murari-goswami / -the-classic-1980-atari-game-boxer-. Goto Github PK

Okay so this is also something that I have been working on for sometime. This is basically a slight improvement on the classic 1980 Atari game "boxer". I have used Recurrent neural networks to perform Reinforcement learning, a method for enchancing the performance of an LSTM trained on data using Reinforcement Learning (RL). We create an RL reward that teaches the model to follow certain rules, while allowing it to retain information learned from data. The methodology here is to reinforce the score and not to influence the machine on HOW it is reached. The scores improve with each episode of the game. I have also invoked the Rocky Balboa 8 bit theme song(since the game is called boxer 😁) with it..to provide that 1980s feel😅.. I have increased the frame rate btw. Note: all of this is done using my CPU cores.