[3H05-07] Machine Learning 1

11:35 AM - 11:50 AM JST(2:35 AM - 2:50 AM UTC)

[3H07] Development of basic concept of quality management of the use of AI/ML in nuclear power plants(3) Views on Risk Avoidance of AI/ML in basic concept related to AI quality and future challenges

*Kawai Masaki¹, Onodera Masaki¹, Etoh Junji¹, Itakura Toyokazu¹, Shibagaki Kazuhiro¹, Shintaro Yagi¹, Kono Naoyuki², Kashimata Koichi², Aragane Kenji³, Demachi Kazuyuki⁴ (1. MRI, 2. HGNE, 3. NEL, 4. UTokyo)

Keywords:

AI,Reliability,Machine Learning,Risk

In order to promote the introduction of machine learning (ML) technology to the development of advanced reactors and the operation of existing reactors, we are developing a draft of the basic concept (guideline) for Quality Management of the use of AI/ML in the nuclear power plants. In this presentation, we show the concept of the simplified evaluation method of AISL (risk aversion) which is modified from the perspective of the use in nuclear power plants based on the guideline for AI reliability evaluation in the plant safety field. In addition, we show the future considerations on the basic concept of Quality Management of the use of AI/ML in nuclear power plants.