Bebble Learns

Bebble Logo
Generated Pie Chart
Home Page
Game Screen
Statistics Page

Inspiration

We were inspired by a concept I explored in High School which we then expanded upon to create machine learning in a game.

What it does

The AI, named Bebble, should learn the optimal moves in a game of 11 consistently when the only choices are incrementing by one or two. He will then verse a player and will find out who is smarter.

How we built it

The program puts two bots against each other in a game of 11. Each choice is random and the winner will have their choices saved in a global array which we then referenced later to find the optimal choice for every situation encountered, through the process of machine learning. After the bots verse each other thousands of times, Bebble should win about 90% of the time against a randomized choice bot.

Challenges we ran into

We ran into multiple challenges. The most prominent one being, storing the data and having the machine use the data to learn. We also had some challenges accessing the correct data values in the ray at a specified index. We also had trouble designing the GUI and connecting all of the scenes together through buttons.