Paper Published in “Machine Learning: Science and Technology” | 曽我部研究室 SogabeLab 国立大学法人電気通信大学│AI 人工知能 量子

A paper by Mr. Shinji Sato (former Master’s student, UEC) has been published in Machine Learning: Science and Technology (IOP Publishing).

“Randomized Tensor Network Reservoir Computing: Validity and Learnability Phase Transitions”
Shinji Sato, Daiki Sasaki, Chih-Chieh Chen, Kodai Shiba and Tomah Sogabe
Machine Learning: Science and Technology, Vol. 6, 035011 (2025) DOI 10.1088/2632-2153/aded56

The study introduces a new randomized tensor network reservoir computing (TNRC) method, and demonstrates both theoretically and experimentally how phase transitions in learnability near the edge of chaos emerge. TNRC also shows superior predictive performance compared to conventional echo state networks, paving the way for advances in quantum-inspired machine learning and complex systems research.

This marks the first publication from the Sogabe Laboratory in this journal.