生成式AI驱动下的知识封装与知识传播新形态研究  

Research on New Forms of Knowledge Packaging and Dissemination Driven by Generative Artificial Intelligence

作　　者：易龙 YI Long(School of Humanities,Central South University,410012,Changsha,China)

机构地区：[1]中南大学人文学院,长沙410012

出　　处：《数字出版研究》2024年第2期84-93,共10页DIGITAL PUBLISHING RESEARCH

基　　金：2023年度湖南省教育厅科学研究优秀青年项目“生成式人工智能背景下的知识封装与传播形态研究”(项目编号:23B0019)。

摘　　要：知识封装技术的发展推动了出版形态的更新,从数据库技术、搜索引擎再到大语言模型技术,知识表示、提取、生成和传播的方式不断演化。生成式AI可对各类数据文本进行学习并推理预测,进而生成知识性内容输出。一方面,生成式AI技术因其知识生成的潜在效果而被普遍用于知识封装产品之中;另一方面,也必须关注生成式AI可能带来的一些问题,如数据可耗尽性、输入依赖性、不平等性、解释性不足及稳健性不足等。生成式AI通过对话式知识提取、多模态知识表达和知识智能体等方式,正在深层重塑知识传播形态,理解其演进特征将有助于更好地驾驭这一新型知识系统。Development of knowledge packaging technology has driven the renewal of publishing forms.Ways in which knowledge is represented,extracted,generated and disseminated continue to evolve from database and search engine technologies to large language model technology.Generative artificial intelligence(GAI)can learn from various data texts and make inferential predictions,thereby outputting knowledge-based content.On the one hand,GAI is widely used in knowledge-packaging products due to its potential for knowledge generation;on the other hand,issues such as data exhaustibility,input dependency,inequality,weak interpretability,and weak robustness highlight the need to pay attention to the potential consequences of GAI.It is profoundly reshaping knowledge dissemination form through dialogical knowledge extraction,multimodal knowledge expression and knowledge agents.Understanding its evolutionary characteristics may lend a hand to better navigate this new type of knowledge system.

关 键 词：生成式AI 知识封装 出版形态 知识传播 大语言模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] G230.7[自动化与计算机技术—控制科学与工程]