Reduced dimensionality representations of soft matter
Many phenomena in soft condensed matter occur on length and/or time scales that are unfeasible to study theoretically or computationally if every atom in the system is represented explicitly.
One strategy to address this issue to simplify the representation of a molecular system by systematically eliminating unimportant degrees-of-freedom to improve computational efficiency without sacrificing accuracy.
This project will use methods from statistical mechanics, machine learning, and artificial intelligence to develop algorithms to automate the dimensionality reduction process.
- Identifying the optimal mapping of a molecular system onto a smaller number of ‘coarse-grained’‚ interaction sites, with a focus on anisotropic molecules; or
- Identifying low-dimensional equations of motion from the trajectory of a high-dimensional dynamical system.