Research / Physics-Informed AI for Scientific Instrumentation / Predicting instrument response
Physics-Informed Machine Learning for Accelerated Discovery and Dynamics Analysis in Ultrafast X-Ray Diffraction
We use simulated atomic motion to predict how X-ray scattering changes over time, then compare those predictions with experimental measurements.

Project overview
Ultrafast X-ray photon correlation spectroscopy (XPCS) tracks how materials change over time, from atomic to microscopic scales. We train models on simulated atomic motion to predict the temporal correlation functions measured in XPCS experiments. That gives us a way to test simulations against measurements of supercooled liquids and materials undergoing the glass transition.