NEAT JS

This is a NEAT implementation originating from Code Bullet's demonstration on YouTube

Getting Started

Currently the project is a front-end NEAT implementation in JavaScript for demonstration purposes. The project uses the P5.js library to draw a flappy-bird game and test the NEAT implemenation.

How to start the evolution process:

1. Navigate to the src/ folder using your terminal
2. Create a server to host the project with python -m SimpleHTTPServer
3. Open a browser and type http://127.0.0.1:8000

SimpleHTTPServer expects there to be an index.html file in the root directory to serve as the initial file

Interacting with the implementation

Open your browser's console

As of de5da3149a78c7b19bed8e53d376893540d3eb66 the evolution process is hardcoded to use a dataset to evolve, because there are no visible agents, fitness output is being logged to the console for a four input XOR dataset. The network within the blue flappy-bird background does update to reflect changes in the topology of the network.

How to format input / output data to evaluate fitness

The format for the dataset used internally to evaluate fitness and evolve is the same as Neataptic and Carrot

The dataset is a series of objects with two keys, input and output which correspond to the inputs and outputs of the neural network. Currently the inputs are hardcoded at 4 and the output is hardcoded at 1. Below is a 4 input XOR gate for demonstration.

    // XOR is true only if exactly one of the inputs is true
    dataset = [
      { input: [0, 0, 0, 0], output: [0] },
      { input: [0, 0, 0, 1], output: [1] },
      { input: [0, 0, 1, 0], output: [1] },
      { input: [0, 0, 1, 1], output: [0] },
      { input: [0, 1, 0, 0], output: [1] },
      { input: [0, 1, 0, 1], output: [0] },
      { input: [0, 1, 1, 0], output: [0] },
      { input: [0, 1, 1, 1], output: [0] },
      { input: [1, 0, 0, 0], output: [1] },
      { input: [1, 0, 0, 1], output: [0] },
      { input: [1, 0, 1, 0], output: [0] },
      { input: [1, 0, 1, 1], output: [0] },
      { input: [1, 1, 0, 0], output: [0] },
      { input: [1, 1, 0, 1], output: [0] },
      { input: [1, 1, 1, 0], output: [0] },
      { input: [1, 1, 1, 1], output: [0] },
    ];