Computing, Machine Learning, and Data-Intensive Science at Synchrotron Facilities

Computing, Machine Learning, and Data-Intensive Science at Synchrotron Facilities

Research at synchrotron light sources hold transformational promise to address some of the most pressing scientific questions in society. They provide the unique ability to probe matter from atomic to mesoscopic length scales; enable experiments at physically-relevant conditions with respect to parameters such as temperature, chemical environment, mechanical loading, and so on; and can probe phenomena over time-scales spanning many orders of magnitude. While these capabilities have enabled scientific breakthroughs across multiple fields, recent dramatic improvements in x-ray source brightness and advances in detector technology have led to an explosion in sophisticated measurement techniques and raw data volumes, presenting technological and computational challenges as well as scientific opportunities.

To fully realize the promise of these facilities, we need new strategies for managing, interacting with, and extracting relevant physical information from increasingly large and complex datasets. This workshop will bring together experts in x-ray and domain science, machine learning, scientific computing, cyberinfrastructure, and data management to share knowledge and brainstorm ways to advance computational tools and infrastructure for data-intensive x-ray science.

Workshop goals:

• Inform future facility directions and initiatives
• Identify opportunities for collaboration within and across disciplines
• Identify community training and support needs

Topical sessions:

• On-the-fly data analysis
• Machine learning
• Computational- and data-intensive applications
• Data curation
• Workflows, visualization and systems for big data

Ideal Slam talks: