Overview

Our research focuses on two main directions: AI for Materials Science and Computational Materials Science.
We develop data-driven and physics-based models for alloy design, microstructure prediction, phase stability, and property optimization,
while maintaining physical interpretability and scientific insight.

2 research themes
AI for Materials Science representative image
AI for Materials Science

AI for Materials Science

Physics-based machine learning and explainable AI for materials design

We develop interpretable and physics-based machine learning frameworks for alloy design, process optimization, and structure-property prediction. Our work emphasizes scientific interpretability, physically meaningful descriptors, and reliable inverse design strategies for accelerating materials discovery.

Materials AIPhysics-based Machine LearningExplainable AIInverse DesignAlloy Design
Graph Neural NetworksXGBoostSHAPNeural NetworksInterpretable Modeling
High-Entropy Alloy DesignHydrogen Storage Alloy DiscoveryAdditive Manufacturing OptimizationSolidification PredictionMicrostructure-aware Design
Computational Materials Science representative image
Computational Materials Science

Computational Materials Science

Thermodynamics, atomistic simulation, and multiscale computational modeling

We use computational materials science tools to understand alloy thermodynamics, phase stability, atomistic behavior, and microstructure evolution. Our research integrates CALPHAD, molecular dynamics, interatomic potentials, machine-learning potentials, and density functional theory to build predictive and physically grounded frameworks for advanced materials design.

Computational Materials ScienceThermodynamicsAtomistic SimulationMultiscale ModelingPhase Stability
CALPHADMolecular Dynamics2NN MEAMMachine Learning PotentialDFT
Phase Diagram and Thermodynamic ModelingInteratomic Potential DevelopmentMicrostructure EvolutionDiffusion and PrecipitationComposition-Structure-Property Relations
JSON