Localizing and Editing Knowledge in Text-to-Image Generative Models

• URL: http://arxiv.org/abs/2310.13730v1
• Date: Fri, 20 Oct 2023 17:31:12 GMT
• Title: Localizing and Editing Knowledge in Text-to-Image Generative Models
• Authors: Samyadeep Basu, Nanxuan Zhao, Vlad Morariu, Soheil Feizi, Varun Manjunatha
• Abstract summary: knowledge about different attributes is not localized in isolated components, but is instead distributed amongst a set of components in the conditional UNet. We introduce a fast, data-free model editing method Diff-QuickFix which can effectively edit concepts in text-to-image models.
• Score: 62.02776252311559
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 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.

Related papers

• On Mechanistic Knowledge Localization in Text-to-Image Generative Models [44.208804082687294]
  We introduce the concept of Mechanistic Localization in text-to-image models. We measure the direct effect of intermediate layers to output generation by performing interventions in the cross-attention layers of the UNet. We employ LocoEdit, a fast closed-form editing method across popular open-source text-to-image models.
  arXiv  Detail & Related papers  (2024-05-02T05:19:05Z)
• Object-Attribute Binding in Text-to-Image Generation: Evaluation and Control [58.37323932401379]
  Current diffusion models create images given a text prompt as input but struggle to correctly bind attributes mentioned in the text to the right objects in the image. We propose focused cross-attention (FCA) that controls the visual attention maps by syntactic constraints found in the input sentence. We show substantial improvements in T2I generation and especially its attribute-object binding on several datasets.
  arXiv  Detail & Related papers  (2024-04-21T20:26:46Z)
• DragonDiffusion: Enabling Drag-style Manipulation on Diffusion Models [66.43179841884098]
  We propose a novel image editing method, DragonDiffusion, enabling Drag-style manipulation on Diffusion models. Our method achieves various editing modes for the generated or real images, such as object moving, object resizing, object appearance replacement, and content dragging.
  arXiv  Detail & Related papers  (2023-07-05T16:43:56Z)
• DiffUTE: Universal Text Editing Diffusion Model [32.384236053455]
  We propose a universal self-supervised text editing diffusion model (DiffUTE) It aims to replace or modify words in the source image with another one while maintaining its realistic appearance. Our method achieves an impressive performance and enables controllable editing on in-the-wild images with high fidelity.
  arXiv  Detail & Related papers  (2023-05-18T09:06:01Z)
• PAIR-Diffusion: A Comprehensive Multimodal Object-Level Image Editor [135.17302411419834]
  PAIR Diffusion is a generic framework that enables a diffusion model to control the structure and appearance of each object in the image. We show that having control over the properties of each object in an image leads to comprehensive editing capabilities. Our framework allows for various object-level editing operations on real images such as reference image-based appearance editing, free-form shape editing, adding objects, and variations.
  arXiv  Detail & Related papers  (2023-03-30T17:13:56Z)
• Editing Implicit Assumptions in Text-to-Image Diffusion Models [48.542005079915896]
  Text-to-image diffusion models often make implicit assumptions about the world when generating images. In this work, we aim to edit a given implicit assumption in a pre-trained diffusion model. Our method is highly efficient, as it modifies a mere 2.2% of the model's parameters in under one second.
  arXiv  Detail & Related papers  (2023-03-14T17:14:21Z)
• PRedItOR: Text Guided Image Editing with Diffusion Prior [2.3022070933226217]
  Text guided image editing requires compute intensive optimization of text embeddings or fine-tuning the model weights for text guided image editing. Our architecture consists of a diffusion prior model that generates CLIP image embedding conditioned on a text prompt and a custom Latent Diffusion Model trained to generate images conditioned on CLIP image embedding. We combine this with structure preserving edits on the image decoder using existing approaches such as reverse DDIM to perform text guided image editing.
  arXiv  Detail & Related papers  (2023-02-15T22:58:11Z)
• ManiCLIP: Multi-Attribute Face Manipulation from Text [104.30600573306991]
  We present a novel multi-attribute face manipulation method based on textual descriptions. Our method generates natural manipulated faces with minimal text-irrelevant attribute editing.
  arXiv  Detail & Related papers  (2022-10-02T07:22:55Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.