Specially Appointed Researcher Keiichi Tamai, Specially Appointed Associate Professor Tsuyoshi Okubo, and Professor Synge Todo of the Graduate School of Science at The University of Tokyo in collaboration Aisin Corporation find that scaling laws apply to deep neural networks exhibiting absorbing phase transitions.

Researchers have demonstrated that universal scaling laws, which describe how the properties of a system change with size and scale, apply to deep neural networks that exhibit absorbing phase transition behavior, a phenomenon typically observed in physical systems. The discovery not only provides a framework describing deep neural networks but also helps predict their trainability or generalizability. The findings were published in the journal Physical Review Research.

See below for more information.

• Todo Group：https://exa.phys.s.u-tokyo.ac.jp/
• UTokyo FOCUS：https://www.s.u-tokyo.ac.jp/en/press/10865/
• Aritcle URL：https://journals.aps.org/prresearch/abstract/10.1103/jp61-6sp2