After starting some research into machine learning, I wanted to have a go at putting what I had learnt into action. My project was just a simple demonstration of genetic learning. Each capsule is a hover car which detects walls on either side and front. Each generation changes the distance and strength of reactions towards each wall it detects.
As tends to be the way, I had a lot of fun putting this together. While it took me a while to figure out what the cars would actually be 'learning', once I got the hang of it I found myself just watching the cars learn on different tracks, seeing what kind of weird stuff they could learn to drive - or hover - on.
Comments