# Decentralization of Multiagent Policies by Learning What to Communicate

## TL;DR

This paper introduces a neural network-based system enabling multiagent teams to learn effective, task-specific communication protocols from expert demonstrations, improving coordination in dynamic environments.

## Contribution

It proposes a novel architecture and training method for agents to autonomously learn communication semantics from expert policies, reducing manual design effort.

## Key findings

- System handles varying team sizes effectively.
- Performance degrades gracefully under communication constraints.
- Successfully applied to a perimeter defense game.

## Abstract

Effective communication is required for teams of robots to solve sophisticated collaborative tasks. In practice it is typical for both the encoding and semantics of communication to be manually defined by an expert; this is true regardless of whether the behaviors themselves are bespoke, optimization based, or learned. We present an agent architecture and training methodology using neural networks to learn task-oriented communication semantics based on the example of a communication-unaware expert policy. A perimeter defense game illustrates the system's ability to handle dynamically changing numbers of agents and its graceful degradation in performance as communication constraints are tightened or the expert's observability assumptions are broken.