Indiana University and the Luddy School of Informatics, Computing, and Engineering are among 13 new members to the Trillion Parameter Consortium (TPC), an international collaboration of research and technology organizations engaging to build large AI models for scientific applications.
TPC includes scientists from federal laboratories, research institutes, academia and industry who work together to address the challenges of building large-scale AI systems, trained on scientific data and intended for science and engineering applications, and advancing trustworthy and reliable AI for scientific discovery.
The U.S. Department of Energy’s Argonne National Laboratory is a TPC founding partner. The laboratory is a multidisciplinary science and engineering research center focused on working to answer the biggest questions facing humanity.
Beth Plale, Intelligent Systems Engineering chair and Michael A and Laurie Burns McRobbie Bicentennial Professor of Computer Engineering, is Luddy’s primary TPC representative. David Crandall, Luddy Professor of Computer Science and Luddy Artificial Intelligence Center director, is the secondary representative.
“The Luddy School and the wider AI-invested community at Indiana University shares TPC’s commitment to creating a global network of resources and expertise,” Plale said. “The Luddy School expertise in particular spans from trusted AI through open science to high-performance computing.
“The global nature of the TPC consortium gives the Luddy School an important lens for readying the next generation of AI researchers.”
Companies such as OpenAI and Anthropic are spending billions of dollars on machine learning systems. The problem -- the systems are grown and not designed. Scientists don’t completely understand how AI learns. If left unchecked, such systems could eventually conceal dangerous activities from humans. Thus, the need for strict safety guidelines.
Luddy School Dean Joanna Millunchick has pushed AI advances in conjunction with responsible safety principles and processes. She said the Luddy School shares TPC goals of building an open community of researchers creating state of the art large-scale generative AI models for science and engineering; incubating, launching and facilitating coordination and collaboration for specific projects building such models; and creating a global network of resources and expertise to facilitate teaming and training of next-generation AI researchers.
Other new members include Harvard University, Princeton University, the University of Michigan, the University of Toronto, Stonybrook University, the University of California San Diego’s San Diego Supercomputer Center, Sony AI, Deep Forest Sciences and Inria.
The consortium seeks to understand key challenges in advancing AI for science, and to create a community within which teams can find and pursue collaborations, share insights and form partnerships.
Since its formation in 2023, TPC has grown to more than 850 participants from over 100 organizations around the world.
Global partners will meet at the European TPC Kickoff workshop June 19-21 in Barcelona, Spain. The workshop offers participation opportunities by the European AI, HPC and disciplinary science research communities.