Dresden 2026 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
DY: Fachverband Dynamik und Statistische Physik
DY 14: Machine Learning in Dynamics and Statistical Physics II
DY 14.13: Vortrag
Montag, 9. März 2026, 18:15–18:30, HÜL/S186
Optimization and Representability of time-dependent Neural Quantum States: a study of the 1D critical quantum Ising model — •Wladislaw Krinitsin1,2, Mohammad Abedi1,2, Jonas Rigo2, and Markus Schmitt1,2 — 1PGI-8, Forschungszentrum Jülich, Jülich, Germany — 2Faculty for Informatics and Data Science, Regensburg University, Germany
In recent years, neural quantum states have emerged as a competitive and powerful numerical approach for many body systems. While they provide a flexible and scalable ansatz, able to represent any state as suggested by the function-approximation theorem, their practical limitations are still opaque, in particular regarding representability and optimization. In this work we investigate these questions within the framework of variational Monte Carlo on the example of the time evolution of the critical transverse-field Ising model in one dimension. Even for moderate system sizes, the departure from the exact solution occurs very early in the dynamics, allowing us to systematically analyze the representability of the state at each time step as well as the impact of different sampling strategies.
Keywords: Neural Quantum States; Machine Learning; Transverse Field Ising Model; Variational Monte Carlo; Importance Sampling