ML in Fundamental Physics – Overview

News:

• If you have registered before 8am April 20th, you should have received information on how to access the moodle page and the link to the lectures.
• Registration via LSF is necessary. You will then receive further information on how lectures are taking place. Lectures will take place via Zoom.
• We will also use Moodle to coordinate exercises and additional material.
• Please check back at a later stage for more information on course details.

Examination:

There shall be a written exam.

Outline of course:

• Basics of Machine Learning
• Optimisers
• Regression
• Logistic/Multi-class classification
• A survey of classifiers
• Neural Networks
• Unsupervised learning
• Variational Methods
• Generative Adversarial Networks
• Normalising Flows
• Reinforcement Learning
• Applications in Physics

Literature:

• Mehta, Bukov, Wang, Day, Richardson, Fisher, Schwab: A high-bias, low-variance introduction to Machine Learning for physicists (1803.08823)
• MacKay: Information Theory, Inference, and Learning Algorithms (CUP, free online version)
• Goodfellow, Bengio, Courville: Deep Learning (MIT Press, deeplearningbook.org)
• Carleo, Cirac, Cranmer, Daudet, Schuld, Tishby, Vogt-Maranto, Zdeborova: Machine Learning and the Physical Sciences (1903.10563)
• Ruehle: Data science applications to string theory (Physics Reports)

Prerequisites:

• No knowledge of python is required, but it is helpful if you have heard about a for-loop, if statements at some point in your life.
• Willingness to get your hands on coding.