| The Advanced Photon Source (APS) at Argonne organizes an ongoing seminar series, the APS Scientific Computation Seminar Series, focused on various aspects of computation for synchrotron science; talks are related to data analysis, reconstruction, theory, simulation, optimization, machine learning, automation, and visualization. The seminar provides an opportunity to present and learn about state-of-the-art software and the application of computational, AI/ML and mathematical techniques to synchrotron science. |
| Title: | AI-Driven Approaches for X-ray Diffraction: Rapid Plastic Deformation Detection and Scalable ML Infrastructure | | Presenter(s): | Weijian Zheng, Postdoctoral Appointee, Data Science & Learning Division | | Date: | March 17, 2025 | | Time: | 1:00 PM (Central Daylight Savings Time) | | Location: | Join ZoomGov Meeting
https://argonne.zoomgov.com/j/1601444470?pwd=N1phbHZVdCtmcVR5cGh0c1Zhc0…
Meeting ID: 160 144 4470
Passcode: 937918
One tap mobile
+16692545252,,1601444470# US (San Jose) +16468287666,,1601444470# US (New York) Dial by your location
+1 669 254 5252 US (San Jose)
+1 646 828 7666 US (New York)
+1 646 964 1167 US (US Spanish Line) +1 669 216 1590 US (San Jose)
+1 415 449 4000 US (US Spanish Line) +1 551 285 1373 US
Meeting ID: 160 144 4470
Find your local number: https://argonne.zoomgov.com/u/af2crdvQy
| | Abstract: | This presentation will highlight two research efforts—fast anomaly detection for X-ray diffraction and M2ML’s scalable machine learning infrastructure. Both contribute to advancing AI-driven automation and computational efficiency in large-scale scientific research, supporting next-generation facilities like APS-U. Advancements in high-energy X-ray diffraction and machine learning are transforming materials research by enabling faster detection of plastic deformation events in metallic polycrystals. Tracking these deformation events is crucial for understanding material behavior under stress and improving the design of structural materials. Our automated anomaly detection method achieves up to 50× speedup, efficiently identifying these events. We also present a preliminary study on pinpointing individual grains responsible for plastic deformation, leveraging MIDAS simulations for further investigation. At the same time, the growing complexity of machine learning models and expanding data volumes demand scalable solutions. We introduce M2ML, a high-performance framework that streamlines data transfer, model training, fine-tuning, and inferencing across experimental facilities, HPC centers, and edge devices. By integrating Globus services, M2ML automates complex workflows and optimizes distributed computing resources. |
|
