Research map / Image generation / Text-to-image diffusion
Co-occurring Associated REtained concepts in Diffusion Unlearning
Miso Kim, Georu Lee, Yunji Kim, Hoki Kim, Jinseong Park, Woojin Lee
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add CARE-set retain/erase losses (ReCARE) to STEREO-style unlearning | STEREO · Nudity unlearning | CARE score: 0.11 -> 0.94 | Improved |
| add CARE-set retain/erase losses (ReCARE) to STEREO-style unlearning | STEREO · Nudity unlearning (CCE attack) | ASR (%): 19.55 -> 11.14 | Improved |
| add global clustering to CARE-set refinement (ablation) | ReCARE · Nudity unlearning (CCE attack) | ASR (%): 25.00 -> 11.14 | Improved |
| add intra-cluster refinement (ablation) | ReCARE · Nudity unlearning (CCE attack) | ASR (%): 16.36 -> 11.14 | Improved |
| add both refinement stages to CARE-set (ablation) | ReCARE · Nudity unlearning | CARE score: 0.88 -> 0.94 | Improved |
| vary number of k-means clusters n from 4 to 6 (ablation) | ReCARE · Nudity unlearning (CCE attack) | ASR (%): 15.36 -> 11.14 | No clear effect |
| vary pruning strictness alpha (0.005 to 0.015) (ablation) | ReCARE · Nudity unlearning | CARE score: 0.94 -> 0.90 | No clear effect |
| vary Top-K tokens per image (30 to 70) (ablation) | ReCARE · Nudity unlearning | CARE score: 0.97 / 0.94 / 0.96 | No clear effect |
| increase retain weight lambda from 1.0 to 1.5 (ablation) | ReCARE · Nudity unlearning (CCE attack) | ASR (%): 11.14 -> 30.91 | Hurt |
| decrease retain weight lambda from 1.0 to 0.5 (ablation) | ReCARE · Nudity unlearning | FID: 13.85 -> 15.50 | Hurt |
| replace CLIP with SigLIP in CARE score evaluation | CARE score · Nudity unlearning | method ranking | No clear effect |
| use manual anchor 'figure' instead of 'person' for STEREO | STEREO · Nudity unlearning | human detection rate: 0.64 -> 0.92 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SalUn — SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in Both Image Classification and Generation (ICLR 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- The Illusion of Unlearning: The Unstable Nature of Machine Unlearning in Text-to-Image Diffusion Models (CVPR 2025 · similarity 0.88)
- You Can’t Have It Both Ways: Concept Entanglement Limits Diffusion Model Unlearning (NeurIPS 2026 · similarity 0.88)
- Unlearning Concepts in Diffusion Model via Concept Domain Correction and Concept Preserving Gradient (AAAI 2025 · similarity 0.88)
- CURE: Concept Unlearning via Orthogonal Representation Editing in Diffusion Models (NeurIPS 2025 · similarity 0.88)
- Enabling Preference-driven Unlearning in Few-step Distilled Text-to-Image Diffusion Models (NeurIPS 2026 · similarity 0.86)
- Growth Inhibitors for Suppressing Inappropriate Image Concepts in Diffusion Models (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
