Kosuke Yamamoto

Papers

Total Citations

H-Index

About

Kosuke Yamamoto is a researcher in reinforcement learning and robotics, with a focus on developing efficient algorithms for high-dimensional dynamical systems. His most cited work, "Path Integral Policy Improvement With Population Adaptation" (2020, 12 citations), introduces a novel approach to path integral policy improvement (PI²) that enhances sample efficiency and adaptability in complex control tasks. This contribution addresses a critical challenge in reinforcement learning: scaling policy optimization to systems with many degrees of freedom, such as robotic manipulators or autonomous vehicles. Yamamoto’s method leverages population-based adaptation to dynamically adjust exploration and learning parameters, enabling faster convergence and more robust performance compared to traditional PI² variants. His work bridges theoretical advances in stochastic optimal control with practical applications in robotics, offering a pathway to more autonomous and adaptive systems. With 12 citations, this paper has already influenced subsequent research in adaptive reinforcement learning, particularly in domains requiring real-time decision-making under uncertainty. Yamamoto’s ongoing contributions continue to push the boundaries of how machines learn and interact with complex environments, making him a notable voice in the intersection of machine learning and control theory.