Models without Referencecited by 45
  7. 2024Meta-Unlearning on Diffusion Models: Preventing Relearning Unlearned Conceptscited by 39
  8. 2024Adversarial Attacks and Defenses on Text-to-Image Diffusion Models: A Surveycited by 34
  9. 2023All but One: Surgical Concept Erasing with Model Preservation in Text-to-Image Diffusion Modelscited by 33
  10. 2024Jailbreak Attacks and Defenses against Multimodal Generative Models: A Surveycited by 31
  11. 2024Erasing Concepts from Text-to-Image Diffusion Models with Few-shot Unlearningcited by 29
  12. 2024T2IShield: Defending Against Backdoors on Text-to-Image Diffusion Modelscited by 29
29 more
  1. 2024Separable Multi-Concept Erasure from Diffusion Modelscited by 25
  2. 2025PromptGuard: Soft Prompt-Guided Unsafe Content Moderation for Text-to-Image Modelscited by 16
  3. 2024Editing Massive Concepts in Text-to-Image Diffusion Modelscited by 15
  4. 2024Espresso: Robust Concept Filtering in Text-to-Image Modelscited by 14
  5. 2024Prompt Sliders for Fine-Grained Control, Editing and Erasing of Concepts in Diffusion Modelscited by 14
  6. 2024Unstable Unlearning: The Hidden Risk of Concept Resurgence in Diffusion Modelscited by 14
  7. 2025Detect-and-Guide: Self-regulation of Diffusion Models for Safe Text-to-Image Generation via Guideline Token Optimizationcited by 12
  8. 2024Replication in Visual Diffusion Models: A Survey and Outlookcited by 11
  9. 2024Unveiling Concept Attribution in Diffusion Modelscited by 11
  10. 2024Towards Understanding Unsafe Video Generationcited by 10
  11. 2024Growth Inhibitors for Suppressing Inappropriate Image Concepts in Diffusion Modelscited by 10
  12. 2024Six-CD: Benchmarking Concept Removals for Benign Text-to-image Diffusion Modelscited by 9
  13. 2025Erased but Not Forgotten: How Backdoors Compromise Concept Erasurecited by 9
  14. 2024AdvAnchor: Enhancing Diffusion Model Unlearning with Adversarial Anchorscited by 7
  15. 2025TarPro: Targeted Protection against Malicious Image Editingcited by 7
  16. 2024BSPA: Exploring Black-box Stealthy Prompt Attacks against Image Generatorscited by 6
  17. 2024Safety Alignment Backfires: Preventing the Re-emergence of Suppressed Concepts in Fine-tuned Text-to-Image Diffusion Modelscited by 4
  18. 2025ACE: Attentional Concept Erasure in Diffusion Modelscited by 3
  19. 2026Closed-Form Concept Erasure via Double Projectionscited by 3
  20. 2025What Lurks Within? Concept Auditing for Shared Diffusion Models at Scalecited by 2
  21. 2025Multimodal Prompt Decoupling Attack on the Safety Filters in Text-to-Image Modelscited by 2
  22. 2025SecureT2I: No More Unauthorized Manipulation on AI Generated Images from Promptscited by 1
  23. 2025Rethinking Robust Adversarial Concept Erasure in Diffusion Modelscited by 1
  24. 2026Designing to Forget: Deep Semi-parametric Models for Unlearningcited by 1
  25. 2024Training-free Editioning of Text-to-Image Modelscited by 0
  26. 2026Compensation-free Machine Unlearning in Text-to-Image Diffusion Models by Eliminating the Mutual Informationcited by 0
  27. 2026CoreUnlearn: Rethinking Concept Unlearning through Disentangled Component-Level Erasure in Text-guided Diffusion Modelscited by 0
  28. 2026SACE: Concept Erasure at the Semantic Singularity in Visual Autoregressive Modelscited by 0
  29. 2026LU-500: A Logo Benchmark for Concept Unlearningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Degeneration-Tuning: Using Scrambled Grid shield Unwanted Concepts from Stable Diffusion
  2. 2024SAFREE: Training-Free and Adaptive Guard for Safe Text-to-Image And Video Generation
  3. 2026Test-Time Scaling for Safe Text-Guided Image Generation via Intermediate Clean Estimates
  4. 2024Rethinking Training for De-biasing Text-to-Image Generation: Unlocking the Potential of Stable Diffusion
  5. 2025Training-Free Safe Text Embedding Guidance for Text-to-Image Diffusion Models
  6. 2026Erasing Thousands of Concepts: Towards Scalable and Practical Concept Erasure for Text-to-Image Diffusion Models
  7. 2024Controllable Generation with Text-to-Image Diffusion Models: A Survey
  8. 2025A Comprehensive Survey on Concept Erasure in Text-to-Image Diffusion Models