DAVE: A VLM Vision Encoder for Document Understanding and Web Agents

• URL: http://arxiv.org/abs/2512.17221v2
• Date: Thu, 25 Dec 2025 02:25:27 GMT
• Title: DAVE: A VLM Vision Encoder for Document Understanding and Web Agents
• Authors: Brandon Huang, Hang Hua, Zhuoran Yu, Trevor Darrell, Rogerio Feris, Roei Herzig,
• Abstract summary: We introduce DAVE, a vision encoder purpose-built for Vision-language models (VLMs)<n>Our training pipeline is designed to leverage abundant unlabeled data to bypass the need for costly large-scale annotations for document and web images.<n>We use ensemble training to fuse features from pretrained generalist encoders with our own document and web-specific representations.
• Score: 50.05119785399764
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: While Vision-language models (VLMs) have demonstrated remarkable performance across multi-modal tasks, their choice of vision encoders presents a fundamental weakness: their low-level features lack the robust structural and spatial information essential for document understanding and web agents. To bridge this gap, we introduce DAVE, a vision encoder purpose-built for VLMs and tailored for these tasks. Our training pipeline is designed to leverage abundant unlabeled data to bypass the need for costly large-scale annotations for document and web images. We begin with a self-supervised pretraining stage on unlabeled images, followed by a supervised autoregressive pretraining stage, where the model learns tasks like parsing and localization from limited, high-quality data. Within the supervised stage, we adopt two strategies to improve our encoder's alignment with both general visual knowledge and diverse document and web agentic tasks: (i) We introduce a novel model-merging scheme, combining encoders trained with different text decoders to ensure broad compatibility with different web agentic architectures. (ii) We use ensemble training to fuse features from pretrained generalist encoders (e.g., SigLIP2) with our own document and web-specific representations. Extensive experiments on classic document tasks, VQAs, web localization, and agent-based benchmarks validate the effectiveness of our approach, establishing DAVE as a strong vision encoder for document and web applications.

Related papers

• Perception Encoder: The best visual embeddings are not at the output of the network [70.86738083862099]
  We introduce Perception (PE), a vision encoder for image and video understanding trained via simple vision-language learning.<n>We find that contrastive vision-language training alone can produce strong, general embeddings for all of these downstream tasks.<n>Together, our PE family of models achieves best-in-class results on a wide variety of tasks.
  arXiv  Detail & Related papers  (2025-04-17T17:59:57Z)
• QID: Efficient Query-Informed ViTs in Data-Scarce Regimes for OCR-free Visual Document Understanding [53.69841526266547]
  Fine-tuning a pre-trained Vision-Language Model with new datasets often falls short in optimizing the vision encoder.<n>We introduce QID, a novel, streamlined, architecture-preserving approach that integrates query embeddings into the vision encoder.
  arXiv  Detail & Related papers  (2025-04-03T18:47:16Z)
• EVEv2: Improved Baselines for Encoder-Free Vision-Language Models [72.07868838411474]
  Existing encoder-free vision-language models (VLMs) are narrowing the performance gap with their encoder-based counterparts.<n>We develop efficient strategies for encoder-free VLMs that rival mainstream encoder-based ones.<n>We show that properly and hierarchically associating vision and language within a unified model reduces interference between modalities.
  arXiv  Detail & Related papers  (2025-02-10T18:59:58Z)
• Response Wide Shut: Surprising Observations in Basic Vision Language Model Capabilities [30.176918208200604]
  Vision-Language Models (VLMs) have emerged as general purpose tools for addressing a variety of complex computer vision problems. These models have been shown to be highly capable, but also lacking some basic visual understanding skills. This paper sets out to understand the limitations of SoTA VLMs on fundamental visual tasks.
  arXiv  Detail & Related papers  (2024-08-13T08:26:32Z)
• HRVDA: High-Resolution Visual Document Assistant [32.51417315241559]
  We propose a High-Resolution Visual Document Assistant (HRVDA) to bridge the gap between MLLMs and visual document understanding. HRVDA employs a content filtering mechanism and an instruction filtering module to filter out the content-agnostic visual tokens and instruction-agnostic visual tokens. Our model achieves state-of-the-art performance across multiple document understanding datasets.
  arXiv  Detail & Related papers  (2024-04-10T11:10:50Z)
• Enhancing Visual Document Understanding with Contrastive Learning in Large Visual-Language Models [56.76307866160105]
  We propose a contrastive learning framework, termed Document Object COntrastive learning (DoCo) DoCo leverages an auxiliary multimodal encoder to obtain the features of document objects and align them to the visual features generated by the vision encoder of Large Visual-Language Models (LVLMs) We demonstrate that the proposed DoCo serves as a plug-and-play pre-training method, which can be employed in the pre-training of various LVLMs without inducing any increase in computational complexity during the inference process.
  arXiv  Detail & Related papers  (2024-02-29T10:17:27Z)
• 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)
• MaMMUT: A Simple Architecture for Joint Learning for MultiModal Tasks [59.09343552273045]
  We propose a decoder-only model for multimodal tasks, which is surprisingly effective in jointly learning of these disparate vision-language tasks. We demonstrate that joint learning of these diverse objectives is simple, effective, and maximizes the weight-sharing of the model across these tasks. Our model achieves the state of the art on image-text and text-image retrieval, video question answering and open-vocabulary detection tasks, outperforming much larger and more extensively trained foundational models.
  arXiv  Detail & Related papers  (2023-03-29T16:42:30Z)

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.