图书馆杂志

图书馆杂志 ›› 2025, Vol. 44 ›› Issue (415): 15-27.

• 工作研究 • 上一篇    下一篇

人机智能融合下的红色文献知识检索研究

严承希1, 2  刘越男1, 2  余 敏1  张燚明2, 3(1 中国人民大学信息资源管理学院 2 中国人民大学数字人文研究院 3 中国人民大学历史学院)   

  • 出版日期:2025-11-15 发布日期:2025-11-26
  • 作者简介:

    严承希 中国人民大学信息资源管理学院,讲师。研究方向:数字人文与文化遗产、自然语言处理与大模型智能计算。作者贡献:初稿撰写与修改、系统设计与开发。E-mail 20218113@ruc. edu. cn 北京 100872

    刘越男 中国人民大学信息资源管理学院,院长,教授。研究方向:电子文件管理、数学档案馆、数据治理。作者贡献:确定研究选题与框架、修改与完善论文。 北京 100872
    余 敏 中国人民大学信息资源管理学院,博士研究生。研究方向:数字人文、数字叙事。作者贡献:完善研究框架、系统设计、修改与完善论文。 北京 100872
    张燚明 中国人民大学历史学院,讲师。研究方向:中国近现代史、中共党史与抗日战争史。作者贡献:提供数据支持、修改与完善论文。 北京 100872

Research on Knowledge Retrieval of Red Documents under Integrated Human-Machine Intelligence

Yan Chengxi1 2 Liu Yuenan1 2 Yu Min1 Zhang Yiming2 3(1 School of Information Resource Management Renmin University of China 2 Digital HumanitiesCenter Renmin University of China 3 School of History Renmin University of China)   

  • Online:2025-11-15 Published:2025-11-26
  • About author:

    Yan Chengxi1 2 Liu Yuenan1 2 Yu Min1 Zhang Yiming2 3(1 School of Information Resource Management Renmin University of China 2 Digital Humanities Center Renmin University of China 3 School of History Renmin University of China)

摘要:

红色文献检索服务是红色资源深度开发与建设的重要实践方式。目前红色文献检索的应用服务主要围绕元数据展开,相关查询式构建策略与导航设计机制等都未能较好地服务于各类用户需求,特别是缺少对于红色文献内容知识的深度挖掘与展现。为了解决上述问题,本研究立足于红色文献检索服务的特点,结合当前主流的红色文献检索系统的通用特征,提出了一种以用户为中心的红色文献知识检索方案。该方案融入了前沿的人机智能融合技术,在目前的元数据检索模型基础上,不仅加入了红色文献的内容实体知识以丰富检索范围,而且设计了两个查询扩展模块,即基于“基本语义扩展”的语义扩展模块和基于AI 语义扩展的“ 人智交互模块”,以分别提供不同知识层次的查询拓展。根据上述方案,本研究开发了一个红色文献知识检索原型系统,并进行了多组用户实验评测。实验结果表明与其他基于元数据检索模型的红色文献检索平台相比,该原型系统在系统功能的丰富性与用户检索的评估指标上都更具优势。此外,两种新提出的检索扩展模块各自具有鲜明的特点和适用范围,可满足不同的检索需求类型。上述研究成果可为红色文献检索服务的知识化与智能化转型发展提供一定的技术参考。

关键词: 人机智能融合&emsp, 红色文献&emsp, 知识检索&emsp, 查询扩展

Abstract:

The retrieval of red documents is an important pathway for advancing the construction anddeep utilization of red digital resources. At present the red document-based retrieval mainly relies onmetadata in which the query formulation strategies and navigation design mechanisms fail to adequatelymeet diverse user needs. In particular existing approaches lack in-depth mining and content-levelpresentation. To address the problems above this study proposes a user-centered knowledge retrievalsolution for red documents which integrates both the functional features of retrieval systems and theunique characteristics of red documents. The model incorporates cutting-edge human-machine intelligencetechnology into the current metadata-based retrieval framework. Specifically it enriches the retrievalscope through entity-level knowledge extraction from document content and introduces two queryexpansion modules namely a 􀆵 semantic expansion module and a 􀆵 human-AI interaction module .These modules provide users with expanded query terms at multiple levels of knowledge representation.Accordingly we developed a knowledge retrieval prototype system for red documents and conductedmultiple user experiments for evaluation. The results show that compared with existing metadata-drivenretrieval platforms our proposed system has more advantages in functional richness and user satisfaction. In addition the two newly proposed query expansion modules have their own distinct characteristics andapplicability to meet different types of search needs. These findings provide certain technical referencesfor the transformation and development of red document retrieval services towards a more knowledgeableand intelligent direction.

Key words: Integrated human-machine intelligence, Red literature, Knowledge retrieval, Query expansion