Abstract:Neural-network quantum molecular dynamics (NNQMD) simulations based on machine learning are revolutionizing atomistic simulations of materials by providing quantum-mechanical accuracy but orders-of-magnitude faster, illustrated by ACM Gordon Bell prize (2020) and finalist (2021). State-of-the-art (SOTA) NNQMD model founded on group theory featuring rotational equivariance and local descriptors has provided much higher accuracy and speed than those models, thus named Allegro (meaning fast). On massively parallel supercomputers, however, it suffers a fidelity-scaling problem, where growing number of unphysical predictions of interatomic forces prohibits simulations involving larger numbers of atoms for longer times. Here, we solve this problem by combining the Allegro model with sharpness aware minimization (SAM) for enhancing the robustness of model through improved smoothness of the loss landscape. The resulting Allegro-Legato (meaning fast and "smooth") model was shown to elongate the time-to-failure $t_\textrm{failure}$, without sacrificing computational speed or accuracy. Specifically, Allegro-Legato exhibits much weaker dependence of timei-to-failure on the problem size, $t_{\textrm{failure}} \propto N^{-0.14}$ ($N$ is the number of atoms) compared to the SOTA Allegro model $\left(t_{\textrm{failure}} \propto N^{-0.29}\right)$, i.e., systematically delayed time-to-failure, thus allowing much larger and longer NNQMD simulations without failure. The model also exhibits excellent computational scalability and GPU acceleration on the Polaris supercomputer at Argonne Leadership Computing Facility. Such scalable, accurate, fast and robust NNQMD models will likely find broad applications in NNQMD simulations on emerging exaflop/s computers, with a specific example of accounting for nuclear quantum effects in the dynamics of ammonia.

Enhancing Open Quantum Dynamics Simulations Using Neural Network-Based Non-Markovian Stochastic Schrödinger Equation Method

Approach to Solving Spin-Boson Dynamics Via Non-Markovian Quantum Trajectories

Neural Network Approach for Non-Markovian Dissipative Dynamics of Many-Body Open Quantum Systems

Quantum neural network approach to Markovian dissipative dynamics of many-body open quantum systems

Quantum Simulation of Open Quantum Dynamics via Non-Markovian Quantum State Diffusion

Non-markovian neural quantum propagator and its application to the simulation of ultrafast nonlinear spectra

Neural-network Variational Quantum Algorithm for Simulating Many-Body Dynamics

Unsupervised Deep Neural Network Approach To Solve Fermionic Systems

Simulating Open Quantum System Dynamics on NISQ Computers with Generalized Quantum Master Equations

Quantum Neural Network Inspired Hardware Adaptable Ansatz for Efficient Quantum Simulation of Chemical Systems

Solving Quantum Master Equations with Deep Quantum Neural Networks

Automatic Evolution of Machine-Learning based Quantum Dynamics with Uncertainty Analysis

Simulation of Open Quantum Dynamics with Bootstrap-Based Long Short-Term Memory Recurrent Neural Network

Neural-network quantum states for ultra-cold Fermi gases

AI for Quantum Mechanics: High Performance Quantum Many-Body Simulations Via Deep Learning

Simulating Quantum Many-Body States with Neural-Network Exponential Ansatz

A Non-stochastic Optimization Algorithm for Neural-network Quantum States

Bridging the Gap between Deep Learning and Frustrated Quantum Spin System for Extreme-scale Simulations on New Generation of Sunway Supercomputer

Capturing Non-Markovian Dynamics on Near-Term Quantum Computers

Allegro-Legato: Scalable, Fast, and Robust Neural-Network Quantum Molecular Dynamics via Sharpness-Aware Minimization