About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Advances in Ceramic Materials and Processing
|
| Presentation Title |
From Data to Discovery: Active Learning Unlocks Complex Ceramic Design Spaces |
| Author(s) |
Corey Oses |
| On-Site Speaker (Planned) |
Corey Oses |
| Abstract Scope |
Active learning is reshaping the design and processing of complex ceramics, especially for high-entropy and compositionally complex systems. We present a workflow that strategically selects representative structures for simulation, enabling accurate property prediction and phase stability assessment while significantly reducing data requirements. By integrating targeted sampling with improved computational accuracy and advanced structural descriptors, our method uncovers how composition and electronic structure influence ceramic stability and performance. Applied to ceramics for energy and environmental applications, this approach efficiently maps stable compositional spaces and identifies promising synthesis targets, achieving strong agreement with experimental phase diagrams. This work demonstrates how selective data-driven modeling and advanced simulation accelerate the discovery and optimization of next-generation ceramics, driving innovation in material development and processing. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Modeling and Simulation, Energy Conversion and Storage, Machine Learning |