Research

General

I am a computational chemist focusing on the application of machine learning to high-throughput materials screening problems. I focus on researching and applying Gaussian processes machine learning methodologies, implementing techniques described in research papers and designing experiments. I am currently working in the field of materials discovery for Hydrogen Fuel Cell (HFC) technologies, researching nanoparticle systems of potential interest that address shortcomings in stability and activity at the cathode for catalyzing the oxygen reduction reaction. I am particularly interested in global optimization of nanoparticles using search techniques such as genetic algorithms. This allows for the investigation of realistic structures before assessing catalytic activity.

I focus on applications in the limit of small data. Typically the problems we wish to address don’t have large datasets and we can greatly improve the efficiency of studies through an iterative selection of the best calculations to perform next. I apply Gaussian processes regression to determine which calculations to perform that are likely to either be optimal solutions to the problem or develop the underlying model. This is achieved through the employ of different acquisition functions coupled with uncertainty estimates on all predictions.