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

Q: Fachverband Quantenoptik und Photonik

Q 25: Poster – Cold Atoms and Molecules, Matter Waves (joint session Q/A/MO)

Q 25.7: Poster

Dienstag, 11. März 2025, 14:00–16:00, Tent

Evalutation of machine learning algorithms for applications in quantum gas experiments — •Oliver Anton¹, Elisa Da Ros¹, Philipp-Immanuel Schneider^3,4, Ivan Sekulic^3,4, Sven Burger^3,4, and Markus Krutzik^1,2 — ¹Institut für Physik and IRIS, Humboldt-Universität zu Berlin — ²Ferdinand-Braun-Institut, Berlin — ³JCMwave GmbH, Berlin — ⁴Zuse Institute Berlin (ZIB), Berlin

The generation of clouds containing cold and ultra-cold atoms is a complex process that requires the optimization of noisy data in multi dimensional parameter spaces. Optimization of such problems can present challenges both in and outside of the lab due to constrains in time, expertise, or access for lengthy manual optimization.
Machine learning offers a solution thanks to its ability to efficiently optimize high dimensional problems without the need for knowledge of the experiment itself. In this poster, we present the results of benchmarking various optimization algorithms and implementations. Their performance is tested in a cold atom experiment, subjected to inherent noise [1]. Current research aims towards the preparation of the cloud for quantum memory applications [2], by engineering the optical density using the tested algorithms.
[1] O. Anton et al., Machine Learning: Science and Technology 5 025022, 2024
[2] E. Da Ros et al., Physical Review Research 5 033003, 2023

Keywords: Machine learning; BEC; Mobile experiment; Automatisation