P23: AI with Super-Computed Data for Monte Carlo Earthquake Hazard Classification
Abstract: Many problems associated with earthquakes are yet to be solved using heroic computing, which is defined as computing at the largest scale possible using the best supercomputers and algorithms. Thus, a continuous effort has been pursued in HPC to solve these problems. However, even when heroic computing is applied, its practical use is difficult without considering the uncertainties in models. In this study, we constructed an AI methodology that uses super-computed data generated using heroic computing. We applied this AI to an earthquake hazard classification including uncertainty analyses in order to demonstrate its utility. This study can be regarded as an innovative step towards realizing high quality computing for Earthquakes by exploiting the potential of HPC through AI.
Award: Best Poster Finalist (BP): yes
Two-page extended abstract: pdf