Train a neural network to play QPong

[HuangJunye]

Abstract

Train a neural network to play QPong, using genetic algorithm, similar to this project: https://github.com/greerviau/SnakeAI

Description

QPong is a quantum version of Pong made by @HuangJunye 's team in Qiskit Camp 2019. The idea of this project is to train a neural network to play QPong. Here are some ideas:

1. Train a neural network to play classical Pong against normal computer player.
2. Train a neural network to play quantum Pong against normal computer player.
3. Let the classical Pong AI compete with quantum Pong AI

Hopefully there will be some insights to be gained from QPong AI. Even if there will no none, it will be certainly fun to watch the AI playing QPong. :)

Members

• Andrew Tan
• Kyle Chu
• Lee Jun Yao Francis
• Brandon Hsu
• Qiskit Coach: @HuangJunye, Slack: @Junye

Deliverable

• A trained neural network for playing CPong and QPong
• (Optional) a cool Youtube video of the evolution of training similar to this: https://www.youtube.com/watch?v=zIkBYwdkuTk&t

GitHub repo

[fustilio]

Let's do this

[ktnc985]

I'm in

[andrewtanJS]

I'm in!

[brandhsu]

^