Data-Driven Learning of the Generalized Langevin Equation with State-Dependent Memory
Pei Ge, Zhongqiang Zhang, Huan Lei
Abstract
We present a data-driven method to learn stochastic reduced models of complex systems that retain a state-dependent memory beyond the standard generalized Langevin equation with a homogeneous kernel. The constructed model naturally encodes the heterogeneous energy dissipation by jointly learning a set of state features and the non-Markovian coupling among the features. Numerical results demonstrate the limitation of the standard generalized Langevin equation and the essential role of the broadly overlooked state-dependency nature in predicting molecule kinetics related to conformation relaxation and transition.
Topics & Concepts
Langevin equationStatistical physicsKernel (algebra)Dependency (UML)Markov processDissipationRelaxation (psychology)Langevin dynamicsState (computer science)PhysicsCoupling (piping)HomogeneousComputer scienceApplied mathematicsMathematicsArtificial intelligenceAlgorithmThermodynamicsMaterials sciencePure mathematicsStatisticsPsychologyMetallurgySocial psychologyProtein Structure and DynamicsSpectroscopy and Quantum Chemical StudiesMachine Learning in Materials Science