Developing AI-assisted high performance fluid simulation codes
前山 伸也(大学共同利用機関法人 自然科学研究機構 核融合科学研究所)

In this project, we aim at establishing AI-assisted HPC fluid simulations. As a proof of principle, we will demonstrate simple model cases in a scalable manner to production applications. We will focus on two important targets: modeling of subgrid-scale dynamics and data-assimilation procedures. We will establish the strategy to loosely couple HPC fluid simulation codes and AI models in a flexible manner.

研究紹介ポスター／最終報告書