[4Xin2] Poster session 2

[4Xin2-106] Development of Concept Acquisition Support Chatbot Empowered by Knowledge Graphs and Large Language Models

〇Atsushi Omata^1,2, Yuka Enomoto², Neteru Masukawa², Shogo Ishikawa² (1.Hamamatsu University School of Medicine, 2.Shizuoka University)

Keywords:LLM, Chatbot, Knowledge Graph, Concept Acquisition Support

This paper describes a method that leveraging a chatbot, enhanced by Knowledge Graphs and Large Language Models (LLMs), to facilitate the learning of specialized concepts. Addressing the challenge that cultural and linguistic differences pose in translating specialized terms, we utilize Knowledge Graphs to structure these concepts and incorporate them as external knowledge for LLM-based chatbots. This approach is applied to the concept of recovery, gaining prominence in Japan, through the development of a chatbot and a Knowledge Graph. Feedback from usability tests with university students has led to refinements in knowledge graphs the utilization and the design of prompts for better question answering. The findings confirm that this method enables users to acquire accurate concepts and information effectively, highlighting its potential in educational applications.