Naoki Kotani

Osaka Institute of Technology, The University of Osaka

Papers

Total Citations

H-Index

About

Naoki Kotani is a researcher specializing in reinforcement learning, transfer learning, and autonomous robotics, with a focus on improving the efficiency and scalability of learning systems. His work addresses critical challenges in enabling agents to reuse knowledge across tasks, reducing computational overhead, and accelerating real-time adaptation. Kotani’s most-cited paper, “Knowledge Selection based on Value in Transfer Learning” (2015, 5 citations), introduces a method to selectively transfer only valuable knowledge, preventing performance degradation in multitask reinforcement learning. His earlier contribution, “A Novel Clustering Method Curbing the Number of States in Reinforcement Learning” (2009, 4 citations), enhances state-space construction by refining Fuzzy ART clustering, controlling category growth to reduce complexity. More recently, his 2020 paper on differential-wheeled mobile robots demonstrates how knowledge transfer cuts learning time for real-time robotic control. Though his citation counts are modest, Kotani’s work is foundational for researchers tackling the curse of dimensionality in reinforcement learning and practical deployment in robotics. His focus on value-based knowledge selection and state-space efficiency offers pragmatic solutions for autonomous systems, making his research valuable for students and engineers building adaptive, resource-constrained agents.