The ability of artificial intelligence (AI) to successfully learn from large datasets has transformed science and engineering as we know it. AI can accelerate scientific discovery and innovation but often requires more computing power than is available to most researchers. The DOE provides supercomputers to solve the nation’s biggest scientific challenges, and this series aims continue to deepen and expand the AI knowledge base for the next generation of AI practioners.
Building on the ALCF's robust training program in the areas of AI and supercomputing, the training series includes hands-on courses that teach attendees how to use leading-edge supercomputers to develop and apply AI solutions for the world's most challenging problems. The series focuses on the following:
- Understanding the underlying concepts of neural networks and large language models
- Appreciating the need for parallel processing and supercomputing as AI problems scale up
- Broaden awareness of applications of AI/ML to scientific problems
- Exposure to current advances in AI hardware and their integration into next-generation supercomputers
Pre-requisites
This training series is aimed at undergraduate and graduate students enrolled at U.S. universities and community colleges. Attendees are expected to have basic experience with Python.
For the Intro to AI-Driven Science on Supercomputers series, no supercomputing or AI knowledge is required.
For the Advanced Topics in AI series, participants are expected to:
- Have either attended the ALCF Intro to AI-Driven Science on Supercomputers series or have familiarity with the topics covered in the series
- Topics include foundational concepts in parallel computing, neural networks, large-language models, prompt engineering, AI accelerators
Advanced Topics attendees will deepen their understanding of applying AI at scale, learn about coupling science simulations with AI, dig into inference workflows, and explore how AI accelerators are enabling AI for Science.
Workshop Series Format
Each session will have both lecture and hands-on components, along with a talk from an Argonne scientist about the work they do using AI for their science.
Each session occurs on Tuesdays from 3:00-4:30 p.m. CT. Session recordings will be made available shortly after each session.
Attendees who complete all in-class and post-class exercises by the end of the series will receive a certificate of completion and a digital badge.
Session materials are hosted on the ALCF AI Science Training series GitHub [click here].
Recordings for each session will be posted weekly for your review.