See the full schedule here.
Title: Automating and Advancing X-ray Diffraction Data Reduction with MIDAS at the Advanced Photon Source
Keywords: Fast data analysis
Abstract: This presentation provides a comprehensive overview of the cutting-edge tools and methodologies being developed within the Multipurpose Instrument Data Acquisition System (MIDAS) to streamline and enhance X-ray diffraction data reduction at the Advanced Photon Source (APS). Addressing the challenges posed by increasing data volumes and experimental complexity, particularly with the upcoming APS upgrade (APS-U), we introduce a suite of automated solutions designed to improve efficiency, accuracy, and adherence to FAIR (Findable, Accessible, Interoperable, and Reusable) data principles.
The talk will detail several key advancements, including automated calibration and integration workflows that significantly reduce manual intervention and improve consistency. We will introduce innovative approaches to data processing, such as the use of Noise2Void for superior image denoising compared to traditional methods. Furthermore, the presentation will cover the development of a "digital twin" to assist users in experimental design, as well as the implementation of a synchrotron ontology and knowledge graphs to create more traceable and reproducible scientific workflows. These tools are pivotal for managing the vast and complex metadata generated during experiments. We will also showcase the application of machine learning, with BraggNN, for ultra-fast and accurate diffraction spot identification, and a novel Rare Event Indicator for capturing critical information during in-situ experiments. These developments collectively empower researchers to move from raw data to insightful results more effectively, paving the way for new scientific discoveries.
