In multi-agent reinforcement learning (MARL), it is desired to achieve the overall cooperation in a decentralized manner without compromising each agent' s ...

Agents that Learn to Vote for a Joint Action Through Multi-Agent Reinforcement Learning ; Natsuki Matsunami at Nagoya Institute of Technology · Natsuki Matsunami.

2020 9th International Congress on Advanced Applied Informatics (IIAI-AAI) | 978-1-7281-7397-9/20/$31.00 ©2020 IEEE | DOI: 10.1109/IIAI-AAI50415.2020.00173 ...

Natsuki Matsunami, Shun Okuhara, Takayuki Ito: Agents that Learn to Vote for a Joint Action Through Multi-Agent Reinforcement Learning.

We evaluate the learning performance of multiple independent learning agents inter- acting in an iterative plurality voting game, which is a competitive game.

Jul 2, 2019 · The considered MARL system selects the action to take according to the votes from local agents. Each agent determines its vote individually ...

In this paper, we use a basic multiarmed bandit style reinforcement learning al- gorithm [18] for testing whether agents can learn to make a collective decision ...

Abstract. In this paper we tackle the challenge of Multi Agents Reinforcement. Learning (MARL) in a situation of collective social choice. We evaluate.

ABSTRACT. Multi-agent reinforcement learning typically suffers from the prob- lem of sample inefficiency, where learning suitable policies involves.

These agents are called Independent Learners. (ILs). In the second category the agents learn joint actions and they are called Joint. Learners (JLs) 3.