Skip to content

BH4/Deep-Reinforcement-Learning

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Deep-Reinforcement-Learning

My attempt to create a single AI that can play several games.

Inspired entirely by this paper https://arxiv.org/abs/1312.5602 about playing Atari games with the technique named Deep Q Learning. I also started my code by studying and directly editing the project at https://gist.github.com/EderSantana/c7222daa328f0e885093.

In addition to the experience replay mechanism, I have added target networks and double DQN methods to the Deep Q-learning implemented here. Dueling DQN and Prioritized Experience Replay are potential next additions.

Currently this implementation successfully learns the simple catch game from the original project as well as Pong. I previously had a functioning implementation of tic-tac-toe and a functional game of snake (which was having more difficulty), but haven't updated their structure to fit the current one yet. Unfortunately lack of time (mostly my PhD and subsequent job) has delayed this and changes to Tensorflow mean coming back to the project will take even more time.

In the meantime here is the single gif I have of the trained pong game from before my environment stopped working. pong_target_1000

About

My attempt to create a single AI that can play several games.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages