| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Modeling, Simulation and Machine Learning
|
| Presentation Title |
Can Machine Learning Predict the Liquidus Temperature of Binary Alloys? |
| Author(s) |
Yifei He |
| On-Site Speaker (Planned) |
Yifei He |
| Abstract Scope |
We use random forest and consider various feature vectors of the alloy based on known information to predict liquidus temperature of phase diagrams, TL. We found that when features based on physical insights into alloys’ mixing are used, the prediction with 8% error can be achieved compare to 13% when only using the properties of elements as features. The poor predictability even under the best circumstances is most dramatically reflected in the fact that even when over 99.8% of data considered for training of the algorithm, the error of prediction into the remaining 0.2% is only 8%. The major challenges in predicting the TL through ML algorithms originates from the challenge to represent the characteristics of alloys through which we argue is a common challenge in complex alloys. Further, the discreteness of atoms and corresponding features, constitutes the most fundamental challenge in applying ML strategies for complex materials science problems. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Computational Materials Science & Engineering, |