Toward High-level Machine Learning Potential for Water Based on Quantum Fragmentation and Neural Networks

Abstract

Accurate and efficient simulation of liquids, such as water and salt solutions, using high-level wave function theories is still a formidable task for computational chemists owing to the high computational costs. In this study, we develop a deep machine learning potential based on fragment-based second-order Møller–Plesset perturbation theory (DP-MP2) for water through neural networks. We show that the DP-MP2 potential predicts the structural, dynamical, and thermodynamic properties of liquid water in better agreement with the experimental data than previous studies based on density functional theory (DFT). The nuclear quantum effects (NQEs) on the properties of liquid water are also examined, which are noticeable in affecting the structural and dynamical properties of liquid water under ambient conditions. This work provides a general framework for quantitative predictions of the properties of condensed-phase systems with the accuracy of high-level wave function theory while achieving significant computational savings compared to ab initio simulations.

Data-Efficient Machine Learning Potentials from Transfer Learning of Periodic Correlated Electronic Structure Methods: Liquid Water at AFQMC, CCSD, and CCSD(T) Accuracy

“…Machine-learned potentials (MLPs) have emerged as an extremely promising approach to accurately model ab initio potential energy surfaces of condensed-phase systems while being orders of magnitude more computationally efficient to evaluate. For liquid water, MLPs have been successfully developed at various levels of electronic structure ranging from different levels of DFT − to, more recently, using the random phase approximation (RPA) and MP2. , The modeling of liquid water and other molecular systems with more accurate electronic structure methods, such as coupled-cluster theory or quantum Monte Carlo, has been limited, so far, to training on finite clusters of molecules. − When training on small clusters, higher-order many-body interactions must be included by other means such as by using the TTM4-F potential, as is done for the MB-Pol water model. − Other cluster-based models for water have gone on to explicitly include 4-body terms − and also train on larger water clusters . MLPs fit to periodic electronic structure offer the opportunity to readily capture many-body electronic structure effects, since these are naturally included in the electronic structure calculation.…”

Hydrogen Bonding and Infrared Spectra of Ethyl-3-methylimidazolium Bis(trifluoromethylsulfonyl)imide/Water Mixtures: A View from Molecular Dynamics Simulations

“…On the other hand, the application of DFTB methodology for these purposes would need an improvement of parameterization, tailored to a particular system. The other possibility, which we intend to exploit in the future, is the use of machine learning to obtain potential energy surfaces for MD simulations, as done recently, e.g., for bulk water …”

“…The other possibility, which we intend to exploit in the future, is the use of machine learning to obtain potential energy surfaces for MD simulations, as done recently, e.g., for bulk water. 68 …”

Recent advances in quantum fragmentation approaches to complex molecular and condensed‐phase systems

“…Recently, several QM quality force fields of proteins have been built via the fragmentation scheme 176,177 . We have also combined our fragmentation QM method with deep learning methods to build MLPs to boost the computational efficiency, bypassing the on‐the‐fly fragment‐based QM calculations 71,74 …”

“…We have reviewed the principles of the EE‐GMFCC method and its earlier applications in 2014 3 and 2020, 4 respectively. Since then, we have further made significant development to the EE‐GMFCC method and greatly extended its applications to more complex systems 64–74 . For example, quantitative descriptions of the excited‐state properties of fluorescent proteins and RNAs have been achieved within the framework of EE‐GMFCC 64,67 .…”