SKM 2021 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
DY: Fachverband Dynamik und Statistische Physik
DY 4: Poster Session II: Nonlinear Dynamics, Simulations and Machine Learning
DY 4.9: Poster
Dienstag, 28. September 2021, 17:30–19:30, P
Machine learning generators of open system Lindblad dynamics — •Francesco Carnazza1, Dominik Zietlow2, Federico Carollo1, Sabine Andergassen1, Georg Martius2, and Igor Lesonovsky1,3 — 1Institut für Theoretische Physik and Center for Quantum Science, Universität Tübingen, Auf der Morgenstelle 14, 72076 Tübingen, Germany — 2Max Planck Institute for Intelligent Systems, Max-Planck-Ring 4, 72076 Tübingen, Germany — 3School of Physics and Astronomy, University of Nottingham, Nottingham, NG7 2RD, UK
In recent years artificial neural network methods have established themselves as a versatile tool to encode the state of both closed and open quantum systems. We are interested in the question whether they can learn the generator of an effective open quantum dynamics which governs a small system of interest that is embedded within a larger one (Paolo Mazza et al. 2020 https://doi.org/10.1103/PhysRevResearch.3.023084). The model we consider is a spin chain where the system of interest is formed by two spins which are coupled to a "bath" consisting of the rest of the chain. The whole chain is evolved according to an transverse field Ising Hamiltonian. From the reduced density matrix, obtained by tracing out the bath degrees of freedom, the two-body correlations are determined, which are subsequently used to train the network. A simple architecture is adopted in order to have the possibility to "look inside" the network and to see whether the learned dynamics is indeed governed by a time-local Lindblad generator.