Physics-informed graph neural networks enhance scalability of variational nonequilibrium optimal control
arXiv:2204.05493 · doi:10.1063/5.0095593
Abstract
When a physical system is driven away from equilibrium, the statistical distribution of its dynamical trajectories informs many of its physical properties. Characterizing the nature of the distribution of dynamical observables, such as a current or entropy production rate, has become a central problem in nonequilibrium statistical mechanics. Asymptotically, for a broad class of observables, the distribution of a given observable satisfies a large deviation principle when the dynamics is Markovian, meaning that fluctuations can be characterized in the long-time limit by computing a scaled cumulant generating function. Calculating this function is not tractable analytically (nor often numerically) for complex, interacting systems, so the development of robust numerical techniques to carry out this computation is needed to probe the properties of nonequilibrium materials. Here, we describe an algorithm that recasts this task as an optimal control problem that can be solved variationally. We solve for optimal control forces using neural network ansätze that are tailored to the physical systems to which the forces are applied. We demonstrate that this approach leads to transferable and accurate solutions in two systems featuring large numbers of interacting particles.
arXiv comment23 pages, 6 figures
References in corpus (54)
- Deep Potential Molecular Dynamics: a scalable model with the accuracy of quantum mechanics
- SchNet - a deep learning architecture for molecules and materials
- DeePMD-kit: A deep learning package for many-body potential energy representation and molecular dynamics
- The large deviation approach to statistical mechanics
- Motility-Induced Phase Separation
- Flocks, herds, and schools: A quantitative theory of flocking
- A Gallavotti-Cohen Type Symmetry in the Large Deviation Functional for Stochastic Dynamics
- Thermodynamic uncertainty relation for biomolecular processes
- Long-lived Giant Number Fluctuations in a Swarming Granular Nematic
- Dissipation bounds all steady-state current fluctuations
- Non-equilibrium statistical mechanics: From a paradigmatic model to biological transport
- Dynamic first-order phase transition in kinetically constrained models of glasses
- First-order dynamical phase transition in models of glasses: an approach based on ensembles of histories
- Defect annihilation and proliferation in active nematics
- Direct evaluation of large-deviation functions
- Statistical Mechanics of Active Ornstein Uhlenbeck Particles
- Introduction to dynamical large deviations of Markov processes
- A numerical approach to large deviations in continuous-time
- Irreversibility and biased ensembles in active matter: Insights from stochastic thermodynamics
- Variational and optimal control representations of conditioned and driven processes
- Population dynamics method with a multi-canonical feedback control
- Effective interactions and large deviations in stochastic processes
- Optimizing active work: Dynamical phase transitions, collective motion, and jamming
- Go with the Winners: a General Monte Carlo Strategy
- Finite-Size Scaling of a First-Order Dynamical Phase Transition: Adaptive Population Dynamics and an Effective Model
- Large fluctuations and dynamic phase transition in a system of self-propelled particles
- Learning force fields from stochastic trajectories
- Using matrix product states to study the dynamical large deviations of kinetically constrained models
- Response theory: a trajectory-based approach
- Exact fluctuations of nonequilibrium steady states from approximate auxiliary dynamics
- Learning to grow: control of material self-assembly using evolutionary reinforcement learning
- Entropy production fluctuations encode collective behavior in active matter
- How dissipation constrains fluctuations in nonequilibrium liquids: Diffusion, structure and biased interactions
- Dissipation controls transport and phase transitions in active fluids: Mobility, diffusion and biased ensembles
- A reinforcement learning approach to rare trajectory sampling
- Adaptive sampling of large deviations
- Computation of Large Deviation Statistics via Iterative Measurement-and-Feedback Procedure
- Collective motion in large deviations of active particles
- Dynamical phase behavior of the single- and multi-lane asymmetric simple exclusion process via matrix product states
- Learning nonequilibrium control forces to characterize dynamical phase transitions
- Variational control forces for enhanced sampling of nonequilibrium molecular dynamics simulations
- Reinforcement learning of rare diffusive dynamics
- Dynamical Phase Transitions in a 2D Classical Nonequilibrium Model via 2D Tensor Networks
- Direct evaluation of rare events in active matter from variational path sampling
- Evolutionary reinforcement learning of dynamical large deviations
- Direct evaluation of dynamical large-deviation rate functions using a variational ansatz
- Dynamical large deviations of two-dimensional kinetically constrained models using a neural-network state ansatz
- Efficiency of one-dimensional active transport conditioned on motility
- Dynamical computation of the density of states and Bayes factors using nonequilibrium importance sampling
- Heterogeneity-induced large deviations in activity and (in some cases) entropy production
- Thermodynamic Control of Activity Patterns in Cytoskeletal Networks
- Constructing Auxiliary Dynamics for Nonequilibrium Stationary States by Variance Minimization
- Exact Large Deviations of the Current in the Asymmetric Simple Exclusion Process with Open Boundaries
- Watch and learn -- a generalized approach for transferrable learning in deep neural networks via physical principles