METTA: Memory-Enhanced Test-Time Adaptation for Meta-RL
A proposed memory-enhanced architecture for meta-reinforcement learning to adapt quickly to unseen tasks across long histories.
A proposed memory-enhanced architecture for meta-reinforcement learning to adapt quickly to unseen tasks across long histories.