Short description: This course covers applications of machine learning methods to quantum (many-body) systems. The module starts with an introduction of the relevant physical systems, ranging from individual qubits to (quantum) many-body systems. The first chapter then focuses on machine learning methods for data analysis, ranging from principal component analysis through symbolic regression and support vector machines to advanced neural network architectures. Various applications of the learned methodology to different quantum systems, e.g. for phase classification and characterization are discussed in the context of research publications. The second chapter deals with the representation of quantum many-body states through neural networks, known as neural quantum states. Variational Monte Carlo is introduced to find ground states and techniques for simulating time evolution and finite temperature are discussed, and applications to different quantum many-body systems in the literature are presented. The third chapter tackles optimization problems, ranging from qubit control to error correction and parameter ramps in quantum simulation experiments, through reinforcement learning and Bayesian optimization techniques.
Requirements: BSc in physics or comparable.
Recommended Textbooks:Â
A. Dawid et al., Modern applications of machine learning in quantum sciences, arXiv:2204.04198 (to appear in Cambridge University Press)
H. Lange et al., From Architectures to Applications: A Review of Neural Quantum States, Quantum Science and Technology 9 (2024)
Sutton & Barto, Reinforcement Learning: an Introduction, Bradford Books