CRCE: Coreference-Retention Concept Erasure in Text-to-Image Diffusion Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866918026236919808 |
|---|---|
| author | Xue, Yuyang Moroshko, Edward Chen, Feng Sun, Jingyu McDonagh, Steven Tsaftaris, Sotirios A. |
| author_facet | Xue, Yuyang Moroshko, Edward Chen, Feng Sun, Jingyu McDonagh, Steven Tsaftaris, Sotirios A. |
| contents | Text-to-Image diffusion models can produce undesirable content that necessitates concept erasure. However, existing methods struggle with under-erasure, leaving residual traces of targeted concepts, or over-erasure, mistakenly eliminating unrelated but visually similar concepts. To address these limitations, we introduce CRCE, a novel concept erasure framework that leverages Large Language Models to identify both semantically related concepts that should be erased alongside the target and distinct concepts that should be preserved. By explicitly modelling coreferential and retained concepts semantically, CRCE enables more precise concept removal, without unintended erasure. Experiments demonstrate that CRCE outperforms existing methods on diverse erasure tasks, including real-world object, person identities, and abstract intellectual property characteristics. The constructed dataset CorefConcept and the source code will be release upon acceptance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_14232 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | CRCE: Coreference-Retention Concept Erasure in Text-to-Image Diffusion Models Xue, Yuyang Moroshko, Edward Chen, Feng Sun, Jingyu McDonagh, Steven Tsaftaris, Sotirios A. Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Machine Learning Text-to-Image diffusion models can produce undesirable content that necessitates concept erasure. However, existing methods struggle with under-erasure, leaving residual traces of targeted concepts, or over-erasure, mistakenly eliminating unrelated but visually similar concepts. To address these limitations, we introduce CRCE, a novel concept erasure framework that leverages Large Language Models to identify both semantically related concepts that should be erased alongside the target and distinct concepts that should be preserved. By explicitly modelling coreferential and retained concepts semantically, CRCE enables more precise concept removal, without unintended erasure. Experiments demonstrate that CRCE outperforms existing methods on diverse erasure tasks, including real-world object, person identities, and abstract intellectual property characteristics. The constructed dataset CorefConcept and the source code will be release upon acceptance. |
| title | CRCE: Coreference-Retention Concept Erasure in Text-to-Image Diffusion Models |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2503.14232 |