Covers: theory of Linear Function Approximation
Estimated time needed to finish: 30 minutes
Questions this item addresses:
  • How to approximate the state and state-action value using Linear Function Approximation?
How to use this item?

It's highly recommended to watch the full lecture video to understand the concept of function approximation in RL.

Fail to play? Open the link directly: https://www.youtube.com/watch?v=buptHUzDKcE
Author(s) / creator(s) / reference(s)
Prof. Emma Brunskill
0 comment
Recipe
publicShare
Star(0)

Model-Free Function Approximation using Linear Functions and Deep Q-Networks

Contributors
Total time needed: ~3 hours
Objectives
Introduces function approximators in Reinforcement Learning to obtain compact representation that generalizes across states and actions in the environment.
Potential Use Cases
Strategic Games, Robotics
Who is This For ?
ADVANCEDAnyone who is interested in learning the advanced concepts and applications of Reinforcement Learning Algorithms
Click on each of the following annotated items to see details.
PAPER 1. Playing Atari with Deep Reinforcement Learning
  • What is Deep Q-learning?
  • How Deep Q-Learning works?
30 minutes
VIDEO 2. CNNs and Deep Q Learning
  • How to use Neural Networks as function approximators in Q-learning?
30 minutes
ARTICLE 3. Double Deep Q Networks
  • What is Double Q-Learning?
20 minutes
PAPER 4. Deep Reinforcement Learning with Double Q-learning
  • What is Double Q-Learning?
30 minutes
VIDEO 5. Value Function Approximation
  • How to approximate the state and state-action value using Linear Function Approximation?
30 minutes
ARTICLE 6. Understanding Q-Learning and Linear Function Approximation
  • How to approximate the value of state/action using Linear Function Approximation?
15 minutes

Concepts Covered

0 comment