paper-with-me

홈 › Papers

Removing Undesirable Concepts in Text-to-Image Diffusion Models with Learnable Prompts

2024-03-18 · Anh Bui, Khanh Doan, Trung Le, Paul Montague, Tamas Abraham, Dinh Phung

Diffusion models have shown remarkable capability in generating visually impressive content from textual descriptions. However, these models are trained on vast internet data, much of which contains undesirable elements such as sensitive content, copyrighted material, and unethical or harmful concepts. Therefore, beyond generating high-quality content, it is crucial to ensure these models do not propagate these undesirable elements. To address this issue, we propose a novel method to remove undesirable concepts from text-to-image diffusion models by incorporating a learnable prompt into the cross-attention module. This learnable prompt acts as additional memory, capturing the knowledge of undesirable concepts and reducing their dependency on the model parameters and corresponding textual inputs. By transferring this knowledge to the prompt, erasing undesirable concepts becomes more stable and has minimal negative impact on other concepts. We demonstrate the effectiveness of our method on the Stable Diffusion model, showcasing its superiority over state-of-the-art erasure methods in removing undesirable content while preserving unrelated elements.

📄 PDF Abstract BibTeX arXiv:2403.12326

Code (2)

tuananhbui89/adaptive-guided-erasure pytorch
tuananhbui89/erasing-adversarial-preservation pytorch

Tasks

Transfer Learning

Methods 이 논문이 사용한 방법론

Diffusion Diffusion models generate samples by gradually removing noise from a signal, and their training objective can be expressed as a reweighted variational lower-bound…

Similar Papers 제목 키워드 기반

Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation

2024-10-21 · Anh Bui, Long Vuong, Khanh Doan, Trung Le 외

Diffusion models excel at generating visually striking content from text but can inadvertently produce undesirable or harmful content when trained on unfiltered internet data. A practical solution is to selectively remov…

EraseAnything: Enabling Concept Erasure in Rectified Flow Transformers

2024-12-29 · Daiheng Gao, Shilin Lu, Shaw Walters, Wenbo Zhou 외

Removing unwanted concepts from large-scale text-to-image (T2I) diffusion models while maintaining their overall generative quality remains an open challenge. This difficulty is especially pronounced in emerging paradigm…

Contrastive Learning

Erasure or Erosion? Evaluating Compositional Degradation in Unlearned Text-To-Image Diffusion Models

2026-04-06 · Arian Komaei Koma, Seyed Amir Kasaei, Ali Aghayari, AmirMahdi Sadeghzadeh 외 arxiv

Post-hoc unlearning has emerged as a practical mechanism for removing undesirable concepts from large text-to-image diffusion models. However, prior work primarily evaluates unlearning through erasure success; its impact…

Text-to-Image GenerationSpatial Reasoning

AdvAnchor: Enhancing Diffusion Model Unlearning with Adversarial Anchors

2024-12-28 · Mengnan Zhao, Lihe Zhang, Xingyi Yang, Tianhang Zheng 외

Security concerns surrounding text-to-image diffusion models have driven researchers to unlearn inappropriate concepts through fine-tuning. Recent fine-tuning methods typically align the prediction distributions of unsaf…

model

Concept Removal for Frontier Image Generative Models

2026-06-24 · Aditya Kumar, Pierre Joly, Adam Dziedzic, Franziska Boenisch arxiv

Image generative models are trained on massive, largely uncurated internet-scale datasets that contain undesirable visual concepts. Efficiently removing such concepts from the model generations without degrading the qual…