.png){kind=icon}
.png){kind=icon}
Return
Machine-learned interatomic potentials for modeling multicomponent metallic systems: a comprehensive review
.png){kind=icon}
OA
AI
Y
H
A
P
M
R
DOI：10.1080/21663831.2026.2684721.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Multicomponent alloys demonstrate outstanding mechanical, chemical, and physical performance, but their vast compositional space cannot be efficiently navigated using conventional trial-and-error strategies. Progress is further hindered by the scarcity of accurate interatomic potentials for such complex chemistries, limiting predictive atomistic modeling. Machine-learned interatomic potentials (MLIPs) have recently emerged as powerful tools that deliver near-quantum accuracy while extending accessible length and time scales, thereby accelerating alloy design and discovery. This review examines recent advances in MLIP frameworks for multicomponent alloys, focusing on structural and thermodynamic predictions while highlighting key unresolved challenges, including transferability, extrapolation reliability, uncertainty quantification, and efficient dataset generation.
Keywords:
Machine-learned interatomic potentials
high-entropy and multicomponent alloys
atomistic simulations
active learning
workflows for MLIP development
Journal
IF:
7.9
Papers:
1.1K
Citations:
6.4K
Organization
Cited Papers