Hannover 2020 – wissenschaftliches Programm

Die DPG-Frühjahrstagung in Hannover musste abgesagt werden! Lesen Sie mehr ...

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

Q: Fachverband Quantenoptik und Photonik

Q 6: Quantum Effects (Disorder and Entanglement)

Q 6.7: Vortrag

Montag, 9. März 2020, 12:45–13:00, f342

Quantum manybody systems with neural networks — •Felix Behrens, Stefanie Czischek, Martin Gärttner, and Thomas Gasenzer — KIP, Heidelberg

The idea of connecting artificial neural networks and quantum mechanics gained a lot of interest over the last years. A representation of arbitrary quantum many-body states using a specific kind of artificial neural network, the restricted Boltzmann machine, has been introduced in [1]. With a generative model approach, any state can be represented. In the framework of Positive Operator Valued Measures (POVM), time evolution eg. following Heisenberg equation, can be represented for a probability distribution. We implement this ansatz with standard machine learning techinques for Restricted Boltzmann Machine (RBM). Given some, hypothetically measured, data, the RBM facilitates fast sampling from the underlying probability. Those samples can in principle be used for a Monte-Carlo like integration of the time evolution and for measuring any operator. The next qurstion to ask is, how statistical noise in the RBM implementation violates physical constraints on the state and its expectation values and how to restrict the RBM representation to its physical subspace. [1] Juan Carrasquilla, arXiv:1810.10584, Oct 2018.