Towards Multimodal Understanding via Stable Diffusion as a Task-Aware Feature Extractor

• URL: http://arxiv.org/abs/2507.07106v1
• Date: Wed, 09 Jul 2025 17:59:47 GMT
• Title: Towards Multimodal Understanding via Stable Diffusion as a Task-Aware Feature Extractor
• Authors: Vatsal Agarwal, Matthew Gwilliam, Gefen Kohavi, Eshan Verma, Daniel Ulbricht, Abhinav Shrivastava,
• Abstract summary: We study whether pre-trained text-to-image diffusion models can serve as instruction-aware visual encoders.<n>We find diffusion features are both rich in semantics and can encode strong image-text alignment.<n>We then investigate how to align these features with large language models and uncover a leakage phenomenon.
• Score: 32.34399128209528
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent advances in multimodal large language models (MLLMs) have enabled image-based question-answering capabilities. However, a key limitation is the use of CLIP as the visual encoder; while it can capture coarse global information, it often can miss fine-grained details that are relevant to the input query. To address these shortcomings, this work studies whether pre-trained text-to-image diffusion models can serve as instruction-aware visual encoders. Through an analysis of their internal representations, we find diffusion features are both rich in semantics and can encode strong image-text alignment. Moreover, we find that we can leverage text conditioning to focus the model on regions relevant to the input question. We then investigate how to align these features with large language models and uncover a leakage phenomenon, where the LLM can inadvertently recover information from the original diffusion prompt. We analyze the causes of this leakage and propose a mitigation strategy. Based on these insights, we explore a simple fusion strategy that utilizes both CLIP and conditional diffusion features. We evaluate our approach on both general VQA and specialized MLLM benchmarks, demonstrating the promise of diffusion models for visual understanding, particularly in vision-centric tasks that require spatial and compositional reasoning. Our project page can be found https://vatsalag99.github.io/mustafar/.

Related papers

• Unlocking the Capabilities of Large Vision-Language Models for Generalizable and Explainable Deepfake Detection [18.125287697902813]
  Current Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities in understanding multimodal data.<n>We present a novel framework that unlocks LVLMs' potential capabilities for deepfake detection.
  arXiv  Detail & Related papers  (2025-03-19T03:20:03Z)
• Seeing is Understanding: Unlocking Causal Attention into Modality-Mutual Attention for Multimodal LLMs [7.03771340666549]
  Vision-language misalignment in Multimodal Large Language Models (MLLMs) is a critical challenge.<n>We propose MapleLeaf AKI, a novel MLLM that unlocks causal attention into modality-mutual attention (MMA) to enable image tokens to attend to text tokens.<n>Our MMA design is intended to be generic, allowing for application across various modalities, and scalable to accommodate diverse multimodal scenarios.
  arXiv  Detail & Related papers  (2025-03-04T13:18:33Z)
• Massive Values in Self-Attention Modules are the Key to Contextual Knowledge Understanding [58.364933651703524]
  We show that concentrated massive values consistently emerge in specific regions of attention queries.<n>These massive values play a critical role in interpreting contextual knowledge.<n>We trace the emergence of massive values and find that such concentration is caused by Rotary Positional.
  arXiv  Detail & Related papers  (2025-02-03T17:47:03Z)
• Performance Gap in Entity Knowledge Extraction Across Modalities in Vision Language Models [36.18155629835474]
  Vision-language models (VLMs) excel at extracting and reasoning about information from images.<n>Yet, their capacity to leverage internal knowledge about specific entities remains underexplored.<n>This work investigates the disparity in model performance when answering factual questions about an entity described in text versus depicted in an image.
  arXiv  Detail & Related papers  (2024-12-18T18:22:30Z)
• Seeing Syntax: Uncovering Syntactic Learning Limitations in Vision-Language Models [18.87130615326443]
  Vision-language models (VLMs) serve as foundation models for image captioning and text-to-image generation.<n>Recent studies have highlighted limitations in VLM text encoders, particularly in areas like compositionality and semantic understanding.
  arXiv  Detail & Related papers  (2024-12-11T05:37:04Z)
• Rethinking Visual Prompting for Multimodal Large Language Models with External Knowledge [76.45868419402265]
  multimodal large language models (MLLMs) have made significant strides by training on vast high-quality image-text datasets. However, the inherent difficulty in explicitly conveying fine-grained or spatially dense information in text, such as masks, poses a challenge for MLLMs. This paper proposes a new visual prompt approach to integrate fine-grained external knowledge, gleaned from specialized vision models, into MLLMs.
  arXiv  Detail & Related papers  (2024-07-05T17:43:30Z)
• Understanding Information Storage and Transfer in Multi-modal Large Language Models [51.20840103605018]
  We study how Multi-modal Large Language Models process information in a factual visual question answering task. Key findings show that these MLLMs rely on self-attention blocks in much earlier layers for information storage. We introduce MultEdit, a model-editing algorithm that can correct errors and insert new long-tailed information into MLLMs.
  arXiv  Detail & Related papers  (2024-06-06T16:35:36Z)
• From Redundancy to Relevance: Information Flow in LVLMs Across Reasoning Tasks [33.476693301050275]
  We conduct experiments with truncation strategies across various LVLMs for visual question answering and image captioning tasks. By exploring the information flow from the perspective of visual representation contribution, we observe that it tends to converge in shallow layers but diversify in deeper layers.
  arXiv  Detail & Related papers  (2024-06-04T13:52:54Z)
• Chain-of-Spot: Interactive Reasoning Improves Large Vision-Language Models [81.71651422951074]
  Chain-of-Spot (CoS) method is a novel approach that enhances feature extraction by focusing on key regions of interest. This technique allows LVLMs to access more detailed visual information without altering the original image resolution. Our empirical findings demonstrate a significant improvement in LVLMs' ability to understand and reason about visual content.
  arXiv  Detail & Related papers  (2024-03-19T17:59:52Z)
• On the Multi-modal Vulnerability of Diffusion Models [56.08923332178462]
  We propose MMP-Attack to manipulate the generation results of diffusion models by appending a specific suffix to the original prompt.<n>Our goal is to induce diffusion models to generate a specific object while simultaneously eliminating the original object.
  arXiv  Detail & Related papers  (2024-02-02T12:39:49Z)
• MouSi: Poly-Visual-Expert Vision-Language Models [132.58949014605477]
  This paper proposes the use of ensemble experts technique to synergize the capabilities of individual visual encoders. This technique introduces a fusion network to unify the processing of outputs from different visual experts. In our implementation, this technique significantly reduces the positional occupancy in models like SAM, from a substantial 4096 to a more efficient and manageable 64 or even down to 1.
  arXiv  Detail & Related papers  (2024-01-30T18:09:11Z)

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.