Reasoning

Learning Context-aware Task Reasoning for Efficient Meta-reinforcement Learning

This work proposes a dual-agents reasoning strategy under the Variational EM framework to achieve efficient exploration in the meta-RL problem.