政策・メディア研究科 中森 辰洋

構成可能システムの安全で高速な自律的最適化

My research develops an automatic optimization system for configurable systems that have many parameters: AI inference systems, database management systems, etc. It focuses on making these systems faster, safer, and adaptive to new environments. By reusing data from similar settings, the system can find good configurations quickly without causing failures. This work aims to make complex systems run efficiently on their own, reducing manual tuning and improving reliability across AI platforms and cloud.