Hyperparameter Optimization in Machine Learning - GPU computing & Machine learning regular meeting at DESY -

by Kai Krajsek (Jülich Supercomputing Center)

Monday 28 Oct 2019, 16:00 → 17:00 Europe/Berlin

Flash Seminar Room (Bldg. 28c)

Despite the successes of recent years, the application of machine learning to a particular task can become a laborious process of selecting appropriate methods and configuring them. In this talk, methods are discussed that help to optimize such hyperparameters. After an introduction to the basics of hyperparameter optimization, established methods such as grid search and Bayesian optimization will be presented. The last part of the talk deals with the relationship of hyperparameter optimization to areas such as meta-learning and AutoML.

Slides krajsek_desy_2019.pdf