Abstract:Macroscopic observables of a system are of keen interest in real applications such as the design of novel materials. Current methods rely on microscopic trajectory simulations, where the forces on all microscopic coordinates need to be computed or measured. However, this can be computationally prohibitive for realistic systems. In this paper, we propose a method to learn macroscopic dynamics requiring only force computations on a subset of the microscopic coordinates. Our method relies on a sparsity assumption: the force on each microscopic coordinate relies only on a small number of other coordinates. The main idea of our approach is to map the training procedure on the macroscopic coordinates back to the microscopic coordinates, on which partial force computations can be used as stochastic estimation to update model parameters. We provide a theoretical justification of this under suitable conditions. We demonstrate the accuracy, force computation efficiency, and robustness of our method on learning macroscopic closure models from a variety of microscopic systems, including those modeled by partial differential equations or molecular dynamics simulations.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to accurately learn macro - dynamics when only some microscopic coordinate forces are obtained. Specifically, current methods rely on calculating or measuring the forces of all microscopic coordinates, which is very computationally expensive for real - world systems. Therefore, this paper proposes a new method to learn macro - dynamics by calculating the forces of only some microscopic coordinates, thereby significantly reducing the computational cost. ### Problem Background Macroscopic observables (such as the design of new materials) are very important in practical applications, but they usually depend on the state of the entire microscopic system. For example, calculating the mean - square displacement requires information on all microscopic coordinates during the simulation process. With the increase in simulation and experimental data, learning macroscopic properties from microscopic observations has become an active research area. However, for the accurate calculation of macroscopic properties of large - scale systems, the computational cost is extremely high because accurate force calculations for all microscopic coordinates are required. In particular, in ab - initio molecular simulations, accurate forces need to be calculated through density functional theory (DFT), which limits its application to relatively small systems (several hundred to several thousand atoms). ### Proposed Method To solve this dilemma, this paper proposes a new method that can learn macro - dynamics when only some microscopic coordinate forces are obtained. This method is based on the sparsity assumption: the force on each microscopic coordinate depends only on a few other coordinates. The specific steps are as follows: 1. **Mapping the training process**: Map the training process on the macroscopic coordinates back to the microscopic coordinates, and use partial force calculations as random estimates to update the model parameters. 2. **Sparsity assumption**: Assume that the force on each microscopic coordinate depends only on a few other coordinates, so that the cost of partial force calculations does not increase with the size of the microscopic system. 3. **Theoretical proof**: Under appropriate conditions, it is theoretically proven that the forces on some microscopic coordinates can be used as random estimates to update the underlying model parameters. ### Experimental Verification The authors verified the accuracy, force - calculation efficiency, and robustness of this method through a series of experiments. The experiments include: - **Predator - prey system**: Verified the correctness and force - calculation efficiency of the method. - **Lennard - Jones system**: Verified the robustness of the method to different underlying structures and demonstrated its applicability in larger systems. - **Allen - Cahn system**: Further verified the robustness of the method to systems with larger microscopic dimensions. ### Main Contributions 1. **New method**: Developed a new method that can learn macro - dynamics from partial microscopic force calculations, significantly reducing the force - calculation cost. 2. **Theoretical proof**: Theoretically proven that the forces on some microscopic coordinates can be used as random estimates to update the underlying model parameters. 3. **Experimental verification**: Verified the accuracy, force - calculation efficiency, and robustness of the method through various microscopic dynamics and underlying model structures. ### Conclusion This paper successfully solves the high - cost problem of calculating all microscopic coordinate forces in large - scale systems and provides an efficient and accurate method for learning macro - dynamics.

Learning Macroscopic Dynamics from Partial Microscopic Observations

Learning Effective Molecular Models from Experimental Observables.

Systematic parameter inference in stochastic mesoscopic modeling

Accurate macroscale modelling of spatial dynamics in multiple dimensions

Learning fast, accurate, and stable closures of a kinetic theory of an active fluid

Couple microscale periodic patches to simulate macroscale emergent dynamics

OnsagerNet: Learning Stable and Interpretable Dynamics Using a Generalized Onsager Principle

Estimation of spatial and time scales of collective behaviors of active matters through learning hydrodynamic equations from particle dynamics

Incorporating physical constraints in a deep probabilistic machine learning framework for coarse-graining dynamical systems

Learning dynamical models from stochastic trajectories

Data-driven Closures & Assimilation for Stiff Multiscale Random Dynamics

On-the-fly active learning of interpretable Bayesian force fields for atomistic rare events

Machine learning of microscopic structure-dynamics relationships in complex molecular systems

Active- and transfer-learning applied to microscale-macroscale coupling to simulate viscoelastic flows

Learning the Non-Equilibrium Dynamics of Brownian Movies

Variational system identification of the partial differential equations governing microstructure evolution in materials: Inference over sparse and spatially unrelated data

Interface learning in fluid dynamics: statistical inference of closures within micro-macro coupling models

Learning macroscopic parameters in nonlinear multiscale simulations using nonlocal multicontinua upscaling techniques

Quantum Mechanics for Closure of Dynamical Systems

Data-driven Effective Modeling of Multiscale Stochastic Dynamical Systems