Building machine learning and quantum solutions for chemistry and materials problems

I work on the interface of chemistry, applied math, and computer science as a research assistant here at MIT, with an overall theme of applying machine learning to improve the computational high throughput screening workflow and accelerate chemical discovery. Specifically, I* Developed the first set of machine learning classifiers that monitor quantum chemistry calculations on the fly in computational high throughput screening, saving more than half of the computational resources and time that would have beed wasted on failed calculations* Developed the first semi-supervised learning classifier to identify strong static correlation in materials, achieving state-of-the-art for this classification task* Integrated transfer learning and uncertainty quantification in computational high throughput screening, reducing the error of machine learning accelerated chemical discovery to 1 kcal/mol chemical accuracy* Discovered functional materials with multi-objective active learning, such as redox couples in redox flow battery, single-site catalysts for methane-to-methanol conversion, and robust transition metal chromophores

I have been developing a recommender for selecting density functional approximations in quantum chemistry calculations to achieve a higher accuracy in virtual high throughput screening

I uncovered novel heat transport behaviors in non-commutative quantum heat engine with heat-flux extended hierarchical equation of motion.

I enabled numerical-exact calculations of open quantum dynamics via extending the domain of applicability of hierarchical equation of motion, and studied the quantum phase transition of the spin-boson model