A deep neural network approach to the identification and quantification of powder X-ray diffraction patterns was applied and proved successful for the quantitative description of complex mineralogical assemblages consisting of up to four minerals with different structures, including different space groups, for which data augmentation is not straightforward.

Machine-learning approaches can greatly facilitate single-particle-imaging experiments at X-ray free-electron-laser facilities by providing real-time images from the coherent X-ray diffraction data stream, using methods presented in this article.

Machine learning methods have been applied to solve simple Patterson maps. The results demonstrate the potential to use neural networks for solving the phase problem in crystallography.

A deep-machine-learning technique based on a convolutional neural network (CNN) is introduced. It has been employed for the classification of crystal system, extinction group and space group for given powder X-ray diffraction patterns of inorganic materials.

A 3D U-net deep convolutional neural network has been developed and tested to segment precipitates in synchrotron-based X-ray tomography experiments. Comparison of predicted segmentation showed a good agreement with manual segmentation.

Open-source deep-learning algorithms (Mask R-CNN) are used to automatically identify intracellular crystals in living insect cell cultures.

Convolutional neural networks are useful for classifying grazing-incidence small-angle X-ray scattering patterns. They are also useful for classifying real experimental data.

Deep learning provides one possible avenue to reduce the data stream generated by serial macromolecular X-ray crystallography. Convolutional neural networks can be trained to recognize the presence or absence of Bragg spots, forming a criterion to veto events prior to downstream data processing.

A neural network for coherent X-ray diffractive imaging experiments is presented that can restore noisy and masked simulated diffraction intensities from biological macromolecules.

A neural network trained with a few thousand simulations using random errors is demonstrated to predict accurately the lens error profile that accounts for all aberrations of a compound refractive lens in a synchrotron beamline.