https://youtu.be/GYlbUPGT6qk

Abstract

Today’s scientists are overwhelmed by the pace of new research, making it harder to stay current and avoid duplication. In this talk, I’ll present a new AI-driven approach to accelerate discovery by automating the analysis of scientific literature and data using high-performance computing (HPC). We’ll begin with a brief look at how AI has evolved—from early systems to today’s large language models—and why ORNL is well positioned to lead future breakthroughs, thanks to its cutting-edge HPC and AI expertise.

Our method uses advanced AI to build knowledge graphs from vast collections of research, then applies machine learning and graph analytics to identify promising directions and generate hypotheses, which can be tested via simulation. This approach accelerates science while also guiding the design of energy-efficient, neuromorphic computing systems optimized for AI workloads.

Biographical Sketch

Dr. Thomas E. Potok is a nationally recognized expert in artificial intelligence and data science. At ORNL he founded and led the Computational Data Analytics Group and then, after a reorganization, founded the Data and AI Systems Section, which he led until May of this year. This organization has been guiding major programs in machine learning, text analytics, and neuromorphic computing.

As a principal investigator, he has secured over $36 million in research funding from the Department of Energy, the Department of Defense, and DARPA (the Defense Advanced Research Projects Agency). His work has earned five R&D 100 Awards—often called the “Oscars of Invention”—for innovations ranging from high-speed text analysis systems to AI tools for neuromorphic and quantum computing.

Dr. Potok founded the International Conference on Neuromorphic Systems (ICONS) nine years ago; this year’s conference (https://iconsneuromorphic.cc) is in July in Seattle. He also founded the ORNL Core University AI Workshop five years ago; AI-CORE (https://research.ncsu.edu/initiatives/ornlcuaiws24/) will be at the University of Tennessee this year.

Before moving to Oak Ridge, he worked for 14 years at IBM’s Software Solutions Laboratory in Research Triangle Park in North Carolina, where he conducted research in software engineering productivity. He has B.S., M.S., and Ph.D. degrees in computer engineering from North Carolina State University. 

In addition to his national research contributions, Dr. Potok co-founded VortexT Analytics, successfully commercializing AI technologies for real-world applications. He holds 17 patents, has published over 150 research papers, and is a Battelle Distinguished Inventor, a recognition of his lasting impact on AI and computing.