260072 VU Data Science for Physicists (2026S)

Data science is fundamental for machine-learning applications and artificial intelligence. Considering the massive scientific, societal, and economic impact of this field, data science is one of the most critical skills needed to succeed in future science and industry. As a physicist, it is imperative to gain an understanding of data science that goes beyond buzzwords.

The staggering growth of computational power and the associated creation of large databases make data one of the most valuable resources of our century. Information stored in such large databases can be extracted by data science methods—that is, by combining statistics with advanced computational methods, including machine learning.

The course focuses on state-of-the-art data science methods in physics. We will discuss a combination of interdisciplinary activities, such as scientific, statistical, and computational tools, required in science as well as industry. This course aims to provide students with an introduction to basic theoretical concepts and practical skills oriented toward data science in physics. Specifically, the lectures cover the following topics:

Syllabus:
0 Introduction -- General information; What is data science in physics?
1 Linear Algebra -- Matrix decompositions, eigenvalue problems, and Principal Component Analysis.
2 Differential Equations -- Finite difference, integrating first-order and partial differential equations.
3 Fitting and Testing -- Regression methods and model testing.
4 Optimization -- Local and global optimization methods.
5 Statistics -- Statistical/Bayesian hypothesis testing and sampling methods.
6 Machine Learning -- Introduction to machine-learning and accuracy of machine-learning models.
7 Machine Learning -- Classification, the perception model, Bayesian classifiers, and trees.
8 Machine Learning -- Neural networks.
9 Machine Learning -- Kernel methods.

The course is structured in theoretical lectures (on Thursdays), followed by practical exercise classes (on Tuesdays).