CoDreamer: Communication-Based Decentralised World Models

Edan Toledo, Amanda Prorok

Sample efficiency is a critical challenge in reinforcement learning. Model-based RL has emerged as a solution, but its application has largely been confined to single-agent scenari…