Argonne is developing cutting-edge artificial intelligence (AI) tools and methods to help enhance scientific discovery. The research takes several different avenues — from large language models to generative AI. Regardless of the road taken, the challenges faced are numerous. To help users and developers, Argonne researchers will offer three half-day tutorials: one on middleware for supporting agentic workflows across heterogeneous research environments, another on performance tuning of ML/AI applications, and a third on leveraging and evaluating large language models for scientific computing. They’ll also give a workshop talk on high-performance user support tools.   

As AI expands its boundaries, opportunities arise for scientific breakthroughs, but questions also arise. Argonne researchers and their colleagues will participate in three workshops exploring issues such as “How can we enable wider deployment of AI in high-performance computing? “What new requirements does in situ AI present? Are foundation models trustworthy?” and “What are the benefits of incorporating AI in extreme-scale workflows?”