Presentation Information
[1A14]TASKI:AI-based nuclear knowledge management system(2) AI Agent for Data Driven Research
*Takahiro NISHIHARA1,2, Tomohiro OKAMURA1,2, Hidekazu ASANO1, Masahirko NAKASE1,2 (1. Science Tokyo, 2. NEUChain inc.)
Keywords:
AI Agent,Machine Learnning,Accelerated Simulation,Actinide Chemistry,Foundation Model
Here's the English abstract (119 words):
The nuclear field has accumulated vast amounts of undocumented data, requiring efficient management and utilization systems. The TASKI project develops AI agents to automate data management, analysis, and organization for multiple applications in our laboratory. For instance, the Acceleration of Actinide Chemistry Experiment (AACE) system employs transfer learning and Human-in-the-Loop machine learning to predict experimental results from surrogate to actual substances, enabling rational experimental planning and reducing experimental iterations. This research advances Large Language Model (LLM)-based AI agent development to enhance data preprocessing and analytical capabilities, aiming to maximize experimental efficiency while reducing researcher burden. This presentation demonstrates AI agent construction methodologies and application examples, discussing the potential impact and applicability across the nuclear field.
The nuclear field has accumulated vast amounts of undocumented data, requiring efficient management and utilization systems. The TASKI project develops AI agents to automate data management, analysis, and organization for multiple applications in our laboratory. For instance, the Acceleration of Actinide Chemistry Experiment (AACE) system employs transfer learning and Human-in-the-Loop machine learning to predict experimental results from surrogate to actual substances, enabling rational experimental planning and reducing experimental iterations. This research advances Large Language Model (LLM)-based AI agent development to enhance data preprocessing and analytical capabilities, aiming to maximize experimental efficiency while reducing researcher burden. This presentation demonstrates AI agent construction methodologies and application examples, discussing the potential impact and applicability across the nuclear field.
Comment
To browse or post comments, you must log in.Log in