My love for AI and Machine Learning.

What it does

Neural Networks learn to play Pong using NEAT (natural selection).

The input layer neurons correspond to: (The attributes visible to the neural networks):

  • Ball X position
  • Ball Y position
  • Ball's angle of travel (2 neurons)
  • Paddle's Y position

The neural networks are paired, face each other for ten seconds, reproduce based on their score.

In this demo, a short animation will be played through, showcasing its features.

How I built it

C++, SDL for images

Challenges I ran into

Conflicting image and render types could not be put on the screen at the same time which put a damper on things

Accomplishments that I'm proud of

It produces working neural networks, the Header I created for the neural networks is contained in a nice template for future, and it's cute, funny, and educational.

What I learned

Even 24 hours isn't enough sometimes.

What's next for Love Between Neural Networks

Better and faster learning!

Built With

Share this project: