lo-fi_swipe_brick_breaker_ML

Inspiration

Joe is addicted to a game called swipe brick breaker so we tried to beat him using a machine learning model

What it does

Uses a machine learning model to play the game swipe brick breaker

How we built it

We used processing to build the graphics interface, with arraylists used to store the bricks and the circles, which each give another ball to shoot. We made an AI through machine learning. The AI is able to learn which angle is the best to shoot based on the game state. The neural network we created was optimized through a genetic algorithm which optimized the weights to use for the network. This had a population of 100 with a 0.01 mutation rate.

Challenges we ran into

Graphics were rather difficult to implement and oftentimes would not smoothly run. The program is also pretty slow when simulating 100 different individual neural networks