Crate artha

A dead simple neural network built as a learning exercise.

Getting Started

    let xs = matrix![[2.,9.],[1.,5.],[3.,6.]];
    let max = xs.max().unwrap();
    let xs: Matrix<f64> = xs.iter().map(|a|vec![a[0] / max[0], a[1] / max[1]]).collect();
    let ys = matrix![[92.], [86.], [89.]];
    let ys: Matrix<f64> = ys.iter().map(|a|vec![a[0] / 100.]).collect();
    let mut nn = NeuralNetwork::new(2,1,vec![3]);
    logln!("Input: ", xs);
    logln!("Actual Output: ", ys);
    let predicted = nn.train(&xs, &ys, 100000);
    logln!("Predicted Output: ", predicted);
    logln!("Loss: ", predicted.mse_diff(&ys).unwrap());

This program is a direct translation of https://dev.to/shamdasani/build-a-flexible-neural-network-with-backpropagation-in-python into rust.

Due to my naive custom Matrix implementation, this network is significantly slower that the tutorial. I'll definitely look into optimizing matrix opeartions and other segments of my code. Also checkout 3Blue1Browns's excellent series on Neural Network https://www.youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi

Besides optimization, I am also hoping to implmenting a network that recognized handwritten digits and who knows what else from there. But for now, this is a fairly inaccurate rookie version that I could build on my own.

• If you have any questions or suggestions, feel free to submit issues, or contact me in other ways.
• If you found my sub-par rust skills offensive, please do provide some constructive criticism.

Re-exports

pub use self::matrix::Matrix;

Modules

matrix

A 2D Matrix system to make our neural network implementaion easier.

Macros

debug	Removes the need for specifying the debug format string in print!.
debugln	Removes the need for specifying the debug format string in println!.
log	Removes the need for specifying the display format string in println!.
logln	Removes the need for specifying the display format string in print!.
matrix	Creates a matrix containing the arguments.

Structs

NeuralNetwork

A dead simple Neural Network