Theory, Computation and Machine Learning

Several of our faculty members are involved in doing fundamental research using theoretical and numerical tools. Some of the examples involve predicting laminar to turbulent transition in pipe flows for complex fluids, pattern formation in thin film flows due to instability, models to understand cell biology, etc. The computational techniques and methods employed and/or developed vary from electronic to continuum levels.

Molecular simulations have become a powerful tool to understand and predict the structural, thermodynamical, and dynamical properties of materials. Our department has exceptional depth and breadth in the area of statistical mechanics methods and molecular simulation techniques ranging from quantum chemistry to molecular dynamics, Monte Carlo, and coarse-grained methods. Several of our faculty also work in the area of multiscale simulations. The materials being investigated include soft materials, such as polymers and colloids, ionic materials, composite materials, semiconductors, metals, and liquid systems. The emphasis on developing a fundamental understanding of a range of problems, such as understanding the thermodynamic aspects of phase transitions in the condensed phases; the effect of polar, hydrogen bond, and hydrophobic interactions on the structural and dynamical phenomena at the nanoscale; self-assembly in soft condensed, and design of new catalysts.

Few of us are utilizing machine learning/AI methods to accelerate the learning and discovery of materials such as suitable MOF for carbon capture or catalyst for CO2 conversion. The goal is to develop tools that integrate different AI/ML with molecular simulation methods, to accelerate learning, rationale design of materials, and create a pipeline for high-throughput screening of materials/molecules.