RIKEN and Argonne National Laboratory sign MOU on AI for Science

In the United States-Japan Joint Leaders’ Statement released today, the collaboration between RIKEN and Argonne National Laboratory on AI for Science was welcomed.

On April 5, the two institutes signed a Memorandum of Understanding (MOU) for carrying out joint research in the area of AI for Science. RIKEN President Makoto Gonokami and ANL Director Paul Kearns signed the MOU at an online ceremony also attended by ANL Associate Director Rick Stevens, RIKEN Executive Director Makiko Naka, RIKEN Center for Computational Science Director Satoshi Matsuoka, and RIKEN TRIP Headquarters Advanced General Intelligence for Science Program Program Director Makoto Taiji.

In addition, on April 9, the Ministry of Education, Culture, Sports, Science and Technology of Japan and the Department of Energy of the USA revised their project arrangement (PA) concerning High Performance computing and Artificial Intelligent and placed AI for Science in the Japan -US governmental collaboration framework.

RIKEN and ANL have taken a central role in the AI for Science framework of Japan-US governmental framework.

The MOU will involve collaboration though means such as the Exchange of scientific and technological information, including possibly the joint development and validation of models, exchange of information on model learning and use techniques; Exchange of data sets for research; Exchange of computing resources for research, and Collaborative Cooperation on research and roadmap exercise; Exchange of researchers and staff; Exchange of scientific and technological information, including delivering lectures and holding joint seminars and symposia; and Exchange of students (in particular, doctoral students).

RIKEN and Argonne National Laboratory have agreed to establish a cooperative relationship to support AI for Science at the two institutions through the development and evaluation of new computing system software and science and engineering applications, including generative model and other AI, application, operation of computing system, data management, AI architecture and remote methodologies.

Under this AI for Science objective, the two institutes plan to conduct research and development of domain-oriented AI infrastructure models for scientific research.

Based on the objective of AI for Science, the two institutes will promote research and development of AI infrastructure models for scientific research oriented to specific scientific domains using various scientific research data. They will carry out research and development on applications for specific areas such as life science, materials science, the two organizations will also work on research and development in fundamental technologies, including for data generation, automation and acceleration of related experiments and simulations, performance evaluation of foundation models, system software development, data management technology and computing system operation technology. They will also share world-leading research trends and provide mutual use of their world-class computing resources and research data sets, including RIKEN's supercomputer Fugaku and Argonne National Laboratory's Aurora, in order to accelerate research and development by leveraging each other's strengths.

The United States is a frontrunner in this research field, and the Department of Energy has launched a new project on AI for Science, with Argonne National Laboratory being one of the core institutions. RIKEN, for its part, is leading Japan in this research field with the launch in FY2024 of the TRIP-AGIS project, which aims to develop foundation models for scientific research. This collaboration between the two institutions is intended to enable Japan and the U.S. to take the lead in promoting AI for Science research, a highly competitive area worldwide. RIKEN will work with Argonne National Laboratory to advance AI for Science by leveraging RIKEN's strengths in diverse scientific knowledge and data, as well as its cutting-edge research environment, including Fugaku.

The Signing Ceremony

(From left: RIKEN President Makoto Gonokami, Argonne Director Paul Kearns)

(From left: RIKEN Executive Director Makiko Naka, RIKEN President Makoto Gonokami, RIKEN Center for Computational Science Director Satoshi Matsuoka, RIKEN TRIP Headquarters Advanced General Intelligence for Science Program Program Director Makoto Taiji)

Quote from RIKEN President Makoto Gonokami

RIKEN and Argonne National Laboratory have had a number of deep discussions on how to deepen our relationship, starting with a top-level meeting last October and continuing through a workshop on specific ideas for collaboration between the two institutions. I would like to thank Argonne National Laboratory for making it possible to sign this MOU based on these close exchanges.

With the signing of this MOU, the AI for Science collaboration between the two institutions has officially begun. RIKEN will collaborate with Argonne National Laboratory under our concept of Transformative Research Innovation Platform of Riken Platforms (TRIP), which aims to organically link RIKEN's leading-edge platforms in various fields.

Generative AI is creating many changes in our society in a variety of ways. Ensuring that generative AI can contribute to the happiness of all humankind is a responsibility of today’s researchers. RIKEN and ANL will place our faith in the power of science to continue to carry out research in pursuit of universal truths, and under the framework of Japan-US relations, will contribute to the development of AI for Science around the world by leveraging our mutual strengths through diverse efforts that transcend national borders.

Quote from ANL Director Paul Kearns

We have formed a partnership that can make pivotal discoveries and transformative advancements. We’re addressing some of the most complex challenges facing society, and I’m confident that we can achieve an even bigger impact by working together.

About AI for Science

RIKEN’s Initiative for AI for Science

Inquiry

For inquiries, please contact
RIKEN International Affairs Division
Email: iad-collab [at] riken.jp