Fetching the paper…

RoMFAC: A robust mean-field actor-critic reinforcement learning against adversarial perturbations on states · Around