Visiting Student - XSD - Koda, Ryota - 10.21.26

Argonne National Laboratory • Lemont, IL

3d

About The Position

This visiting student project will develop a real-time workflow for 3D X-ray ptychography using a sample-specific surrogate model to reduce reconstruction time. In 3D ptychography, a sample is illuminated from multiple angles. A 2D image is recovered at each angle using iterative phase retrieval, such as ePIE, and the images are combined into a 3D volume. Because phase retrieval is repeated for every direction, it dominates the total runtime.

The proposed workflow builds a surrogate model on the fly rather than relying on a training dataset prepared in advance. Directional subsampling will select a subset of projection angles. ePIE will run on those angles, and the resulting diffraction-pattern and object-image pairs will form a small training set. A surrogate model trained on this set will predict initial object estimates for the remaining angles. After preprocessing, these predictions will initialize a second ePIE pass. The workflow will run on GPUs.

Additional work will include developing a foundation surrogate model for diverse samples and testing other surrogate architectures. The entire appointment must be conducted within the United States.

Requirements

• Must be 18 years or older when the appointment begins.
• Must conduct the entire appointment within the United States.
• Must be currently enrolled in undergraduate or graduate studies at an accredited institution, have graduated from an accredited institution within the past 3 months, or be actively enrolled in a graduate program at an accredited institution.

Responsibilities

• Develop a real-time 3D X-ray ptychography workflow using a sample-specific surrogate model.
• Use directional subsampling to select projection angles and run ePIE on the selected angles.
• Build a training set from diffraction-pattern and object-image pairs and train an on-the-fly surrogate model.
• Use surrogate predictions as initial object estimates for the remaining angles in a second ePIE pass after preprocessing.
• Run the workflow on GPUs.
• Develop a foundation surrogate model for diverse samples.
• Test other surrogate architectures.

What This Job Offers

• Job Type: Full-time
• Career Level: Entry Level
• Education Level: No Education Listed