Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

TT: Fachverband Tiefe Temperaturen

TT 33: Correlated Magnetism – Spin Liquids

TT 33.6: Vortrag

Mittwoch, 19. März 2025, 16:15–16:30, H33

Variational Monte Carlo Simulations of Two-dimensional Quantum Spin Liquids — •Florian Michael and Benedikt Fauseweh — TU Dortmund University, Dortmund, Germany

In this project we use state-of-the-art variational algorithms to train neural quantum states for the quantum spin liquid phase of the J1-J2 Heisenberg model on a square lattice. Specifically, this approach makes use of a hybrid architecture of a restricted Boltzmann machine and pair-product states, capturing both global and local correlations efficiently. To further increase the precision of the wave function representation as well as mitigate finite-size effects, we apply quantum number projections and impose twisted boundary conditions.

The project is implemented within the NetKet framework, leveraging the automatic differentiation and just-in-time compilation of JAX as well as GPU accelerated high-performance clusters. The goal is to further advance the application of neural quantum states in quantum many-body physics and gain new insights on properties of quantum spin liquids that are currently difficult to simulate due to their long-range entanglement and absence of magnetic order.

Keywords: variational Monte Carlo; quantum spin liquids; neural quantum states; NetKet; Heisenberg model