Meta-Reinforcement Learning Using Model Parameters

Gabriel Hartmann, Amos Azaria

In meta-reinforcement learning, an agent is trained in multiple different environments and attempts to learn a meta-policy that can efficiently adapt to a new environment. This pap…