Action-Driven Processes for Continuous-Time Control

Ruimin He, Shaowei Lin

At the heart of reinforcement learning are actions -- decisions made in response to observations of the environment. Actions are equally fundamental in the modeling of stochastic p…