图书馆杂志

图书馆杂志 ›› 2024, Vol. 43 ›› Issue (400): 4-12.

• 专题 •    下一篇

面向全流程管理的专家协同知识生产系统设计与实现——以家谱联合编目系统为例

朱武信 夏翠娟 吴建伟 (上海图书馆)   

  • 出版日期:2024-08-15 发布日期:2024-08-27
  • 作者简介:朱武信 上海图书馆(上海科学技术情报研究所), 工程师。研究方向:数字人文。作者贡献:数据调研、内容撰写。E-mail:wxzhu@libnet.sh.cn 上海 200031 夏翠娟 上海图书馆(上海科学技术情报研究所),研究员。研究方向:知识组织、数字人文、数字记忆。作者贡献:论文选题和框架指导、文章整体修改定稿。上海 200031 吴建伟 上海图书馆(上海科学技术情报研究所),研究馆员。研究方向:历史文献学。作者贡献:参与文中著录规范制定、专业知识指导。上海 200031

Design and Implementation of Expert Collaborative Knowledge Production System for Full Process Management: Genealogy Joint Cataloging System

Zhu Wuxin, Xia Cuijuan, Wu Jianwei (Shanghai Library)   

  • Online:2024-08-15 Published:2024-08-27
  • About author:Zhu Wuxin, Xia Cuijuan, Wu Jianwei (Shanghai Library)

摘要: 在数智时代,传统编目系统面临着挑战,难以满足基于智慧数据生成的知识生产要求。上海图书馆对传统编目流程进行改进,融入基于众包理念的专家协同知识生产与全流程管理理念,设计了面向全流程管理的专家协同知识生产系统——家谱联合编目系统。该系统实现了:利用上海图书馆数字人文建设已有的人、机构、地、时、事等基础知识库,实现了编目过程中跨网域的规范控制,不同机构的家谱编目专家,共享基础知识库中已有的知识、共同补充丰富编目过程中缺失的知识,实时构建知识关联和语义链接;并将编目流程和知识生产与家谱知识服务平台打通,由此知识生产与知识服务无缝地链接起来,实现了从采购、捐赠、著录、查重、质量检测、规范控制、知识服务、关联数据发布的全流程管理;在此基础上,探索了基于大语言模型的AI 家谱编目数据审校的路径。

关键词: 数字人文 全流程数据加工 联合编目 众包 AIGC

Abstract:

In the age of digital intelligence, traditional cataloging systems are facing challenges and are unable to meet the requirements for knowledge production based on intelligent data generation. The Shanghai Library has improved the traditional cataloging process, incorporated the concept of expert collaborative knowledge production and full-process management based on the concept of crowdsourcing, and designed an expert collaborative knowledge production system for full-process management— the Genealogy Joint Cataloging System. This system could realize the standard control across network domains in the cataloging process by using the existing basic knowledge base of people, institutions, places, times, events, etc. of the Library. Genealogy cataloging experts from different institutions can share basic knowledge. The existing knowledge in the library can jointly supplement and enrich the knowledge missing in the cataloging process, build knowledge associations and semantic links in real time, and connect the cataloging process and knowledge production with the genealogy knowledge service platform, so that knowledge production and knowledge services are seamlessly linked. Together, we have achieved full-process management of procurement, donation, description, duplication checking, quality inspection, normative control, knowledge services, and related data release. On this basis, the path of AI genealogy cataloging data review based on the big language model was explored.

Key words: Digital humanities, Full-process data processing, Joint cataloging, Crowdsourcing, AIGC