Active subspace learning for coarse-grained molecular dynamics.

ABSTRACT: We introduce Active Subspace Coarse-Graining (ASCG), an interpretable framework for systematic bottom-up coarse-graining trained from atomistic molecular dynamics simulations that simultaneously defines the coarse-grained mapping, effe ctive interactions, and the equations of motion within one unified mathematical framework. We employ active subspace learning to identify linear projections of atomistic degrees of freedom that maximally describe gradients of the potential energy, yielding a reduced set of coarse-grained variables that capture the dominant collective motions across the potential of mean force. Effective coarse-grained forces and noise terms are obtained directly from the same projection, eliminating the need for separate parameterization schemes. We demonstrate the ASCG meth

SUBMITTER: Wojnar A 

PROVIDER: S-EPMC12633002 | biostudies-literature | 2025 Dec

REPOSITORIES: biostudies-literature