Current AI computations are significantly different from “classical” onboard computations in both form and output. Thus both detecting and surviving faults is going to require new thinking in both architecture and algorithms, This talk will discuss some of these relationships as we see them now and possible directions of research going forward..

PETER M. KOGGE is the McCourtney Professor of Computer Science and Engineering at the University of Notre Dame, a retired IBM Fellow, a founder of a startup Lucata, Inc, and Notre Dame site director for the multi-university Center for Quantum Technologies. His research interests are in massively parallel computing paradigms and the interaction of non-traditional technology with computing. He holds over 40 patents and is author of several books, including the 2023 “The Zen of Exotic Computing.” His space background includes design of the Space Shuttle IOP and several stints as Visiting Scientist at JPL. In 2008, he led DARPA’s Exascale technology study group, which resulted in a widely referenced report on how the then next 10 years of technology growth would affect exascale computing. Dr. Kogge is an IEEE and AAAS fellow, and has received the Daniel Slotnick best paper award (1994), the IEEE/ACM Seymour Cray award (2012), the IEEE Charles Babbage award (2014), the IEEE Computer Pioneer award (2015), the Gauss best paper award from the Int. Supercomputing Conf. (2015), and the George Cotter award from Chesapeake Large Scale Analytics Conference (2024).