Research map / Image generation / Text-to-image diffusion
Continual Unlearning for Text-to-Image Diffusion Models: A Regularization Perspective
Justin Lee, Zheda Mai, Jinsu Yoo, Chongyu Fan, Cheng Zhang, Wei-Lun Chao
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 |
|---|---|---|---|
| unlearn concepts sequentially instead of one request | ConAbl and SculpMem · UnlearnCanvas continual style/object unlearning | RA-I and RA-C | Hurt |
| unlearn all requests simultaneously instead of sequentially | ConAbl · UnlearnCanvas continual unlearning | retention accuracy | Improved |
| add L1/L2 update-norm regularization (ablation) | ConAbl sequential unlearning · UnlearnCanvas | RA-I, RA-C | Improved |
| add selective fine-tuning (top-k% parameters) (ablation) | ConAbl sequential unlearning · UnlearnCanvas | RA-C | Improved |
| add TIES model merging of per-concept unlearned models (ablation) | ConAbl sequential unlearning · UnlearnCanvas | retention accuracy | Improved |
| add semantic-aware gradient projection (ablation) | ConAbl sequential unlearning · UnlearnCanvas | RA-I | Improved |
| add gradient projection (vs SelFT and merging) on object unlearning (ablation) | ConAbl sequential unlearning · UnlearnCanvas object unlearning | RA-C | Hurt |
| combine gradient projection with other add-on regularizers (ablation) | ConAbl sequential unlearning · UnlearnCanvas | harmonic mean of UA, RA-I, RA-C | Improved |
| match optimization steps between sequential and simultaneous unlearning (ablation) | ConAbl · UnlearnCanvas style unlearning | parameter drift: 2,100 each after six concepts | No clear effect |
Closest papers by meaning
- Boosting Alignment for Post-Unlearning Text-to-Image Generative Models (NeurIPS 2024 · similarity 0.93)
- Fine-Grained Erasure in Text-to-Image Diffusion-based Foundation Models (CVPR 2025 · similarity 0.92)
- The Illusion of Unlearning: The Unstable Nature of Machine Unlearning in Text-to-Image Diffusion Models (CVPR 2025 · similarity 0.91)
- Unlearning Concepts in Diffusion Model via Concept Domain Correction and Concept Preserving Gradient (AAAI 2025 · similarity 0.91)
- Enabling Preference-driven Unlearning in Few-step Distilled Text-to-Image Diffusion Models (NeurIPS 2026 · similarity 0.91)
- Machine Unlearning for Image-to-Image Generative Models (ICLR 2024 · similarity 0.90)
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.
