Research

Paper

AI LLM March 17, 2026

Evo-Retriever: LLM-Guided Curriculum Evolution with Viewpoint-Pathway Collaboration for Multimodal Document Retrieval

Authors

Weiqing Li, Jinyue Guo, Yaqi Wang, Haiyang Xiao, Yuewei Zhang, Guohua Liu, Hao Henry Wang

Abstract

Visual-language models (VLMs) excel at data mappings, but real-world document heterogeneity and unstructuredness disrupt the consistency of cross-modal embeddings. Recent late-interaction methods enhance image-text alignment through multi-vector representations, yet traditional training with limited samples and static strategies cannot adapt to the model's dynamic evolution, causing cross-modal retrieval confusion. To overcome this, we introduce Evo-Retriever, a retrieval framework featuring an LLM-guided curriculum evolution built upon a novel Viewpoint-Pathway collaboration. First, we employ multi-view image alignment to enhance fine-grained matching via multi-scale and multi-directional perspectives. Then, a bidirectional contrastive learning strategy generates "hard queries" and establishes complementary learning paths for visual and textual disambiguation to rebalance supervision. Finally, the model-state summary from the above collaboration is fed into an LLM meta-controller, which adaptively adjusts the training curriculum using expert knowledge to promote the model's evolution. On ViDoRe V2 and MMEB (VisDoc), Evo-Retriever achieves state-of-the-art performance, with nDCG@5 scores of 65.2% and 77.1%.

Metadata

arXiv ID: 2603.16455
Provider: ARXIV
Primary Category: cs.CV
Published: 2026-03-17
Fetched: 2026-03-18 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.16455v1</id>\n    <title>Evo-Retriever: LLM-Guided Curriculum Evolution with Viewpoint-Pathway Collaboration for Multimodal Document Retrieval</title>\n    <updated>2026-03-17T12:36:56Z</updated>\n    <link href='https://arxiv.org/abs/2603.16455v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.16455v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Visual-language models (VLMs) excel at data mappings, but real-world document heterogeneity and unstructuredness disrupt the consistency of cross-modal embeddings. Recent late-interaction methods enhance image-text alignment through multi-vector representations, yet traditional training with limited samples and static strategies cannot adapt to the model's dynamic evolution, causing cross-modal retrieval confusion. To overcome this, we introduce Evo-Retriever, a retrieval framework featuring an LLM-guided curriculum evolution built upon a novel Viewpoint-Pathway collaboration. First, we employ multi-view image alignment to enhance fine-grained matching via multi-scale and multi-directional perspectives. Then, a bidirectional contrastive learning strategy generates \"hard queries\" and establishes complementary learning paths for visual and textual disambiguation to rebalance supervision. Finally, the model-state summary from the above collaboration is fed into an LLM meta-controller, which adaptively adjusts the training curriculum using expert knowledge to promote the model's evolution. On ViDoRe V2 and MMEB (VisDoc), Evo-Retriever achieves state-of-the-art performance, with nDCG@5 scores of 65.2% and 77.1%.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <published>2026-03-17T12:36:56Z</published>\n    <arxiv:comment>Accepted by CVPR2026</arxiv:comment>\n    <arxiv:primary_category term='cs.CV'/>\n    <author>\n      <name>Weiqing Li</name>\n    </author>\n    <author>\n      <name>Jinyue Guo</name>\n    </author>\n    <author>\n      <name>Yaqi Wang</name>\n    </author>\n    <author>\n      <name>Haiyang Xiao</name>\n    </author>\n    <author>\n      <name>Yuewei Zhang</name>\n    </author>\n    <author>\n      <name>Guohua Liu</name>\n    </author>\n    <author>\n      <name>Hao Henry Wang</name>\n    </author>\n  </entry>"
}