Path-space information metrics and variational inference for coarse-grained non-equilibrium steady states
Tuesday, April 10, 2018 - 2:00pm - 3:00pm
We discuss information-theoretic tools for obtaining optimized coarse-grained molecular models for systems at equilibrium as well as with non-equilibrium steady states. The presented approach compares microscopic behavior of molecular systems to parametric or non-parametric coarse-grained models using the relative entropy between distributions on the path space. It allows us to formulate a corresponding path space variational inference problem and obtain controlled approximation of observables on the coarse-grained model. The methods become entirely data-driven when the microscopic dynamics are replaced with corresponding correlated data in the form of time series.