Abstract:Kernel models of potential energy surfaces (PES) for polyatomic molecules are often restricted by a specific choice of the kernel function. This can be avoided by optimizing the complexity of the kernel function. For regression problems with very expensive data, the functional form of the model kernels can be optimized in the Gaussian process (GP) setting through compositional function search guided by the Bayesian information criterion. However, the compositional kernel search is computationally demanding and relies on greedy strategies, which may yield sub-optimal kernels. An alternative strategy of increasing complexity of GP kernels treats a GP as a Bayesian neural network (NN) with a variable number of hidden layers, which yields NNGP models. Here, we present a direct comparison of GP models with composite kernels and NNGP models for applications aiming at the construction of global PES for polyatomic molecules. We show that NNGP models of PES can be trained much more efficiently and yield better generalization accuracy without relying on any specific form of the kernel function. We illustrate that NNGP models trained by distributions of energy points at low energies produce accurate predictions of PES at high energies. We also illustrate that NNGP models can extrapolate in the input variable space by building the free energy surface of the Heisenberg model trained in the paramagnetic phase and validated in the ferromagnetic phase. By construction, composite kernels yield more accurate models than kernels with a fixed functional form. Therefore, by illustrating that NNGP models outperform GP models with composite kernels, our work suggests that NNGP models should be a preferred choice of kernel models for PES.

What problem does this paper attempt to address?

The paper attempts to address the problem of how to improve the accuracy and generalization ability of models in the modeling of potential energy surfaces (PES) for polyatomic molecules with limited data. Specifically, the paper compares two methods: Gaussian Process (GP) models based on composite kernel functions and Neural Network Gaussian Process (NNGP) models. 1. **Background**: - For quantum dynamics simulations of polyatomic molecules, accurate global potential energy surface models are required, which are usually obtained by fitting the results of electronic structure calculations. - Machine learning methods, especially artificial neural networks (NN) and kernel models (such as Kernel Ridge Regression KRR and Gaussian Process Regression GP), have been widely used to construct these potential energy surface models. - Methods to improve model accuracy include increasing the amount of training data and optimizing model complexity. However, given the extremely time-consuming nature of potential energy calculations, the goal is to optimize the model on a given small dataset to achieve high accuracy. 2. **Core Problem**: - Although GP models based on composite kernel functions can improve accuracy by optimizing the form of the kernel function, this method is computationally expensive and relies on greedy strategies, which may lead to suboptimal solutions. - The Neural Network Gaussian Process (NNGP) model, as an alternative strategy, treats GP as a Bayesian neural network with a variable number of hidden layers, allowing for more efficient training and better generalization accuracy without relying on specific kernel function forms. 3. **Research Objectives**: - Directly compare the performance of GP models based on composite kernel functions and NNGP models in constructing global potential energy surfaces for polyatomic molecules. - Evaluate whether the NNGP model can achieve more efficient training and higher generalization accuracy on small datasets, especially in extrapolation tasks. 4. **Main Findings**: - The NNGP model outperforms the GP model based on composite kernel functions in terms of training efficiency and generalization accuracy. - The NNGP model excels in extrapolation tasks, such as accurately predicting the potential energy surface in high-energy regions from models trained on low-energy point distributions. - The NNGP model demonstrates stronger learning ability and better data efficiency when dealing with high-dimensional potential energy surfaces. In summary, this paper aims to explore how to more effectively construct global potential energy surface models for polyatomic molecules under limited data conditions by comparing NNGP models and traditional GP models, and it demonstrates the significant advantages of the NNGP model in this regard.

Neural network Gaussian processes as efficient models of potential energy surfaces for polyatomic molecules

Interpolation and extrapolation of global potential energy surfaces for polyatomic systems by Gaussian processes with composite kernels

Gaussian Processes with Spectral Delta kernel for higher accurate Potential Energy surfaces for large molecules

Efficient non-parametric fitting of potential energy surfaces for polyatomic molecules with Gaussian processes

Neural networks vs Gaussian process regression for representing potential energy surfaces: A comparative study of fit quality and vibrational spectrum accuracy.

Gaussian Processes for Surrogate Modeling of Discharged Fuel Nuclide Compositions

Representing Globally Accurate Reactive Potential Energy Surfaces with Complex Topography by Combining Gaussian Process Regression and Neural Networks.

Quantum Gaussian process model of potential energy surface for a polyatomic molecule

Assessing Gaussian Process Regression and Permutationally Invariant Polynomial Approaches To Represent High-Dimensional Potential Energy Surfaces.

Gaussian Moments as Physically Inspired Molecular Descriptors for Accurate and Scalable Machine Learning Potentials

Ab-Initio Potential Energy Surfaces by Pairing GNNs with Neural Wave Functions

Gradient domain machine learning with composite kernels: improving the accuracy of PES and force fields for large molecules

Global Neural Network Potential with Explicit Many-Body Functions for Improved Descriptions of Complex Potential Energy Surface.

A Gaussian Process Based Δ-Machine Learning Approach to Reactive Potential Energy Surfaces.

The Bigger the Better? Accurate Molecular Potential Energy Surfaces from Minimalist Neural Networks

Building nonparametric $n$-body force fields using Gaussian process regression

Additive Kernels for Gaussian Process Modeling

Composite Gaussian Processes: Scalable Computation and Performance Analysis

Construction of Reactive Potential Energy Surfaces with Gaussian Process Regression: Active Data Selection

Physics Inspired Approaches To Understanding Gaussian Processes

Extrapolative Bayesian Optimization with Gaussian Process and Neural Network Ensemble Surrogate Models