Regularization

Overview


The Javascript neural network api provides functions for adding regularization to a model. The following demonstrate sample code using regularization.

L1


L1 regularization is provided by the affineL1 function 


let ba = await import('/lib/backpropagation/v1.0.0/backpropagation.mjs');

let layers = [];
layers.push(ba.affineL1({
    inputs:20,
    outputs:5,
    lambda:0.1
}));

L2


L2 regularization is provided by the affineL2 function 


let ba = await import('/lib/backpropagation/v1.0.0/backpropagation.mjs');

let layers = [];

layers.push(ba.affineL2({
    inputs:20,
    outputs:5,
    lambda:0.1
}));

Dropout


Dropout is provided by the affineDropout function 


let ba = await import('/lib/backpropagation/v1.0.0/backpropagation.mjs');

let layers = [];

layers.push(ba.affineDropout({
    inputs:20,
    outputs:5,
    probability:0.05
}));