Over the past couple of years, ptychography has become increasingly computationally intensive with larger detectors running at higher speeds. Today, it is routine to run Ptychography experiments which produce > Gb/s data. This trend will continue, and traditional iterative reconstruction algorithms will struggle to keep pace with the speed of data generation even when GPU-accelerated and running on HPC resources. Recently, ML methods have proven to be successful - speeding up image reconstruction by orders of magnitude and allowing for advanced algorithms to improve image quality. Nevertheless, challenges and caveats exist with respect to the reliability of results predicted by ML methods.
In this workshop, we would like to present recent developments as well as current challenges and pain points of machine learning and advanced optimization applied to ptychography. We will also hear from a couple of speakers applying ML to other x-ray techniques and their potential application to ptychography workflows. We would like to create a space for open discussion to lead the way for higher applicability and acceptance of ML-supported data analysis in this field with direct integration of workflows at the beamlines.
Organization Committee:
- Juliane Reinhardt (LBNL)
- Ester Tsai (BNL)
- Mathew Cherukara (ANL)
Preliminary Agenda:
| (Central Time) | Title | Speaker |
|---|---|---|
| 11:00-11:15 | Welcome + Setting the stage | |
| 11:15-11:35 | Real-time & Large FOV Ptychography through AI@Edge | Tao Zhou (ANL) |
| 11:35-11:55 | Super-resolution, denoising, AI-guided phasing parameter selection | Yi Jiang (ANL) |
| 11:55-12:15 | Unsupervised ML techniques for fast feature tracking in ptychography | Dergan Lin (ANL) |
| Short Break | ||
| 12:25-12:45 | TBA | Esther Tsai (BNL) |
| 12:45-13:05 | ML for XPCS | Andi Barbour/Tatiana Konstantinova (BNL) |
| 13:05 till end | Discussion/Demos |
Meeting Link:
https://lbnl.zoom.us/j/99748049371?pwd=QzltamRYYVdFQXo1UndBcG1rOVl2dz09