How can materials dynamically control or remodel their own internal structure to affect their behavior? How can the statistics of structural disorder be biased to produce non-trivial properties? How can one discover novel equilibrium and non-equilibrium assembly mechanisms in highly parameterized systems? Questions like these are a necessary step in the development of synthetic biology, where non-biological materials and nano-scale machines operate with the complexity and functionality found only in biology.
Towards this end, the Goodrich Group uses computational and theoretical tools to discover basic soft matter principles that could one day lead to new functional materials as well as deepen our understanding of complex biological matter. The goal is to first understand general or even universal mechanisms that are not overly sensitive to the details of a given experimental system, and then work with experimentalists to test these ideas in practice. The group deploys and develops a number of numerical techniques, from molecular dynamics and Monte Carlo to machine learning and automatic differentiation. Specifically, the researchers are at the forefront in the development of trainable physics models, which provide a new and powerful way to explore high-dimensional systems and discover complex, non-trivial phenomena.