Value-function Approximation
Source:: 2018 Reinforcement Learning An Introduction
- Extending Reinforcement Learning to function approximation also makes it applicable to partially observable problems, in which the full state is not available to the agent.
# Value-function Approximation
- Function approximation methods expect to receive examples of the desired input–output behavior of the function they are trying to approximate.
- Not all function approximation methods are equally well suited for use in reinforcement learning.
- Methods that are able to from online and nonstationary settings are preferable in Reinforcement Learning.