Artificial neural networks have been shown to be powerful approximations for quantum many-body problems. In this project, the aim is to study and develop various deep learning methods based on restricted Boltzmann machines and neural networks to study quantum mechanical many-particle systems.
Deep learning approaches have proven to be a flexible tool to approximate quantum many-body states in condensed matter and chemistry problems, even when non-perturbative interactions are prominent. The project has its main focus on many-particle systems in condensed matter physics and nuclear physics, using suitable approaches for modelling the ground-state wave function of quantum mechanical many-body systems. The project relies on efficient stochastic sampling and optimization schemes, including various Monte Carlo sampling methods.
The project can also be extended to include quantum machine learning approaches based on quantum Boltzmann machines and other deep learning methods.
Requirements
- The applicant is expected to hold a Master’s degree or equivalent in Computational or Theoretical Physics with specialization in quantum mechanical many-body problems with a strong background in Computational Physics/Science.
- Candidates with a background in Machine Learning will be preferred.
Supervisors
Professor Morten Hjorth-Jensen
Call 2: Project start autumn 2022
This project is in call 2, starting autumn 2022.