A novel machine learning enabled hybrid optimization framework for efficient and transferable coarse-graining of a model polymer
This work presents a novel framework governing the development of an efficient, accurate, and transferable coarse-grained (CG) model of a polyether material. The proposed framework combines the two fundamentally different classical optimization approaches for the development of coarse-grained model parameters; namely bottom-up and top-down approaches. This is achieved through integrating the optimization algorithms into a machine learning (ML) model, trained using molecular dynamics (MD) simulation data. In the bottom-up approach, bonded interactions of the CG model are optimized using deep neural networks (DNN), where atomistic bonded distributions are matched. The atomistic distributions emulate the local chain structure. In the top-down approach, optimization of nonbonded potentials is accomplished by reproducing the temperature-dependent experimental density. We demonstrate that CG model parameters achieved through our machine-learning enabled hybrid optimization framework fulfills the thermodynamic consistency and transferability issues associated with the classical approaches to coarse-graining model polymers. We demonstrate the efficiency, accuracy, and transferability of the developed CG model, using our novel framework through accurate predictions of chain size as well as chain dynamics, including the limiting behavior of the glass transition temperature, diffusion, and stress relaxation spectrum, where none were included in the potential parameterization process. The accuracy of the predicted properties are evaluated in the context of molecular theories and available experimental data.