Unsupervised real-world knowledge extraction via disentangled variational autoencoders for photon diagnostics

We present real-world data processing on measured electron time-of-flight data via neural networks.Specifically, the use of disentangled variational autoencoders on data from a diagnostic instrumentfor online wavelength monitoring at the free electron laser FLASH in Hamburg. Without a-prioriknowledge the network is able to find representations of single-shot FEL spectra, which have a lowsignal-to-noise ratio. This reveals, in a directly human-interpretable way, crucial information aboutthe photon properties. The central photon energy and the intensity as well as very detector-specificfeatures are identified. The network is also capable of data cleaning, i.e. denoising, as well as theremoval of artefacts. In the reconstruction, this allows for identification of signatures with very lowintensity which are hardly recognisable in the raw data. In this particular case, the network enhancesthe quality of the diagnostic analysis at FLASH. However, this unsupervised method also has thepotential to improve the analysis of other similar types of spectroscopy data