Aufsatz in einer Fachzeitschrift
Modelling of ultrafast coherent strong-field dynamics in potassium with neural networks
Details zur Publikation
Autor(inn)en: | Selle, R.; Brixner, T.; Bayer, T.; Wollenhaupt, M.; Baumert, T. |
Publikationsjahr: | 2008 |
Seitenbereich: | 074019-1 - 074019-7 |
Jahrgang/Band : | 41 |
Zusammenfassung, Abstract
We investigate the applicability of neural networks ( NNs) for the automated generation of effective computer models for coherent light - matter interactions. The simulation of Autler - Townes doublets from strong-field ionization of potassium atoms is chosen as a test system that exhibits distinct quantum-mechanical effects. Shaped femtosecond laser pulses are employed for studying the response of a quantum-mechanical system to a large variety of different electric fields, and the resulting data can be used for training a NN. We show that a NN is able to approximate the investigated process in parameter regions sampled by the training data and that it can be employed for the interpolation of control landscapes.
We investigate the applicability of neural networks ( NNs) for the automated generation of effective computer models for coherent light - matter interactions. The simulation of Autler - Townes doublets from strong-field ionization of potassium atoms is chosen as a test system that exhibits distinct quantum-mechanical effects. Shaped femtosecond laser pulses are employed for studying the response of a quantum-mechanical system to a large variety of different electric fields, and the resulting data can be used for training a NN. We show that a NN is able to approximate the investigated process in parameter regions sampled by the training data and that it can be employed for the interpolation of control landscapes.
