In contrast to the already successful approaches in litera-ture, we have used hand-crafted features with popular re-inforcement learning algorithms to make the agent learn how to play the game of Breakout. Having domain specif-ic constraints, we recommend anyone to refrain them-selves from considering this approach as the results are not convincing as compared to the already existing tech-niques.
rohit57/Breakout
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|