Reinforcement Learning of Implicit and Explicit Control Flow in Instructions
Ethan A. Brooks 1 Janarthanan Rajendran 1 Richard L. Lewis 2 Satinder Singh 1
Abstract task instructions that require the agent to learn control flow
Learning to flexibly follow task instructions in dy- either because the instructions themselves include explicit
namic environments poses interesting challenges conditionals such as the if- and while- statements fa ...


雷达卡




京公网安备 11010802022788号







