Cooperative learning AAAI 2024
FoX: Formation-aware exploration in multi-agent reinforcement learning
UNIST · Machine Learning & Intelligent Control Lab
Read the paper ↗RESEARCH OVERVIEW
Guide exploration through diverse team formations in partially observable cooperative tasks.
FoX organizes exploration around formations to reduce redundant search. Agents learn to recognize their formation from local observations and explore meaningful alternatives, with evaluations in sparse SMAC and Google Research Football tasks.
Publication details & source ↗