| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
2025 Technical Division Student Poster Contest
|
| Presentation Title |
SPG-17: HEA Design Agent: AI-Powered Tools for Predicting and Optimizing High-Entropy Alloy Properties |
| Author(s) |
You-Xun Kevin Chang, Kuan-Hung Chao, Yi-Hsiang Chen, Fan-Yi Ouyang, Shun-Chi Wu |
| On-Site Speaker (Planned) |
You-Xun Kevin Chang |
| Abstract Scope |
High entropy alloys (HEAs) exhibit outstanding mechanical strength, thermal stability, and corrosion resistance, making them highly desirable for advanced engineering applications. However, their vast compositional diversity and complex property interactions present significant challenges for design and production. To overcome the challenges, this study introduces an agentic AI for efficient HEA discovery, with a specific focus on predicting and optimizing Young’s modulus. At the core of this approach is a large language model, specifically GPT, which intelligently extracts high-quality experimental data on HEA compositions and their mechanical properties from scientific literature. This automation speeds up dataset creation and ensures comprehensive coverage. Using this dataset, XGBoost predicts Young’s modulus precisely, while a genetic algorithm optimizes atomic combinations for specific targets. Experiments reveal less than 10% error between predictions and measurements, showcasing the design agent's accuracy. This agentic AI transforms HEA design, facilitating rapid, data-driven exploration of advanced materials for diverse applications. |
| Proceedings Inclusion? |
Undecided |
| Keywords |
Computational Materials Science & Engineering, Machine Learning, High-Entropy Alloys |