Use Case 3 – Machine Learning Based Spectra Classification
At the European XFEL, X-ray pulses can be generated with only 220ns separation in time and a maximum of 27000 pulses per second. In experiments at European XFEL, spectral changes can indicate the change of the system under investigation and so the progress of the experiment. An immediate feedback on the actual status (e.g., time resolved status of the sample) would be essential to quickly judge how to proceed with the experiment. The two major spectral changes that we aim to capture are either the change of intensity distribution (e.g., drop or appearance) of peaks at certain locations, or the shift of those on the spectrum.
Machine Learning (ML) opens up new avenues for data-driven analysis in spectroscopy by offering the possibility to quickly recognize such specific changes on-the-fly during data collection. ML requires lots of data which are clearly annotated. Hence, it is important that research outputs align with the FAIR principles. In case of XFEL experiments, it is suggested to introduce NeXus data format standards in future experiments.
In this work an example is presented, of a possible use of Neural Network based ML for accurately classifying the system state if data is properly provided. A solution is shown, to automatically find the regions (or bins) with high separability where the spectra classes differ significantly. By teaching individual neural networks for each bin, and by combining them with a weighting technique, a robust classification of any new spectral curve can be quickly obtained.
Christian Plueckthun; Zuzana Konopkova; Sandor Brockhauser; Yue Sun
. Pennicard, D., Smoljanin, S., Pithan, F., Sarajlic, M., Rothkirch, A., Yu, Y., Liermann, H.P., Morgenroth, W., Winkler, B., Jenei, Z. and Stawitz, H., 2018. LAMBDA 2M GaAs—A multi-megapixel hard X-ray detector for synchrotrons. Journal of Instrumentation, 13(01), p.C01026, DOI: 10.1088/1748-0221/13/01/C01026
PaNOSC related work packages:
More on PaNOSC github:
Should you have questions, comments or suggestions related to this use case, please write an email to firstname.lastname@example.org