Multi-agent Reinforcement Learning

The article introduces Fact-based Agent modeling (FAM), a new method in multi-agent reinforcement learning. FAM uses ‘facts’ – rewards and observations obtained by agents after taking actions – as a…