Reliably characterized pulses are the starting point of any application of ultrafast techniques. Unfortunately, experimental constraints do not always allow for optimizing the characterization conditions. This dictates the need for refined analysis methods. Here we show that neural networks can provide a viable characterization when applied to data from interferometry for direct electric-field reconstruction (SPIDER). We have adopted a cascade of convolutional networks, addressing the multiparameter structure of the interferogram with a reasonable computing power. In particular, the necessity of precalibration is reduced, thus pointing toward the introduction of neural networks in more generic arrangements.
Gianani, I., Walmsley, I.A., Barbieri, M. (2024). SPIDERweb: a neural network approach to spectral phase interferometry. OPTICS LETTERS, 49(19), 5415-5418 [10.1364/OL.534767].
SPIDERweb: a neural network approach to spectral phase interferometry
Gianani I.
;Barbieri M.
2024-01-01
Abstract
Reliably characterized pulses are the starting point of any application of ultrafast techniques. Unfortunately, experimental constraints do not always allow for optimizing the characterization conditions. This dictates the need for refined analysis methods. Here we show that neural networks can provide a viable characterization when applied to data from interferometry for direct electric-field reconstruction (SPIDER). We have adopted a cascade of convolutional networks, addressing the multiparameter structure of the interferogram with a reasonable computing power. In particular, the necessity of precalibration is reduced, thus pointing toward the introduction of neural networks in more generic arrangements.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.