PeakNet: Bragg peak finding in X-ray crystallography experiments with U-Net

Serial crystallography at X-ray free electron laser (XFEL) sources has experienced tremendous progress in achieving high data rate in recent times. While this development offers potential to enable novel scientific investigations, such as imaging molecular events at logarithmic timescales, it also poses challenges in regards to real-time data analysis, which involves some degree of data reduction to only save those features or images pertaining to the science on disks. If data reduction is not effective, it could directly result in a substantial increase in facility budgetary requirements, or even hinder the utilization of ultra-high repetition imaging techniques making data analysis unwieldy. Furthermore, an additional challenge involves providing real-time feedback to users derived from real-time data analysis. In the context of serial crystallography, the initial and critical step in real-time data analysis is finding X-ray Bragg peaks from diffraction images. To tackle this challenge, we present PeakNet, a Bragg peak finder that utilizes neural networks and runs about four times faster than Psocake peak finder, while delivering significantly better indexing rates and comparable number of indexed events. We formulated the task of peak finding into a semantic segmentation problem, which is implemented as a classical U-Net architecture. A key advantage of PeakNet is its ability to scale linearly with respect to data volume, making it well-suited for real-time serial crystallography data analysis at high data rates.

SpeckleNN: A unified embedding for real-time speckle pattern classification in X-ray single-particle imaging with limited labeled examples

With X-ray free-electron lasers (XFELs), it is possible to determine the three-dimensional structure of noncrystalline nanoscale particles using X-ray single-particle imaging (SPI) techniques at room temperature. Classifying SPI scattering patterns, or "speckles", to extract single hits that are needed for real-time vetoing and three-dimensional reconstruction poses a challenge for high data rate facilities like European XFEL and LCLS-II-HE. Here, we introduce SpeckleNN, a unified embedding model for real-time speckle pattern classification with limited labeled examples that can scale linearly with dataset size. Trained with twin neural networks, SpeckleNN maps speckle patterns to a unified embedding vector space, where similarity is measured by Euclidean distance. We highlight its few-shot classification capability on new never-seen samples and its robust performance despite only tens of labels per classification category even in the presence of substantial missing detector areas. Without the need for excessive manual labeling or even a full detector image, our classification method offers a great solution for real-time high-throughput SPI experiments.

fairDMS: Rapid Model Training by Data and Model Reuse

Extracting actionable information from data sources such as the Linac Coherent Light Source (LCLS-II) and Advanced Photon Source Upgrade (APS-U) is becoming more challenging due to the fast-growing data generation rate. The rapid analysis possible with ML methods can enable fast feedback loops that can be used to adjust experimental setups in real-time, for example when errors occur or interesting events are detected. However, to avoid degradation in ML performance over time due to changes in an instrument or sample, we need a way to update ML models rapidly while an experiment is running. We present here a data service and model service to accelerate deep neural network training with a focus on ML-based scientific applications. Our proposed data service achieves 100x speedup in terms of data labeling compare to the current state-of-the-art. Further, our model service achieves up to 200x improvement in training speed. Overall, fairDMS achieves up to 92x speedup in terms of end-to-end model updating time.

Bridge Data Center AI Systems with Edge Computing for Actionable Information Retrieval

Extremely high data rates at modern synchrotron and X-ray free-electron lasers (XFELs) light source beamlines motivate the use of machine learning methods for data reduction, feature detection, and other purposes. Regardless of the application, the basic concept is the same: data collected in early stages of an experiment, data from past similar experiments, and/or data simulated for the upcoming experiment are used to train machine learning models that, in effect, learn specific characteristics of those data; these models are then used to process subsequent data more efficiently than would general-purpose models that lack knowledge of the specific dataset or data class. Thus, a key challenge is to be able to train models with sufficient rapidity that they can be deployed and used within useful timescales. We describe here how specialized data center AI systems can be used for this purpose.