Photonic information processing based on wave chaos and dynamical deep learning

2022 Fiscal Year Final Research Report

• Project/Area Number: 20H04255
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Single-year Grants
• Section: 一般
• Review Section: Basic Section 61040:Soft computing-related
• Research Institution: Kanazawa University
• Principal Investigator: Sunada Satoshi 金沢大学, 機械工学系, 教授 (10463704)
• Co-Investigator(Kenkyū-buntansha): 原山 卓久 早稲田大学, 理工学術院, 教授 (70247229)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: リザバーコンピューティング / 波動カオス / 動的深層学習 / 物理深層学習

Outline of Final Research Achievements

The purpose of this study is to realize an information processing system based on high-dimensional dynamical systems involving wave chaos. The two main results are as follows: (i) We developed a large-capacity photonic reservoir chip based on a wave chaotic microcavity and demonstrated high-speed processing at a rate of 12.5 GS/s. (ii) We also developed a theoretical framework that enables dynamical systems to perform deep learning-like information processing based on its optimal control. We applied this theory to optoelectronic systems with time delays and showed that pattern recognition is possible with a few control signals.

Free Research Field

光情報処理、非線形物理

Academic Significance and Societal Importance of the Research Achievements

本研究で開発した微小共振器構造に基づく光リザバーチップは、微小領域にて大容量のリザバー計算を高速かつ高効率に実行できる。さらに最適制御理論に基づき制御を合わせることで、リザバー計算の枠組みを超えた高度な情報処理が可能となる。これにより、AI処理の超高速化や省エネ化に貢献し、且つこれまで捉えることのできなった高速現象の異常検知・認識への応用も期待できる。また、光通信や光計測分野をはじめとしたさまざまな分野への波及も期待できる。