基于生成式逻辑的古籍文献自动化置标语义框架构建与应用研究  

Research on the Construction and Application of Automated Semantic Framework for Ancient Texts Based on Generative Logic

作　　者：文玉锋[1] 赵悦言 Wen Yufeng;Zhao Yueyan

机构地区：[1]西北师范大学管理学院,甘肃兰州730070 [2]四川大学公共管理学院,四川成都610065

出　　处：《图书与情报》2024年第2期126-134,共9页Library & Information

基　　金：国家社会科学基金一般项目“古籍文本挖掘与知识计算研究”(项目编号:23BTQ020);四川大学2024年中央高校自主立项项目“数智赋能公共文化服务创新机制研究--以智慧文旅平台为例”(项目编号:2024自研-公管12)研究成果之一。

摘　　要：目前,我国古籍文献的数字化以文献扫描、粗粒度文件管理等浅层知识服务为主,生成式人工智能技术的发展为古籍文献数字化的深度化提供了新的机遇。文章基于框架语义学理论构建置标语义逻辑结构框架,以生成式逻辑向大语言模型提出问题,递归提取古籍语料中深层语义内容,并将其输出为符合置标语义框架的结构化数据,使古籍文本在基础语义层面获得统一的处理逻辑。古籍自动置标语义框架能够实现大规模自动化古籍文献内容结构生成式表征,为古籍整理智能化转型提供一种自动可行的技术方案。Currently,the digitalization of ancient Chinese texts primarily involves basic document scanning and coarse-grained file management.However,the development of generative artificial intelligence technologies offers new opportunities for advancing the depth of digitalization of ancient texts.Based on the theory of frame semantics,this article constructs a semantic logical structure framework for semantic tagging.It poses questions to large language models using generative logic,recursively extracts deep semantic content from ancient text corpora,and outputs it as structured data conforming to the semantic tagging framework.This enables ancient texts to be processed with a unified semantic logic at the basic semantic level.The automatic semantic tagging framework for ancient texts can achieve large-scale,automated generative representation of the content structure of ancient documents,providing an automatically viable technical solution for the intelligent transformation of ancient text organization.

关 键 词：古籍文本 生成式表征 自动置标语义框架 大语言模型 

分 类 号：G255.1[文化科学—图书馆学]