wy Lv3

Function approximation in reinforcement learning (And deep reinforcement learning)

Value Function Approximation

Agent state update

Classes of Function Approximation

(Deep) neural nets often perform quite well, and remain a popular choice

Gradient-based algorithms

Gradient Descent | Approximate Values By Stochastic Gradient Descent

Linear function approximation

Feature Vectors

Feature construction example’:’ coarse coding

Convergence and divergence

Batch Reinforcement Learning

Deep reinforcement learning

DQN


Planning and models

Model-Based RL

Learn a model from experience

Plan value functions using the learned model.

Model

……….

  • Title:
  • Author: wy
  • Created at : 2023-07-23 15:58:45
  • Updated at : 2024-07-11 22:05:30
  • Link: https://yuuee-www.github.io/blog/2023/07/23/RL/step5/RLstep5/
  • License: This work is licensed under CC BY-NC-SA 4.0.
Comments
On this page