Namespace SiaNet.Regularizers

Classes

BaseRegularizer

Regularizers allow to apply penalties on layer parameters or layer activity during optimization. These penalties are incorporated in the loss function that the network optimizes.

The penalties are applied on a per-layer basis.The exact API will depend on the layer, but the layers Dense, Conv1D, Conv2D and Conv3D have a unified API.

L1

L1 regularization technique also called Lasso Regression

Lasso Regression (Least Absolute Shrinkage and Selection Operator) adds “absolute value of magnitude” of coefficient as penalty term to the loss function.

L1L2

Combined regularizer for Lasso and Ridge regression technique

L2

L2 regularization technique also called Ridge Regression.

Ridge regression adds “squared magnitude” of coefficient as penalty term to the loss function. Here the highlighted part represents L2 regularization element.