Hyperparameters And Validation Sets In Deep Learning

Hyperparameters And Validation Sets In Deep Learning

Introduction Most machine learning algorithms have several settings that we will use to regulate the behavior of the training algorithm. These settings are called hyperparameters. The values of hyperparameters aren’t adopted by the training algorithm itself (though we will design a nested learning procedure where one learning algorithm learns the simplest hyperparameters for an additional learning algorithm).  Description Within the polynomial regression example, there’s one hyperparameter: the degree of the polynomial, which acts as a … Read more