Learning to Communicate with Deep Multi-Agent Reinforcement Learning

TL;DR

The paper addresses learning communication protocols among multiple agents to maximize a shared utility in partially observable environments. It introduces two approaches: Reinforced Inter-Agent Learning (RIAL) using deep Q-learning and Differentiable Inter-Agent Learning (DIAL), which allows backpropagation through (noisy) communication channels, enabling centralised learning with decentralised execution. The authors demonstrate end-to-end learning of communication protocols in complex environments inspired by riddles and multi-agent computer vision tasks, and they highlight engineering innovations essential for success. These contributions advance the development of scalable, learnable inter-agent communication for coordinated decision-making under partial observability.

Abstract

We consider the problem of multiple agents sensing and acting in environments with the goal of maximising their shared utility. In these environments, agents must learn communication protocols in order to share information that is needed to solve the tasks. By embracing deep neural networks, we are able to demonstrate end-to-end learning of protocols in complex environments inspired by communication riddles and multi-agent computer vision problems with partial observability. We propose two approaches for learning in these domains: Reinforced Inter-Agent Learning (RIAL) and Differentiable Inter-Agent Learning (DIAL). The former uses deep Q-learning, while the latter exploits the fact that, during learning, agents can backpropagate error derivatives through (noisy) communication channels. Hence, this approach uses centralised learning but decentralised execution. Our experiments introduce new environments for studying the learning of communication protocols and present a set of engineering innovations that are essential for success in these domains.