ANN Zoo
Pigs learn to fly
Here we have a flock of pigs learning to play the classic game Flappybird . Each pig tries to learn to fly through a series of pipes. At each moment, the pigs must decide to fly higher or not. If the pigs hit the top or the bottom of the course, they stop flying. The pigs are rewarded for making it as far through the course as they can.
Once all the pigs have stopped flying, then they learn a little and try again. You can speed up the learning process by increasing the Speed hyperparameter. A genetic algorithm is used to select pigs for the next flying lesson based on how far they made it in the last lesson.
Stats
Learning pigs:High score: 0
All time high score: 0
Actions
Options
The brain of each pig is a neural network with inputs selected below (in the Inputs section) and a single output which is whether to flap right now or not. The pigs use their brains for every frame. The network can be visualized with neurons as circles and the connections between those neurons as lines. The thicker the lines, the stronger the connection. Black connections are positive , and red connections are negative.
The bars to the left of the input neurons indicate what the first pig 'sees' right now.
Hyperparameters
Speed:Size:
Neurons:
Mutation rate:
Mutation size:
Inputs
Here you can select what each pig gets to 'see'. If you deselect an input, then the lesson will start over for all the pigs.