Regularization for Deep Learning: A Taxonomy

TL;DR

This paper offers a comprehensive taxonomy of deep learning regularization methods, categorizing them into data, architecture, error, regularization terms, and optimization, to unify understanding and guide future research.

Contribution

It introduces a systematic taxonomy that unifies diverse regularization techniques in deep learning, highlighting their relationships and providing practical guidance.

Findings

Reveals fundamental links between different regularization methods

Provides a structured categorization to facilitate understanding and development

Offers practical recommendations for users and developers

Abstract

Regularization is one of the crucial ingredients of deep learning, yet the term regularization has various definitions, and regularization methods are often studied separately from each other. In our work we present a systematic, unifying taxonomy to categorize existing methods. We distinguish methods that affect data, network architectures, error terms, regularization terms, and optimization procedures. We do not provide all details about the listed methods; instead, we present an overview of how the methods can be sorted into meaningful categories and sub-categories. This helps revealing links and fundamental similarities between them. Finally, we include practical recommendations both for users and for developers of new regularization methods.