Localizing and Editing Knowledge in Text-to-Image Generative Models
Text-to-Image Diffusion Models such as Stable-Diffusion and Imagen have achieved unprecedented quality of photorealism with state-of-the-art FID scores on MS-COCO and other generation benchmarks. Given a caption, image generation requires fine-grained knowledge about attributes such as object structure, style, and viewpoint amongst others. Where does this information reside in text-to-image generative models? In our paper, we tackle this question and understand how knowledge corresponding to distinct visual attributes is stored in large-scale text-to-image diffusion models. We adapt Causal Mediation Analysis for text-to-image models and trace knowledge about distinct visual attributes to various (causal) components in the (i) UNet and (ii) text-encoder of the diffusion model. In particular, we show that unlike generative large-language models, knowledge about different attributes is not localized in isolated components, but is instead distributed amongst a set of components in the conditional UNet. These sets of components are often distinct for different visual attributes. Remarkably, we find that the CLIP text-encoder in public text-to-image models such as Stable-Diffusion contains only one causal state across different visual attributes, and this is the first self-attention layer corresponding to the last subject token of the attribute in the caption. This is in stark contrast to the causal states in other language models which are often the mid-MLP layers. Based on this observation of only one causal state in the text-encoder, we introduce a fast, data-free model editing method Diff-QuickFix which can effectively edit concepts in text-to-image models. DiffQuickFix can edit (ablate) concepts in under a second with a closed-form update, providing a significant 1000x speedup and comparable editing performance to existing fine-tuning based editing methods.
Code (0)
등록된 구현이 없습니다.
Tasks
AttributeImage GenerationModel EditingMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
X-Edit: Detecting and Localizing Edits in Images Altered by Text-Guided Diffusion Models
Text-guided diffusion models have significantly advanced image editing, enabling highly realistic and local modifications based on textual prompts. While these developments expand creative possibilities, their malicious …
Face SwappingSSIMLocalizing and Editing Knowledge in Large Audio-Language Models
Large Audio-Language Models (LALMs) have shown strong performance in speech understanding, making speech a natural interface for accessing factual information. Yet they are trained on static corpora and may encode incorr…
Rethinking Structure Preservation in Text-Guided Image Editing with Visual Autoregressive Models
Visual autoregressive (VAR) models have recently emerged as a promising family of generative models, enabling a wide range of downstream vision tasks such as text-guided image editing. By shifting the editing paradigm fr…
Reinforcement LearningImage EditingEditMGT: Unleashing Potentials of Masked Generative Transformers in Image Editing
Recent advances in diffusion models (DMs) have achieved exceptional visual quality in image editing tasks. However, the global denoising dynamics of DMs inherently conflate local editing targets with the full-image conte…
Style TransferImage EditingDirectDrag: High-Fidelity, Mask-Free, Prompt-Free Drag-based Image Editing via Readout-Guided Feature Alignment
Drag-based image editing using generative models provides intuitive control over image structures. However, existing methods rely heavily on manually provided masks and textual prompts to preserve semantic fidelity and m…
Image ManipulationImage Editing