I develop physics-informed and hybrid ML models that combine simulations and experimental signals for inverse problems, metrology, and defect inspection in semiconductor manufacturing. I'm also interested in interpretability and uncertainty quantification for reliable, auditable AI.
Selected Highlights
Recent publications and patents in Scientific Machine Learning
Research
Explore my research pillars in Scientific Machine Learning, interpretable AI, and data-driven discovery in physical sciences.
Publications
Browse my peer-reviewed publications in machine learning, materials science, and computational physics.
Patents
View my patent portfolio covering ML-driven metrology, defect inspection, and image enhancement for semiconductor manufacturing.
Featured Publications
NLS: An accurate and yet easy-to-interpret prediction method
Neural Networks, 2023
Explainable machine learning algorithms to predict glass transition temperature
Acta Materialia, 2020
MeLIME: Meaningful local explanation for machine learning models
arXiv preprint, 2020
Text-mined dataset of inorganic materials synthesis recipes
Scientific Data, 2019
Towards rational design of carbon nitride photocatalysts of cyanamide 'defects' as catalytically relevant sites
Nature Communications, 2016
Featured Patents
Method and System for Predicting Metrology Data Using a Neural Network
WO2025247596A1 · 2025
Defect Identification and Segmentation Without Labeled Data
EP4607460A2 · 2025
Measurement Charge Removal Using Diffusion Models
EP4607453A2 · 2025
Contrastive Deep Learning for Defect Inspection
WO2025108661A1 · 2025