About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Artificial Intelligence Applications in Integrated Computational Materials Engineering (AI-ICME)
|
| Presentation Title |
An Extensible, Data-Driven Platform for Automated Atomic Structure Analysis |
| Author(s) |
Ryan Sills, Yating Fang, Qian Qian Zhao, Pallavi Biswas, Ahmed Aziz Ezzat, Alexander Stukowski, Daniel Utt, Andre Merzky, Shantenu Jha |
| On-Site Speaker (Planned) |
Ryan Sills |
| Abstract Scope |
A major challenge when analyzing atomistic simulations is identifying what atomic structures are present. Existing methods for structure analysis, such as common neighbor analysis and the dislocation extraction algorithm, are narrow in scope (e.g., can only identify certain crystal structures) and cannot easily be generalized to arbitrary structures, such as point defects and grain boundaries. In this talk, we present a data-driven platform called ATOMIC (Atomistic Tool for Open, Machine-learned Identification and Characterization) that aims to serve as an infinitely extensible atomic structure analysis tool for the ICME and atomistic modeling communities. The key innovations underlying ATOMIC are (1) a data-driven classification scheme where new classes (structures) can be added without retraining for all prior classes, (2) an automatic data labeling scheme that obviates the need for human input during data generation via molecular dynamics simulations, and (3) a computational workflow manager for distributing computing tasks across diverse resource pools. |
| Proceedings Inclusion? |
Planned: |