Fetching the paper…

Model-Based RL for Mean-Field Games is not Statistically Harder than Single-Agent RL · Around