摘要
探索融合本地化部署的大语言模型与知识图谱,以提升智能教学设计与支持。以“网络爬虫与商业预测分析”课程为例,开展数据预处理、模型优化、知识图谱构建及应用服务集成,实现知识点结构化表示、动态更新、智能问答与个性化资源推荐。构建了学生画像与教师决策支持模块。研究结果显示,本地化部署的大语言模型能够提供实时精准的知识支持,而知识图谱引入优化了资源管理和学习路径规划。这种融合模式提升了学生交互体验和知识获取效率,能够帮助教师更好地把握学生需求,制定针对性教学策略。此外,该模式还强化了数据隐私保护与资源的自主可控性,符合高校数字化转型的安全合规要求。
This study explores the integration of locally deployed large language models(LLM)and knowledge graphs(KG)to enhance intelligent teaching design and support.Using a course on Web crawling and business prediction analysis as a case study,the research involves data preprocessing,model optimization,KG construction,and application service integration.This approach enables the structured representation and dynamic updating of knowledge points,and provides AI-powered question answering and personalized resource recommendations.Additionally,student profiling and teacher decision support modules are developed.The findings show that the local deployment of LLM offers real-time and accurate knowledge support,while the introduction of KG optimizes resource management and learning path planning.This integrated model not only improves student interaction experiences and knowledge acquisition efficiency,but also assists teachers in better understanding student needs,thereby enabling the development of more targeted teaching strategies.Moreover,it strengthens data privacy protection and resource controllability,aligns with the safety and compliance requirements of university digital transformation.The integration of LLM and KG represents a significant advancement in smart education,offers a promising direction for future educational innovation.
作者
毛志新
冯睿
张智
刘文侠
MAO Zhixin;FENG Rui;ZHANG Zhi;LIU Wenxia(College of Business Administration,Shanghai Business School,Shanghai 200235,China;School of Business,Changshu Institute of Technology,Changshu 215500,Jiangsu,China)
出处
《实验室研究与探索》
北大核心
2025年第5期141-147,共7页
Research and Exploration In Laboratory
基金
2024年教育部人文社科项目(24YJC630291)
2022年上海市哲学社会科学基金项目(2022ZGL003)
2024年江苏高校哲学社会科学研究项目(2024SJYB1057)。
关键词
本地大语言模型
知识图谱
课程设计
智能问答
数据隐私
local large language model
knowledge graph
course design
AI-powered question answering
data privacy
