Research Projects

This project introduces a novel and computationally efficient framework that combines first-principles calculations with machine learning (ML) to accelerate the discovery of quantum defects. The central goal is to build a scalable computational pipeline capable of predicting key thermodynamic, electronic, optical, and magnetic properties of defects. By training ML surrogate models on a curated dataset of defect properties, we will implement an active learning strategy to iteratively identify the most informative regions of the defect space. This approach will continuously expand a database of promising quantum defects and their associated properties. Candidates identified through this pipeline will undergo further analysis using advanced theoretical methods such as time-dependent density functional theory (TDDFT) and quantum defect embedding theory (QDET). Additional studies will explore surface effects and external stimuli, including magnetic fields and pressure, to support experimental validation and quantum sensing applications. The resulting database and ML models will also benefit other domains where point defects are critical, such as photovoltaics and optoelectronics. The novelty of this approach lies in its seamless integration of high-fidelity simulations with data-driven learning, enabling systematic and efficient exploration of the vast defect landscape.