We study computational and statistical methods in materials science

Computation, statistics, data science, machine learning and artificial intelligence are at the heart of the effort to accelerate discovery and design in materials science. In the CSMS Lab, we work on both the fundamental development of the these techniques, as well as their application to real-world materials problems. We actively collaborate with experimentalists and theoreticians alike to accelerate their research, automate workflows, draw insight from data and simulation results, and develop solutions that merge data with prior scientific knowledge.

Research Focus Areas

Optimal Experimental Design and Bayesian Optimization

How to select experiments with some objective in mind, with the goal of minimizing the number of experiments needed to achieve the stated objective?

Prior Knowledge Formation, Elicitation, and Representation

Prior knowledge simplifies problems, reducing dimensionality and data requirements. How do we encode what scientists know about their problem, and quantify their confidence in this knowledge?

Machine Learning Enabled Atomistic Simulations

In addition to classical simulation work (such as kinetic Monte Carlo and molecular dynamics), we are interested in augmenting materials simulation with machine learning to gain computational speedups.

Featured Applications


ARES: The Robot Scientist


Working with materials scientists at the Air Force Research Labortory, we're using Optimal Experimental Design to provide the brains behind ARES, an autonomous research robot that can execute and characterize CVD experiments. Using our software, ARES autonomously runs experiments to maximize material properties.