[1A13-16] Knowledge Management System

4:15 PM - 4:30 PM JST(7:15 AM - 7:30 AM UTC)

[1A13] TASKI:AI-based nuclear knowledge management system(1) Application of large language models to the nuclear energy field

*Tomohiro Okamura^1,2, Takahiro Nishihara^1,2, Hidekazu Asano¹, Masahiko Nakase^1,2 (1. Science Tokyo, 2. NEUChain inc.)

Keywords:

Nuclear Knowledge Management,Large Language Models (LLMs),Knowledge & Skill Transfer,AI Agents,TASKI

In the nuclear sector, the vast volumes of design, operational, and maintenance data coupled with the loss of tacit knowledge through generational turnover make reliable transfer of expertise a critical challenge. Recently, large language models (LLMs) capable of understanding domain specific terminology in context and performing reasoning have emerged as a promising key to this problem. Institute of Science Tokyo and NEUChain Technologies, Inc. are jointly advancing TASKI, an AI-based nuclear knowledge-management system. TASKI integrates a comprehensive knowledge base built from existing reports, regulatory documents, and research papers with search algorithms, AI agents, numerical analysis, and machine-learning capabilities, aiming to provide seamless access to knowledge, semi-automate analytical and documentation tasks, and support human-resource development. This presentation will detail the TASKI project and discuss domestic and international case studies of LLM applications in the nuclear field, examining their effects and future potential.