Fumito Uwano


The University of Electro-Communications
Research Fellow of Japan Society for the Promotion of Science (DC1)

Abstract of my research
We study Multi-agent reinforcement learning (MARL) as solving problems which some autonomous agents cause by reinforcement learning. In particular, we examine the methods for the agents to learn cooperative policy as a fully decentralized MARL, and we attach importance to No-communication and theoretical method. No-communication is important to avoid communication delay and synchronization, and to guarantee the performance is important to establish the capability of methods.
Lately, I already proposed the two learning methods. The first method is called Profit Minimizing Reinforcement Learning (PMRL); it forces agents to learn how to reach the farthest goal, and then the agent closest to the goal is directed to the goal. The second method is called Yielding Action Reinforcement Learning (YARL); it forces agents to learn through a Q-learning process, and if the agents have a conflict, the agent that is closest to the goal learns to reach the next closest goal. The simulations performed on the maze problem for the agent cooperation task revealed that the two methods successfully enabled the agents to exhibit cooperative behavior, even if the size of the maze and the number of agents change.

Recent works
“Strategy for Learning Cooperative Behavior with Local Information for Multi-agent Systems,”
Proceedings of The 21st International Conference on Principle and Practice of Multi-Agent Systems, pp.663-670, 2018Download »
“Multi-Agent Cooperation Based on Reinforcement Learning with Internal Reward in Maze Problem,”
SICE Journal of Control, Measurement, and System Integration, Vol. 11, pp.321-330, 2018Download »
“Comparison Between Reinforcement Learning Methods with Different Goal Selections in Multi-Agent Cooperation,”
Journal of Advanced Computational Intelligence and Intelligent Informatics Vol.21 No.5 p. 917-929Download »

Subsidiary works

Space Informatics

“Recovery System Based on Exploration-biased Genetic Algorithm for Stuck Rover in Planetary Exploration,”
Journal of Robotics and Mechatronics Vol.29 No.5
“Analyzing Triangle Matching Method Based on Craters for Spacecraft Localization,”
Proceedings of The International Symposium on Artificial Intelligence, Robotics and Automation in Space, June, 2018Download »

Heart Rate Estimation (Well-being Computing)

“Ensemble Heart Rate Extraction Method for Biological Data from Water Pressure Sensor,”
Proceedings of 2018 AAAI Spring Symposium Series, p.304-309, May, 2018

Learning Classifier Systems

“Generalizing rules by random forest-based learning classifier systems for high-dimensional data mining,”
Proceedings of the Genetic and Evolutionary Computation Conference Companion, July, 2018

Opinion Sharing (Multi-agent Network)

“Weighted Opinion Sharing Model for Cutting Link and Changing Information among Agents as Dynamic Environment,”
SICE Journal of Control, Measurement, and System Integration, Vol. 11, pp.331-340, 2018Download »

Hello. I’m Fumito Uwano from the University of Electro-Communications.
I belongs to this laboratory in order to study many methods for Artificial Intelligence. Especially, I interested in multi-agent reinforcement learning.

Contact me
(when you send me an e-mail, please convert (at) to @)