Learning Communication Skills in Multi-task Multi-agent Deep Reinforcement Learning

Changxi Zhu, Mehdi Dastani, Shihan Wang; 7, 2026.

Presented at the Reinforcement Learning Conference (RLC), Montréal, Quebec, Canada, August 15–17, 2026.

Abstract

In multi-agent deep reinforcement learning (MADRL), agents can communicate with one another to perform a task in a coordinated manner. When multiple tasks are involved, agents can also leverage knowledge from one task to improve learning in other tasks. In this paper, we propose Multi-task Communication Skills (MCS), a MADRL method with communication that learns and performs multiple tasks simultaneously, with agents interacting through learnable communication protocols. Specifically, MCS adopts a task-invariant communication protocol that enables communication under varying numbers of agents and tasks with different observation and action spaces. To enhance coordinated behaviors among agents, we further correlate communicated messages to agents' policies for each task. We adapt three existing multi-agent benchmark environments to multi-task settings. Empirical results demonstrate that MCS achieves better performance than multi-task MADRL baselines without communication, as well as single-task MADRL baselines with and without communication.

[abs][pdf]

BibTeX

@article{zhu2026learning,
    title={Learning Communication Skills in Multi-task Multi-agent Deep Reinforcement Learning},
    author={Changxi Zhu and Mehdi Dastani and Shihan Wang},
    journal={Reinforcement Learning Journal},
    volume={7},
    pages={},
    year={2026}
}