If HPC had a band of rock stars, Rick Stevens might well play lead guitar and do the vocals. The associate lab director and Exascale Computing Initiative leader at Argonne National Lab (and professor at the University of Chicago) has one of the most impressive records of sustained achievement in the HPC community. Here, in part one of a two-part episode of the @HPCpodcast, Stevens takes up the topic of AI is changing the face of science from the points of view of a scientist and a computer scientist.
In 2019, Stevens led a series of town halls attended by more than 1,000 scientists at at Argonne, Oak Ridge and Berkeley national laboratories on the relevance and applications of AI in scientific research attended by over 1,000 scientists and engineers. In this far ranging conversation with Stevens, he looks at some of the issues (probability vs precision, data management, automated reasoning and surrogate machine learning vs. physics-based HPC simulations) raised during his town hall series, how AI for science has progressed since then and where it’s going in the post-exascale era of supercomputing in the 2020s and beyond.