人工智能时代的元数据方法论  

Metadata Methodology in Artificial Intelligence Era

作　　者：刘炜[1] 刘倩倩 付雅明 祝蕊 Liu Wei;Liu Qianqian;Fu Yaming;Zhu Rui

机构地区：[1]上海图书馆/上海科学技术情报研究所 [2]南京大学信息管理学院 [3]上海大学文化遗产与信息管理学院

出　　处：《图书馆理论与实践》2023年第4期16-29,共14页Library Theory and Practice

摘　　要：元数据是关于数据的数据,随着技术的进步,元数据获取逐渐成为信息系统数据建模和实现功能的关键性步骤,发展起一套包括实体定义、关系描述、对象分析、属性提取、本体建模,以及数据清洗、消歧、对齐、映射、关联、丰富、导入、导出乃至服务部署、注册发现、运行监测等一系列操作的方法论体系,旨在帮助实现任何信息体的结构化描述、语义编码和机器理解。这些不仅是语义技术(包括关联数据)和知识图谱技术必需的应用,而且已成为信息系统建立独立的、基于知识的内容架构的基本操作和主要方案。文章把与元数据相关的一系列方法体系统称为元数据方法,相关的最佳实践基本体现于语义万维网已经制定、正在制订或正在考虑制订的各项标准规范中。元数据方法在未来基于Web 3.0的多模态元宇宙建设中会继续起到多方面的重要作用,如利用知识模型构建数字孪生,甚至支持对整个虚拟世界的建模等。当然基于人工的描述和编码显然不能适应元宇宙时代用户生产内容(UGC)和ChatGPT带来的人工智能生成内容(AIGC)的内容生产方式,必须有一套方法论帮助自动实现语义形式化。这应该是元数据方法适应未来智慧时代需求的必由之路。Metadata is the data about data.With the progress of technology,metadata acquisition gradually becomes a key step of data modeling and implementation function in information system.Developing a set of operational methodology system including entity definition,relationship description,object analysis,attribute extraction,ontology modeling,the cleaning,disambiguation,alignment,mapping,association,rich,import,export of data,service deployment,registration discovery,and operational monitoring,is to help achieve structured description,semantic coding and machine understanding of any information body.These are not only the necessary application of semantic technology(including related data)and knowledge graph technology,but also have become the basic operation and main scheme for information system to establish an independent,knowledge-based content framework.In this article,a series of methods and systems related to metadata are called metadata methods.The relevant best practice basically reflected in semantic World Wide Web has been developed,being developed or under consideration in each standard specifications.Metadata method will continue to play an important role in the construction of multi-modal metaverse based on Web3.0.For example,it can use knowledge model to build digital twins,and even support the modeling of the entire virtual world.Of course,human-based description and coding obviously cannot adapt to the content production mode of user generation content(UGC)and artificial intelligence generated content(AIGC)brought by Chat GPT in metaverse era.There must be a set of methodology to help automatically achieve semantic formalization.This should be an indispensable way for metadata method to adapt to the needs of future intelligent era.

关 键 词：元数据方法 内容架构 语义建模 知识本体 Web 3.0 ChatGPT 元宇宙 

分 类 号：G254[文化科学—图书馆学]