{"id":"arxiv:2503.14232","kind":"preprint","title":"CRCE: Coreference-Retention Concept Erasure in Text-to-Image Diffusion Models","authors":["Yuyang Xue","Edward Moroshko","Feng Chen","Jingyu Sun","Steven McDonagh","Sotirios A. Tsaftaris"],"author_count":6,"year":2025,"published":"2025-03-18","venue":null,"cited_by":null,"doi":null,"arxiv_id":"2503.14232","nct_id":null,"source_url":"https://arxiv.org/abs/2503.14232","archive_url":"https://space-kitty.com/papers/arxiv:2503.14232","updated":"2025-05-20","pmid":null,"pmcid":null,"openalex_id":null,"pdf_url":"https://arxiv.org/pdf/2503.14232","abstract":"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.","abstract_license":"CC0-1.0 (arXiv descriptive metadata)","license":null,"is_open_access":false,"retracted":false,"peer_reviewed":false,"categories":["cs.CV","cs.AI","cs.CL","cs.LG"],"details":null,"provenance":"arxiv","attribution":"Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.","topics":[],"cited_in_dossiers":[]}