ALS: Presentation on AI Systems for Science

Please join the lab-wide presentation: FASST: Building Frontier AI Systems for Science and the Path to Zettascale this Thursday, July 18, 2024 at 10:00 AM – 12:00 PM in Building 240, Room 1416. You can also join via Zoom using the link below.

Speaker: Rick Stevens, CELS Associate Laboratory Director and Argonne Distinguished Fellow

Abstract: The successful development of transformative applications of AI for science, medicine, and energy research will have a profound impact on the world. The rate of development of AI capabilities continues to accelerate, and the scientific community is becoming increasingly agile in using AI, leading to significant changes in how science and engineering goals will be pursued in the future.

Frontier AI (the leading edge of AI systems) enables small teams to conduct increasingly complex investigations, accelerating tasks such as generating hypotheses, writing code, and automating entire scientific campaigns. However, certain challenges remain resistant to AI acceleration such as human-to-human communication, large-scale systems integration, and assessing creative contributions. Taken together, these developments signify a shift toward more capital-intensive science as productivity gains from AI will drive resource allocations to groups that can effectively leverage AI into scientific outputs. Also, as AI becomes the major driver of innovation in high-performance computing, major shifts in the computing marketplace will occur over the next decade as a growing performance gap is established between systems designed for traditional scientific computing versus those optimized for large-scale AI such as Large Language Models (LLMs).

In response to these trends while also in recognition of the role of government supported research to shape the future research landscape, the U.S. Department of Energy (DOE) has created the Frontier AI for Science, Security, and Technology (FASST) initiative. FASST is a decadal research and infrastructure development initiative aimed at accelerating the creation and deployment of frontier AIsystems for science, energy research, and national security.

ALD Rick Stevens will present a sweeping overview of the FASST initiative, including a review of the initiative’s goals, along with how it’ll transform the research of the DOE national laboratories. The talk will touch on the recently established Trillion Parameter Consortium (TPC), whose aim is to foster a community-wide effort to accelerate the creation of large-scale generative AI for science, as well as the AuroraGPT project, an international collaboration to build a series of multilingual multimodal foundation models for science that are pretrained on deep domain knowledge to enable them to play key roles in future scientific enterprises.

This Lab-wide presentation invites all Laboratory staff, scientists, and students to attend and engage on a groundbreaking direction of science.

Join via Zoom link here: https://argonne.zoomgov.com/j/1610819589?pwd=ce4gIA7SgMHxe9bXH00br1ZLHxej5z.1