Ryubu Hosoki and Kento Sato (High Performance Big Data Research Team) Receive The CCGrid2025 Best Paper Award

Ryubu Hosoki, a student trainee, and Dr. Kento Sato, Team Principal of the High Performance Big Data Research Team at RIKEN R-CCS, have been honored with the Best Paper Award at CCGrid2025.

The CCGrid Best Paper Award is presented at the IEEE International Symposium on Cluster, Cloud and Internet Computing (CCGrid) to recognize outstanding research papers that make innovative and practical contributions in areas such as distributed systems, cloud computing, and internet computing.

Their award-winning research addresses the significant challenge of communication latency in large-scale AI model training, which requires immense computation and data exchange. They developed a novel method called "Comm-Shift" to optimize communication efficiency, thereby accelerating the training process. This advancement enables tackling more extensive and complex problems in applications like generative AI and simulations, promoting the practical implementation and societal integration of AI technologies.

This research was conducted as part of an international collaboration with Tokyo Institute of Science and the French Alternative Energies and Atomic Energy Commission (CEA).