BADGER: Learning to (Learn [Learning Algorithms] through Multi-Agent Communication)

TL;DR

This paper introduces BADGER, a multi-agent meta-learning system where agents learn to communicate and adapt rapidly to new environments by developing shared communication policies, enabling generalization beyond existing methods.

Contribution

It presents a novel memory-based multi-agent meta-learning architecture that facilitates learning communication policies for rapid adaptation and generalization in unseen environments.

Findings

Emergence of rapid adaptation through learned communication

Shared policy enables modular and scalable learning

Improved generalization over existing meta-learning approaches

Abstract

In this work, we propose a novel memory-based multi-agent meta-learning architecture and learning procedure that allows for learning of a shared communication policy that enables the emergence of rapid adaptation to new and unseen environments by learning to learn learning algorithms through communication. Behavior, adaptation and learning to adapt emerges from the interactions of homogeneous experts inside a single agent. The proposed architecture should allow for generalization beyond the level seen in existing methods, in part due to the use of a single policy shared by all experts within the agent as well as the inherent modularity of 'Badger'.