Рет қаралды 1,718,105
Welcome to the Programmed Universe!
In this video you will see how an Artificial Intelligence learned to play Flappy Bird!
An artificial 3-layer Perceptron Multilayer was used!
✅ - Input layer with 4 sensors (Horizontal Distance, Vertical Distance, Pipe Speed, Pipe Opening) + 1 Bias, totaling 5 Neurons.
✅ - Hidden Layer with 6 neurons + 1 Bias, totaling 7 Neurons
✅ Output Layer with 2 neurons (Jump, Parachute)
The activation function used in all neurons was ReLU.
The learning method is commonly called '' Random Mutations ''.
The population size I used was 2000 individuals
The learning time ranged from 0 ~ 30 minutes per Game mode.
The language used throughout the process was Language C.
Code + Executable - github.com/JVictorDias/FlappI...
❤️ - A big hug and a great study for you!