Automatic hyperparameter exploration

Hyperparameter selection is an important aspect of successfully training a model. However, manually experimenting with different combinations of values for the hyperparameters car be a tedious task. What techniques, algorithms or heuristics exist to automatically explore the hyperparameter space to find promising values?


1 Response to “Automatic hyperparameter exploration”

  1. 1 Gabriel Bernier-Colborne February 1, 2013 at 10:35

    One way of selecting hyper-parameters is to use a hyper-learner, which is a pure learning algorithm (an algorithm without hyper-parameters) which passes different combinations of hyper-parameters to an impure learning algorithm in order to find the optimal values. Other techniques can be used, which will be discussed later in this course.

Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: