FlappyLearning and Machine-Learning-Flappy-Bird
These two programs are competitors, both implementing machine learning (specifically neuroevolution and neural networks with genetic algorithms) to teach a computer to play Flappy Bird, offering alternative approaches to the same problem.
About FlappyLearning
xviniette/FlappyLearning
Program learning to play Flappy Bird by machine learning (Neuroevolution)
Implements a genetic algorithm framework (NeuroEvolution.js) that evolves neural network topologies through selection, mutation, and breeding across generations, with configurable population dynamics and elitism rates. The Flappy Bird environment serves as the fitness evaluation domain, where networks receive pixel-based input and output control signals. Built as a browser-based JavaScript implementation enabling real-time visualization of agent learning progression.
About Machine-Learning-Flappy-Bird
ssusnic/Machine-Learning-Flappy-Bird
Machine Learning for Flappy Bird using Neural Network and Genetic Algorithm
Implements neuro-evolution by combining a 3-layer neural network (2 input neurons sensing gap distance/height, 6 hidden neurons, 1 output neuron) with a genetic algorithm that performs selection, crossover, and mutation across generations. Built entirely in HTML5 using the Phaser game framework and Synaptic neural network library, with fitness calculated as distance traveled minus proximity to obstacles.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work