Abstract:While there have been many developments in computational probes of both strongly-correlated molecular systems and machine-learning accelerated molecular dynamics, there remains a significant gap in capabilities in simulating accurate non-local electronic structure over timescales on which atoms move. We develop an approach to bridge these fields with a practical interpolation scheme for the correlated many-electron state through the space of atomic configurations, whilst avoiding the exponential complexity of these underlying electronic states. With a small number of accurate correlated wave functions as a training set, we demonstrate provable convergence to near-exact potential energy surfaces for subsequent dynamics with propagation of a valid many-body wave function and inference of its variational energy whilst retaining a mean-field computational scaling. This represents a profoundly different paradigm to the direct interpolation of potential energy surfaces in established machine-learning approaches. We combine this with modern electronic structure approaches to systematically resolve molecular dynamics trajectories and converge thermodynamic quantities with a high-throughput of several million interpolated wave functions with explicit validation of their accuracy from only a few numerically exact quantum chemical calculations. We also highlight the comparison to traditional machine-learned potentials or dynamics on mean-field surfaces.

What problem does this paper attempt to address?

This paper aims to address the problem of accurately describing non-local electronic structures in molecular dynamics simulations, particularly maintaining high precision on the timescale of atomic motion. Specifically, the research objectives can be summarized as follows: 1. **Bridging the gap between electronic and atomic timescales**: Traditional electronic structure calculation methods can provide high-precision results but only for fixed atomic configurations. Molecular dynamics simulations, however, need to consider the changes in atomic positions over time, requiring numerous continuous electronic structure calculations on the timescale of atomic motion, which is computationally expensive for high-precision methods. 2. **Avoiding exponential complexity**: By developing a practical interpolation scheme to handle the variations of related multi-electron states across the entire atomic configuration space, this method avoids the exponential complexity of underlying electronic states. 3. **Direct interpolation of wave functions**: A different paradigm is proposed, which involves interpolating the many-body electronic wave functions instead of the potential energy surface itself. This allows the propagation of effective many-body wave functions during subsequent dynamics and infers their variational energy while retaining the complexity of mean-field calculations. 4. **Systematic improvability**: By interpolating a small number of accurate correlated wave functions between different atomic configurations as a training set, it is demonstrated that this method can converge to an approximately accurate potential energy surface, achieving systematic improvability in molecular dynamics simulations. 5. **Combining with modern electronic structure methods**: The above interpolation wave function method is combined with modern Density Matrix Renormalization Group (DMRG) methods to achieve convergence to approximate accuracy for strongly correlated potential energy surfaces. 6. **Efficient molecular dynamics simulations**: This method enables molecular dynamics simulations on practical timescales, addressing the computational cost issues of traditional methods on these timescales. In short, this paper proposes a novel approach to overcome the challenges of accurately describing electronic structures in molecular dynamics simulations, particularly on the timescale of atomic motion, by interpolating many-body wave functions to accelerate the computational process while ensuring the accuracy of the results.

Interpolating many-body wave functions for accelerated molecular dynamics on the near-exact electronic surface

Fast and accurate nonadiabatic molecular dynamics enabled through variational interpolation of correlated electron wavefunctions

A combined on-the-fly/interpolation procedure for evaluating energy values needed in molecular simulations

Accurate interatomic force field for molecular dynamics simulation by hybridizing classical and machine learning potentials

Electronic coarse graining enhances the predictive power of molecular simulation allowing challenges in water physics to be addressed

Scalable fine-grained parallelization of plane-wave-based ab initio molecular dynamics for large supercomputers

Accurate Many-Body Electronic Structure Near the Basis Set Limit: Application to the Chromium Dimer

Enabling Large-Scale Condensed-Phase Hybrid Density Functional Theory Based $Ab$ $Initio$ Molecular Dynamics I: Theory, Algorithm, and Performance

Atom-wise formulation of the many-body dispersion problem for linear-scaling van der Waals corrections

Simulation of surface processes

Refining Potential Energy Surface through Dynamical Properties via Differentiable Molecular Simulation

Ab-initio variational wave functions for the time-dependent many-electron Schrödinger equation

Towards Exact Molecular Dynamics Simulations with Machine-Learned Force Fields

In operando active learning of interatomic interaction during large-scale simulations

Extending the atomic decomposition and many-body representation, a chemistry-motivated monomer-centered approach for machine learning potentials

Trajectory Surface Hopping with Tight Binding Density Functional Theory applied to Molecular Motors

Semiclassical Monte-Carlo approach for modelling non-adiabatic dynamics in extended molecules

An Efficient and Accurate Car-Parrinello-like Approach to Born-Oppenheimer Molecular Dynamics

By-passing the Kohn-Sham equations with machine learning

Investigating the Behavior of Diffusion Models for Accelerating Electronic Structure Calculations