Scholarlink

<p>Multimodal industrial documents–such as operation manuals, circuit diagrams, and parameter tables–contain domain knowledge distributed across text, images, and document layout. However, most existing retrieval-augmented generation (RAG) frameworks rely on static retrieval and fusion policies with fixed modality weights and uniform retrieval depth, making them less adaptable to diverse query intents and dynamic cross-modal dependencies. As a result, they often retrieve incomplete evidence and yield suboptimal reasoning in complex long-document scenarios. To address these challenges, we propose MARL-RAGDoc, a hierarchical multi-agent reinforcement learning framework for multimodal retrieval-augmented reasoning. A high-level coordinator agent dynamically allocates modality weights and retrieval depth based on query characteristics, while specialized text, image, and table agents perform fine-grained evidence selection within their respective candidate pools. A collaborative reasoning module integrates the retrieved evidence and provides hierarchical reward signals to continuously optimize retrieval policies. Experimental results on multiple multimodal document benchmarks demonstrate that MARL-RAGDoc consistently outperforms baselines in both retrieval accuracy and reasoning performance, while remaining computationally efficient. Our code and dataset are publicly available at&#xa0;<a href="https://github.com/Yihong-Q/MARL-RAGDoc">https://github.com/Yihong-Q/MARL-RAGDoc</a>.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hierarchical multi-agent reinforcement learning for retrieval-augmented industrial document question answering

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月14日

查看全文
  • Yihong Qian,
  • Baoli Han,
  • Yufeng Yuan,
  • Xiaofeng Zhang,
  • Hang Zhu,
  • Li Ni,
  • Shaojun Zhong

摘要

Multimodal industrial documents–such as operation manuals, circuit diagrams, and parameter tables–contain domain knowledge distributed across text, images, and document layout. However, most existing retrieval-augmented generation (RAG) frameworks rely on static retrieval and fusion policies with fixed modality weights and uniform retrieval depth, making them less adaptable to diverse query intents and dynamic cross-modal dependencies. As a result, they often retrieve incomplete evidence and yield suboptimal reasoning in complex long-document scenarios. To address these challenges, we propose MARL-RAGDoc, a hierarchical multi-agent reinforcement learning framework for multimodal retrieval-augmented reasoning. A high-level coordinator agent dynamically allocates modality weights and retrieval depth based on query characteristics, while specialized text, image, and table agents perform fine-grained evidence selection within their respective candidate pools. A collaborative reasoning module integrates the retrieved evidence and provides hierarchical reward signals to continuously optimize retrieval policies. Experimental results on multiple multimodal document benchmarks demonstrate that MARL-RAGDoc consistently outperforms baselines in both retrieval accuracy and reasoning performance, while remaining computationally efficient. Our code and dataset are publicly available at https://github.com/Yihong-Q/MARL-RAGDoc.