NeuralWorks

NeuralWorks is a lightweight, modular JavaScript library designed to create and train neural networks directly in the browser or on Node.js environments. With a focus on simplicity and flexibility, NeuralWorks empowers developers to build neural networks without the need for complex dependencies.

Features

• Easy to Use: Simple API for building, training, and predicting with neural networks.

• Modular: Easily extendable for custom neural network designs.

• Client-Side Support: Works directly in the browser using vanilla JavaScript, as well as on Node.js.

• Sigmoid Activation: Supports the sigmoid activation function for smooth output scaling.

• Backpropagation: Trains networks using the backpropagation algorithm.

Installation

To install NeuralWorks in your project, run the following command:

npm  install  neuralworks

Project Structure

The project is organized as follows:

• src/: Contains the source code for the library.

  • index.js: Main file containing the core functionality of the library.

• LICENSE: The open-source license for the project.

• README.md: The documentation file you are currently reading, providing details about the library and usage.

• package.json: Defines project metadata, dependencies, and scripts.

Usage

Once installed, you can import and start using the library in your JavaScript or TypeScript code.

Example Usage

const  NeuralNetwork = require('neuralworks');

  

// Create a neural network instance

const  nn = new  NeuralNetwork(2, 4, 1); // 2 input nodes, 4 hidden nodes, 1 output node

  

// Training data (XOR Problem)

const  inputs = [

[0, 0],

[1, 0],

[0, 1],

[1, 1]

];

  

const  targets = [

[0], // Expected output for [0, 0]

[1], // Expected output for [1, 0]

[1], // Expected output for [0, 1]

[0] // Expected output for [1, 1]

];

  

// Training the network

for (let  i = 0; i < 10000; i++) {

const  index = Math.floor(Math.random() * inputs.length);

nn.train(inputs[index], targets[index]);

}

  

// Testing the network

console.log(nn.computeOutput([1, 0])); // Expected output: ~1 (close to 1 for XOR problem)

console.log(nn.computeOutput([0, 0])); // Expected output: ~0

API

NeuralNetwork(inputNodes, hiddenNodes, outputNodes, learningRate?)

Constructor for creating a neural network.

• inputNodes (number): Number of input nodes.

• hiddenNodes (number): Number of hidden nodes.

• outputNodes (number): Number of output nodes.

• learningRate (number, optional): Learning rate (default is 0.01).

computeOutput(inputArray)

• inputArray (array): Array of input values (e.g., [0, 1]).

• Returns: An array of output values, typically between 0 and 1.

train(inputArray, targetsArray)

• inputArray (array): Array of input values.

• targetsArray (array): Array of expected output values.

• Purpose: Trains the network using backpropagation.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

1. Fork the repository.

2. Create your feature branch (git checkout -b feature/your-feature).

3. Commit your changes (git commit -am 'Add new feature').

4. Push to the branch (git push origin feature/your-feature).

5. Create a new Pull Request.

---

Made by David Estrin.