AI/ML-Accelerated Discovery
Adaptive feedback, deep-learning screening, and scalable workflows for exploring large composition and structure spaces.
Computational Materials Science · AI for Discovery
Weiyi Xia
Scientist I at Ames National Laboratory developing AI/ML-assisted, first-principles, and exascale workflows to discover and understand functional materials.
My work bridges predictive computation and experimental validation across magnetic materials, rare-earth and rare-earth-free compounds, intermetallics, ML interatomic potentials, and GW/BSE methodology.
40+
publications across first-principles methods and AI-guided discovery
3
public materials data resources for magnetic, ML, and rare-earth systems
2025
Ames National Laboratory Inventor Incentive Award
Ph.D.
Condensed Matter Physics, University at Buffalo
Research Direction
AI-guided materials discovery with quantum-mechanical grounding.
I develop computational strategies that reduce the distance between a promising composition and a materials claim that can be checked, reproduced, and eventually synthesized. The common thread is a closed loop: generate candidates, learn from first-principles calculations, interrogate physical mechanisms, and feed the result back into the next search.
Magnetic Materials
Rare-earth-free magnets, Fe-Co-X compounds, magnetic anisotropy, spin-state physics, and noncollinear magnetic order.
Interatomic Potentials
Machine-learning potentials for phase stability, forces, energies, and synthesis or growth kinetics in complex materials.
Electronic Structure Methods
Large-scale GW/BSE methodology and quasiparticle/optical-property calculations for low-dimensional and bulk materials.
Scientific Highlights
Representative arcs from prediction to mechanism.
PNAS · npj Comput. Mater. · JMCA · PRM
Closed-loop discovery workflows
Integrated ML screening, adaptive feedback, DFT validation, and high-performance workflows for accelerated materials design.
PNAS · Fe-Co-B
Prediction-to-validation discovery
An adaptive loop connecting candidate generation, ML prioritization, first-principles checks, and experimental validation.
npj Comput. Mater. · La-Co-Pb
Immiscible-element structure search
Machine learning turns an antagonistic chemical space into a navigable map of low-energy crystal motifs.
Selected Publications
A publication record spanning methods, magnets, and materials discovery.
The selected work below emphasizes papers that anchor the current research program: AI-assisted discovery, magnetic materials, exascale workflows, intermetallic compounds, and large-scale electronic-structure methodology.
- 2026 AI-driven and quantum-informed searches for rare-earth-free magnets: Applications to Fe-Co-X (X = B, C, P, Si, S) compounds. J. Magn. Magn. Mater.
- 2026 Accelerated discovery and design of Fe-Co-Zr magnets with tunable magnetic anisotropy through machine learning and parallel computing. Phys. Rev. Mater.
- 2025 Machine-learning and first-principles investigation of lightweight medium-entropy alloys for hydrogen-storage applications. Int. J. Hydrogen Energy
- 2025 Synthesis challenges, thermodynamic stability, and growth kinetics of La-Si-P ternary compounds. J. Mater. Chem. A
- 2025 Machine learning accelerated prediction of Ce-based ternary compounds involving antagonistic pairs. Phys. Rev. Mater.
- 2025 Manipulating ambient pressure superconductivity in metal borocarbides through hole doping. J. Phys.: Condens. Matter
- 2024 Machine learning assisted search for Fe-Co-C ternary compounds with high magnetic anisotropy. APL Machine Learning
- 2024 Giant magnetic anisotropy of Pb atoms in 3d-based magnets. Phys. Rev. B
- 2023 Accelerating materials discovery using integrated deep machine learning approaches. J. Mater. Chem. A
- 2023 Machine learning guided discovery of ternary compounds involving La and immiscible Co and Pb elements. npj Comput. Mater.
- 2022 Accelerating the discovery of novel magnetic materials using machine learning-guided adaptive feedback. PNAS
- 2022 Predicting magnetic anisotropy energies using site-specific spin-orbit coupling energies and machine learning: Application to iron-cobalt nitrides. Phys. Rev. Mater.
- 2020 Combined subsampling and analytical integration for efficient large-scale GW calculations for 2D systems. npj Comput. Mater.
- 2019 Prediction of MXene based 2D tunable band gap semiconductors: GW quasiparticle calculations. Nanoscale
Current Scientific Directions
Materials questions now shaping the next papers and proposals.
Exascale AI for materials discovery
Scalable workflows that combine AI/ML models, DFT validation, phase-diagram construction, and community-ready software.
Critical-element-conscious magnets
Rare-earth-free and critical-element-reduced magnetic materials with attention to anisotropy, stability, and manufacturability.
Synthesis-aware predictions
Free-energy, kinetics, and experimental refinement as first-class constraints in the design of new intermetallic compounds.
Magnetic ground-state complexity
Collinear, noncollinear, and spin-spiral calculations for systems where magnetic frustration changes the materials story.
Open Resources
Databases and reusable research infrastructure.
Trajectory
From electronic-structure methods to AI-assisted discovery.
Scientist I, Ames National Laboratory
AI/ML-assisted discovery, exascale workflows, magnetic materials, and synthesis-aware computational design.
Postdoctoral Research Associate, Ames National Laboratory and Iowa State University
Adaptive genetic algorithms, ML-guided materials discovery, magnetic systems, and experimental collaborations.
Ph.D. in Condensed Matter Physics, University at Buffalo
Large-scale GW quasiparticle calculations for 3D and 2D materials: methodology development and applications.
Contact