$\textit{Revelio}$: Interpreting and leveraging semantic information in diffusion models

• URL: http://arxiv.org/abs/2411.16725v1
• Date: Sat, 23 Nov 2024 03:54:22 GMT
• Title: $\textit{Revelio}$: Interpreting and leveraging semantic information in diffusion models
• Authors: Dahye Kim, Xavier Thomas, Deepti Ghadiyaram,
• Abstract summary: We study $textithow$ rich visual semantic information is represented within various layers and denoising timesteps of different diffusion architectures. We uncover monosemantic interpretable features by leveraging k-sparse autoencoders (k-SAE) We substantiate our mechanistic interpretations via transfer learning using light-weight classifiers on off-the-shelf diffusion models' features.
• Score: 10.68914376295842
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We study $\textit{how}$ rich visual semantic information is represented within various layers and denoising timesteps of different diffusion architectures. We uncover monosemantic interpretable features by leveraging k-sparse autoencoders (k-SAE). We substantiate our mechanistic interpretations via transfer learning using light-weight classifiers on off-the-shelf diffusion models' features. On $4$ datasets, we demonstrate the effectiveness of diffusion features for representation learning. We provide in-depth analysis of how different diffusion architectures, pre-training datasets, and language model conditioning impacts visual representation granularity, inductive biases, and transfer learning capabilities. Our work is a critical step towards deepening interpretability of black-box diffusion models. Code and visualizations available at: https://github.com/revelio-diffusion/revelio

Related papers

• Expanding the Role of Diffusion Models for Robust Classifier Training [14.409018677904571]
  We show that diffusion models offer representations that are both diverse and partially robust.<n>Our representation analysis indicates that incorporating diffusion models into adversarial training encourages more disentangled features.<n>Experiments on CIFAR-10, CIFAR-100, and ImageNet validate these findings.
  arXiv  Detail & Related papers  (2026-02-23T15:06:52Z)
• On the Role of Discreteness in Diffusion LLMs [69.64854287505999]
  We revisit the view of diffusion process and language modeling, and outline five properties that separate diffusion mechanics from language-specific requirements.<n>We identify two central issues: (i) uniform corruption does not respect how information is distributed across positions, and (ii) token-wise marginal training cannot capture multi-token dependencies during parallel decoding.<n>These observations motivate diffusion processes that align more closely with the structure of text, and encourage future work toward more coherent diffusion language models.
  arXiv  Detail & Related papers  (2025-12-27T16:03:08Z)
• Graph Representation Learning with Diffusion Generative Models [0.0]
  We train a discrete diffusion model within an autoencoder framework to learn meaningful embeddings for graph data.<n>We extract the representation from the combination of the encoder's output and the decoder's first time step hidden embedding.<n>Our approach demonstrates the potential of discrete diffusion models to be used for graph representation learning.
  arXiv  Detail & Related papers  (2025-01-22T07:12:10Z)
• Decoding Diffusion: A Scalable Framework for Unsupervised Analysis of Latent Space Biases and Representations Using Natural Language Prompts [68.48103545146127]
  This paper proposes a novel framework for unsupervised exploration of diffusion latent spaces. We directly leverage natural language prompts and image captions to map latent directions. Our method provides a more scalable and interpretable understanding of the semantic knowledge encoded within diffusion models.
  arXiv  Detail & Related papers  (2024-10-25T21:44:51Z)
• Suppress Content Shift: Better Diffusion Features via Off-the-Shelf Generation Techniques [119.02857688205295]
  We discover that diffusion feature has been hindered by a hidden yet universal phenomenon that we call content shift. We locate the cause of content shift as one inherent characteristic of diffusion models, which suggests the broad existence of this phenomenon in diffusion feature. We propose a practical guideline named GATE to efficiently evaluate the potential benefit of a technique and provide an implementation of our methodology.
  arXiv  Detail & Related papers  (2024-10-09T09:43:36Z)
• How Diffusion Models Learn to Factorize and Compose [14.161975556325796]
  Diffusion models are capable of generating photo-realistic images that combine elements which likely do not appear together in the training set. We investigate whether and when diffusion models learn semantically meaningful and factorized representations of composable features.
  arXiv  Detail & Related papers  (2024-08-23T17:59:03Z)
• Diffusion Models and Representation Learning: A Survey [3.8861148837000856]
  This survey explores the interplay between diffusion models and representation learning. It provides an overview of diffusion models' essential aspects, including mathematical foundations. Various approaches related to diffusion models and representation learning are detailed.
  arXiv  Detail & Related papers  (2024-06-30T17:59:58Z)
• Training Class-Imbalanced Diffusion Model Via Overlap Optimization [55.96820607533968]
  Diffusion models trained on real-world datasets often yield inferior fidelity for tail classes. Deep generative models, including diffusion models, are biased towards classes with abundant training images. We propose a method based on contrastive learning to minimize the overlap between distributions of synthetic images for different classes.
  arXiv  Detail & Related papers  (2024-02-16T16:47:21Z)
• Guided Diffusion from Self-Supervised Diffusion Features [49.78673164423208]
  Guidance serves as a key concept in diffusion models, yet its effectiveness is often limited by the need for extra data annotation or pretraining. We propose a framework to extract guidance from, and specifically for, diffusion models.
  arXiv  Detail & Related papers  (2023-12-14T11:19:11Z)
• Do text-free diffusion models learn discriminative visual representations? [39.78043004824034]
  We explore the possibility of a unified representation learner: a model which addresses both families of tasks simultaneously. We develop diffusion models, a state-of-the-art method for generative tasks, as a prime candidate. We find that diffusion models are better than GANs, and, with our fusion and feedback mechanisms, can compete with state-of-the-art unsupervised image representation learning methods for discriminative tasks.
  arXiv  Detail & Related papers  (2023-11-29T18:59:59Z)
• DiffDis: Empowering Generative Diffusion Model with Cross-Modal Discrimination Capability [75.9781362556431]
  We propose DiffDis to unify the cross-modal generative and discriminative pretraining into one single framework under the diffusion process. We show that DiffDis outperforms single-task models on both the image generation and the image-text discriminative tasks.
  arXiv  Detail & Related papers  (2023-08-18T05:03:48Z)
• A Survey of Diffusion Models in Natural Language Processing [11.233768932957771]
  Diffusion models capture the diffusion of information or signals across a network or manifold. This paper discusses the different formulations of diffusion models used in NLP, their strengths and limitations, and their applications.
  arXiv  Detail & Related papers  (2023-05-24T03:25:32Z)
• Discffusion: Discriminative Diffusion Models as Few-shot Vision and Language Learners [88.07317175639226]
  We propose a novel approach, Discriminative Stable Diffusion (DSD), which turns pre-trained text-to-image diffusion models into few-shot discriminative learners. Our approach mainly uses the cross-attention score of a Stable Diffusion model to capture the mutual influence between visual and textual information.
  arXiv  Detail & Related papers  (2023-05-18T05:41:36Z)
• Unleashing Text-to-Image Diffusion Models for Visual Perception [84.41514649568094]
  VPD (Visual Perception with a pre-trained diffusion model) is a new framework that exploits the semantic information of a pre-trained text-to-image diffusion model in visual perception tasks. We show that VPD can be faster adapted to downstream visual perception tasks using the proposed VPD.
  arXiv  Detail & Related papers  (2023-03-03T18:59:47Z)

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.