Optics and Sensing Laboratory Article 2024

Reinforcement learning in cold atom experiments

Thumb ticker sm 20240912 volchkov valentin 1 small
Optics and Sensing Laboratory
Optics & Sensing Laboratory
Mpiis teaser

Cold atom traps are at the heart of many quantum applications in science and technology. The preparation and control of atomic clouds involves complex optimization processes, that could be supported and accelerated by machine learning. In this work, we introduce reinforcement learning to cold atom experiments and demonstrate a flexible and adaptive approach to control a magneto-optical trap. Instead of following a set of predetermined rules to accomplish a specific task, the objectives are defined by a reward function. This approach not only optimizes the cooling of atoms just as an experi- mentalist would do, but also enables new operational modes such as the preparation of pre-defined numbers of atoms in a cloud. The machine control is trained to be robust against external perturbations and able to react to situations not seen during the training. Finally, we show that the time con- suming training can be performed in-silico using a generic simulation and demonstrate successful transfer to the real world experiment.

Author(s): Malte Reinschmidt and József Fortágh and Andreas Günther and Valentin Volchkov
Journal: nature communications
Volume: 15:8532
Year: 2024
Month: January
Bibtex Type: Article (article)
DOI: 10.1038/s41467-024-52775-8
URL: https://doi.org/10.1038/s41467-024-52775-8
Electronic Archiving: grant_archive

BibTex

@article{RLMOT,
  title = {Reinforcement learning in cold atom experiments},
  journal = {nature communications},
  abstract = {Cold atom traps are at the heart of many quantum applications in science and
  technology. The preparation and control of atomic clouds involves complex
  optimization processes, that could be supported and accelerated by machine
  learning. In this work, we introduce reinforcement learning to cold atom
  experiments and demonstrate a flexible and adaptive approach to control a
  magneto-optical trap. Instead of following a set of predetermined rules to
  accomplish a specific task, the objectives are defined by a reward function.
  This approach not only optimizes the cooling of atoms just as an experi-
  mentalist would do, but also enables new operational modes such as the
  preparation of pre-defined numbers of atoms in a cloud. The machine control
  is trained to be robust against external perturbations and able to react to
  situations not seen during the training. Finally, we show that the time con-
  suming training can be performed in-silico using a generic simulation and
  demonstrate successful transfer to the real world experiment.},
  volume = {15:8532},
  month = jan,
  year = {2024},
  slug = {rlmot},
  author = {Reinschmidt, Malte and Fortágh, József and G{\"u}nther, Andreas and Volchkov, Valentin},
  url = {https://doi.org/10.1038/s41467-024-52775-8},
  month_numeric = {1}
}