Our lab works at the interface between computation, statistics and material science. We are specifically interested in making machine learning techniques effective in the regime of sparse and noisy data in which materials science research is frequently conducted. In this regime, we have to be: 1) economical with the data that we do have, 2) strategic in acquiring new data, and 3) resourceful in incorporating information from other sources.

Major Research Thrusts

  • Decision-making for autonomous and semi-autonomous research
  • Prior knowledge formation, elicitation and representation for physics-based systems
  • Machine-learning-enabled molecular dynamics simulations