Robust Multi-agent Communication Based on Decentralization-Oriented Adversarial Training

Open in new window