Data science applications to string theory

We first introduce various algorithms and techniques for machine learning and data science. While there is a strong focus on neural network applications in unsupervised, supervised and reinforcement learning, other machine learning techniques are discussed as well. These include various clustering and anomaly detection algorithms, support vector machines, and decision trees. In addition, we review data science techniques such as genetic algorithms and topological data analysis. This first part of the review makes some reference to concepts in physics, but the explanations and examples do not assume any knowledge of string theory and should therefore be accessible to a wide variety of readers with a physics background. After that, we illustrate applications to string theory. We give an overview of existing string theory data sets and describe how they can be studied using data science techniques. We also explain the computational complexity involved in the investigation of string vacua. Example codes that illustrate the techniques introduced in this review are available from Fabian Ruehle (0000). (C) 2020 The Author. Published by Elsevier B.V.