知识图谱和大语言模型辅助新工科课程教学资源建设方法
On the Method of Constructing Teaching Resources for New Engineering Courses Assisted by Knowledge Graph and Large Language Model
王佐旭1
作者信息
摘要
新工科专业课程资源建设常要求内容新、教学方法新,同时满足学生个性化学习.知识图谱由于其强大的结构化表示能力、高效的知识检索能力和优秀的知识推理能力,适用于工科课程教学资源的建设.同时,大语言模型具有优异的自然语言理解和推理能力,能够基于预训练模型完成知识抽取、自然语言问答等功能,可与知识图谱相结合来构建新工科课程教学资源平台.本文提出了一个知识图谱和大语言模型技术辅助工科课程教学资源建设的方法,结合自然语言处理技术、深度学习技术、预训练语言模型等技术完成工科课程知识的本体构建、知识抽取和知识图谱快速构建,并以制造信息学原理与应用课程为例进行了应用验证.结果表明,该方法可以在较短的时间内完成新工科专业课程资源的整理和建设,对新工科专业建设提供了借鉴意义.
Abstract
The teaching resource construction for new engineering majors often requires new content and new teaching methods,while meeting students'personalized learning.Knowledge graphs are well-suited for this purpose due to its powerful capabilities on knowledge representation,knowledge re-trieval,and knowledge reasoning.Meanwhile,large language models exhibit excellent natural language understanding and reasoning capabilities,enabling them to perform knowledge extraction,natural language question answering,etc.,based on pretrained mod-els.They can be integrated with knowledge graphs to construct a course resource platform for emer-ging engineering disciplines.This article proposes an approach for utilizing knowledge graph and large language model technology to construct teaching resources for engineering courses.The proposed ap-proach integrates natural language processing techniques,deep learning technologies and pre-trained language models to achieve ontology construction,knowledge extraction,and rapid knowledge graph construction for engineering course content.It is exemplified through a case study of a course entitled Manufacturing Informatics.The results show that the organization and construction of the course:Manufacturing Informatics can be completed in a short period of time,which provides a practical ref-erence for the construction of new engineering majors.
关键词
工科课程/知识图谱/大语言模型/教学资源建设/知识抽取Key words
engineering course/knowledge graph/large language model/teaching resource construc-tion/knowledge extraction引用本文复制引用
出版年
2025
NETL
NSTL
万方数据