Objective of FJOH-2026

The main objective of the FJOH-2026 edition is to provide the school participants with a working knowledge of AI methods used in the nuclear sector. The lectures will focus on how data-driven methods (using machine learning algorithms) and model-driven methods, combined with human oversight, are impacting and progressively transforming our R&D landscape, helping us develop new insight and unlock new possibilities. These AI-driven methods are increasingly used for high-fidelity reactor and fuel simulations, large dataset analyses, complex design studies or safety analyses; they help to complement, extend, improve, optimize computer simulations for enhanced predictions and better decision making.

The course includes (i) general lectures on AI methods, best practices and limitations, (ii) specialized lectures on reactor physics, thermal hydraulics, fuels and materials; and (iii) lectures and seminars on AI integration in practice. Special attention will be paid to situations where key data are missing or of insufficient quality, and how machine-learning techniques can help bridge these data gaps.

FJOH-2026 includes plenary lectures, seminars, and technical visits. The FJOH-2026 participants will also have the opportunity to practice their freshly-acquired knowledge as part of group activities. Time for these group activities is set aside in the School schedule.

By the end of the course, the participants should be able to describe (i) the main AI methods currently available and how they are used in reactor physics, thermo-hydraulics, fuels and materials R&D; (ii) the benefits and limitations of these methods.