Flappy Bird AI Documentation
Welcome to the detailed documentation for the Flappy Bird AI project. Here you can dive deep into how the game and the artificial intelligence work together.
📚 Documentation Sections
Understand the brain of the bird.
- Input layers and state normalization.
- Hidden layers and activation functions.
- Output actions.
How do we know if the AI is learning?
- The Reward System (+0.1 vs -100).
- Epsilon-Greedy Strategy.
- Defining success.
The physics and logic behind the game.
- Phaser 3 Arcade Physics.
- Collision detection.
- Frame skipping for training optimization.
The tools used to build this project.
- Phaser 3
- TensorFlow.js
- Vite
What can this project teach you about Machine Learning?
- The Agent-Environment Loop.
- Exploration vs. Exploitation.
- Why “Toy Problems” are important.
← Back to Repository