Speaker
Description
Diverse soft matter systems offer the possibility to tune the rheological response by tailoring microscopic interactions; however, brute-forcing the exploration may be either impractical or extremely costly. Here, we present an efficient approach for the exploration of the parameter space of soft matter systems based on the synergistic combination of coarse-grained modeling, Brownian dynamics simulations and machine learning. As a case study, we choose DNA-based associative fluids and we employ a minimal coarse-grained model, whose predictions are compared to experimentally available rheological curves. The model exhibits a viscoelastic response, characterized by a cross-over from viscous to elastic behavior. Coupling simulations with Gaussian Process Regression and active learning, we explore the design space with high predictive accuracy. Benchmarking against experimental DNA hydrogel data demonstrates that the model captures essential rheological behavior in the strongly associated regime and defines clear limits of applicability.