Using Open AI Gym library to solve Atari Pong game by implementing Reinforcement Learning techniques
Table of contents
Open Table of contents
Atari Pong Environment
Pong is a two-dimensional sports game that simulates table tennis. The player controls an in-game paddle by moving it vertically across the left or right side of the screen.
Players use the paddles to hit a ball back and forth. The goal is for each player to reach eleven points before the opponent;points are earned when one fails to return the ball to the other
- Actions - There are 6 actions noop, fire, right, left, rightfire, leftfire
- Reward - 10 points for passing the ball to the opponent, minus points for missing the ball
- Goal – Make the agent play pong
Algorithm
The model is trained using the Actor Critic method.
Actor critic:
Actor critic combines both value-based and policy-based algorithms, here we have both networks actor and critic the actor-network is used to perform an action for a given state it chooses the best action. The critic network calculates the Q value for that state action and helps the actor to improve the policy for selecting the action
Screenshots
- Final trained model playing the pong game:
Read more - Colab Notebook