Machine learrning optimization of a magneto-optical trap (2020)
- Authors:
- USP affiliated author: SANTOS, LUCAS MARCELO DE SÁ MARQUES DOS - IFSC
- School: IFSC
- Subjects: FÍSICA ATÔMICA; APRENDIZADO COMPUTACIONAL; ÓPTICA
- Keywords: Atomic cooling; Magneto-optical trap; Machine learning
- Language: Inglês
- Abstract: As an essential component of modem production techniques of Bose-Einstein condensates, the magneto-optical trap has an important role in determining the size of the condensate that can be obtained. Extensive work has been done over the years with the goals both of simplifying the construction of the trap and of acquiring a deeper understanding of the mechanisms at play in its interior, which has also allowed, to an extent, the prior determination of adequate choices of experimental parameters that lead to the production of larger and colder samples. However, although a simple description is sufficient to provide a broad understanding of the most relevant parameters cg, the cooling beam detuning and the magnetic field gradient - still other phenomena arise in the trap which are not taken into account by this model, such as collective effects and the presence of an additional laser beam. This absence of complete theoretical modeling precludes need, a priori, optimization of the magneto-optical trap. As a result, optimization is typically started from a set of standard, generally adequate parameters, which are then fine-tuned manually, a process that is of limited precision and, most importantly, too time-consuming to be done often enough to keep the trap near optimal conditions at all times. Machine learning optimization methods may provide an alternative that is both need and time-efficient. While previous work has applied machine leaving to the optimization of a magneto-optical trap with a performance measure based on optical depth, we propose fining this performance measure from the trap loading curve. In order to properly substantiate our proposal, we go over much of the basic physics of the magneto-optical trap and its loading dynamics, establishing the connection between experimental parameters and our performance measure. We then introduced our custom implementation and the methods used for optimization, which resulted in a considerable increase in thefinal trap population for shorter times and starting from unoptimized parameters
- Imprenta:
- Place of publication: São Carlos
- Date published: 2020
-
ABNT
SANTOS, Lucas Marcelo Sá Marques dos. Machine learrning optimization of a magneto-optical trap. 2020. Trabalho de Conclusão de Curso (Graduação) – Instituto de Física de São Carlos, Universidade de São Paulo, São Carlos, 2020. Disponível em: https://bdta.abcd.usp.br/directbitstream/fdb09dde-137a-446a-9ef7-c33cfc2d2ab3/Lucas%20Marcelo%20de%20S%C3%A1%20Marques%20dos%20Santos.pdf. Acesso em: 24 abr. 2024. -
APA
Santos, L. M. S. M. dos. (2020). Machine learrning optimization of a magneto-optical trap (Trabalho de Conclusão de Curso (Graduação). Instituto de Física de São Carlos, Universidade de São Paulo, São Carlos. Recuperado de https://bdta.abcd.usp.br/directbitstream/fdb09dde-137a-446a-9ef7-c33cfc2d2ab3/Lucas%20Marcelo%20de%20S%C3%A1%20Marques%20dos%20Santos.pdf -
NLM
Santos LMSM dos. Machine learrning optimization of a magneto-optical trap [Internet]. 2020 ;[citado 2024 abr. 24 ] Available from: https://bdta.abcd.usp.br/directbitstream/fdb09dde-137a-446a-9ef7-c33cfc2d2ab3/Lucas%20Marcelo%20de%20S%C3%A1%20Marques%20dos%20Santos.pdf -
Vancouver
Santos LMSM dos. Machine learrning optimization of a magneto-optical trap [Internet]. 2020 ;[citado 2024 abr. 24 ] Available from: https://bdta.abcd.usp.br/directbitstream/fdb09dde-137a-446a-9ef7-c33cfc2d2ab3/Lucas%20Marcelo%20de%20S%C3%A1%20Marques%20dos%20Santos.pdf
Download do texto completo
| Tipo | Nome | Link | |
|---|---|---|---|
 |
Lucas Marcelo de Sá Marq... | Direct link |
How to cite
A citação é gerada automaticamente e pode não estar totalmente de acordo com as normas
