Advanced Computing in the Age of AI | Monday, July 15, 2024

BSC Hosts European Kick-Off of the Trillion Parameter Consortium in Barcelona 

June 18, 2024 -- The The Barcelona Supercomputing Center – Centro Nacional de Supercomputación (BSC-CNS) has announced the European Kick-Off of the International Trillion Parameter Consortium (TPC), an event focused on accelerating the development and use of Generative Artificial Intelligence (AI) for science and engineering. The event will take place from June 19-21, 2024, at the Occidental Atenea Mar Hotel in Barcelona, Spain.

Credit: Shutterstock

The European Kick-Off of the TPC will bring together industry leaders, researchers, and practitioners to discuss the transformative potential of Generative AI in scientific and engineering applications, with a special focus on engaging the European AI, HPC, and disciplinary science research communities. The event will feature keynote speeches, panel discussions, and hands-on workshops, providing attendees with insights into the latest trends and innovations.

"The predictive power of Large Language Models (LLMs) has increasingly made them essential across many scientific domains, presenting new challenges but also significant opportunities," said Mateo Valero, BSC Director. "This event is an occasion to explore how Generative AI is reshaping science and engineering, and to accelerate its adoption across various fields."

The International Trillion Parameter Consortium (TPC)

The International Trillion Parameter Consortium (TPC) aims to bring together groups interested in collaborating on building, training, and using large-scale AI models and operating extensive computing systems. Formed in 2023, TPC includes communities focused on AI methods development, hardware and software systems design, and the application of AI to scientific and engineering challenges.

The TPC facilitates collaboration among HPC and AI experts, domain scientists, and HPC centers to maximize the use of limited resources and advance AI applications in science and engineering. TPC aims to grow the community, support career development, and enable scientists to leverage AI for research. Through events like this workshop, TPC promotes shared learning, innovation, and the formation of new collaborative working groups.

"Generative AI projects, from discipline-specific to multi-disciplinary models, offer significant opportunities for collaboration," said Fabrizio Gagliardi, BSC representative at TPC and coordinator of the event. "However, the high computation cost for training these models necessitates shared efforts. By addressing challenges like data sharing, licensing, and ethical AI development, TPC aims to reduce duplication of effort and accelerate advancements in AI."

Plenary Sessions

The event will kick off with an opening plenary session on June 19, 2024, from 15:00 to 17:45. This session will feature distinguished speakers including Rick Stevens (Associate Laboratory Director, Argonne National Laboratory) presenting "A Vision for AI in Science," Satoshi Matsuoka (Director, RIKEN Center for Computational Science) discussing "AI for Science at Riken and Japan," and Mateo Valero (Director, Barcelona Supercomputing Center) speaking on "AI and Architectures at BSC." Attendees can also participate in a hands-on tutorial on Large Language Models for Science and Engineering Applications on the first day of the meeting.

On June 20, 2024, plenary sessions will cover major international and national AI initiatives, with Jérôme Bobin (CEA) providing an update on the International Post-Exascale initiative (InPex), followed by an international panel on AI funding initiatives moderated by Fabrizio Gagliardi (BSC) and featuring panelists from the EU, US, and Japan. Later sessions will include insights from building large-scale AI models for science, with contributions from Ian Foster (Argonne National Laboratory/University of Chicago) and others. The day will conclude with parallel breakout sessions on topics such as model architecture, bioinformatics, community training, and AI hardware strategies.

On June 21, 2024, parallel breakout sessions will continue, covering data workflows, trust evaluation, model architecture, and healthcare applications. The event will close with a plenary session on trustworthy and responsible AI, featuring speakers like Mikel Rodriguez (Google DeepMind) and Bo Li (University of Chicago), followed by closing remarks from BSC hosts.

Founding Partners of TPC

AI Singapore, Allen Institute For AI, AMD, Argonne National Laboratory, Barcelona Supercomputing Center, Brookhaven National Laboratory, CalTech, CEA, Cerebras Systems, CINECA, CSC – IT Center for Science, CSIRO, ETH Zürich, Fermilab National Accelerator Laboratory, Flinders University, Fujitsu Limited, HPE, Intel, Juelich Supercomputing Center, Kotoba Technologies, Inc., LAION, Lawrence Berkeley National Laboratory, Lawrence Livermore National Laboratory, Leibniz Supercomputing Centre, Los Alamos National Laboratory, Microsoft, National Center for Supercomputing Applications, National Energy Technology Laboratory, National Institute of Advanced Industrial Science and Technology (AIST), National Renewable Energy Laboratory, National Supercomputing Centre, Singapore, NCI Australia, New Zealand eScience Infrastructure, Northwestern University, NVIDIA, Oak Ridge National Laboratory, Pacific Northwest National Laboratory, Pawsey Institute, Princeton Plasma Physics Laboratory, RIKEN, Rutgers University, SambaNova, Sandia National Laboratories, Seoul National University, SLAC National Accelerator Laboratory, Stanford University, STFC Rutherford Appleton Laboratory, UKRI, Texas Advanced Computing Center, Thomas Jefferson National Accelerator Facility, Together AI, Tokyo Institute of Technology, Université de Montréal, University of Chicago, University of Delaware, University of Illinois Chicago, University of Illinois Urbana-Champaign, University of New South Wales, University of Tokyo, University of Utah, University of Virginia.

More information about the European Kick-Off of the International Trillion Parameter Consortium (TPC) can be found here.

Source: BSC