Physics > Chemical Physics
[Submitted on 11 Apr 2023 (this version), latest version 30 Apr 2023 (v2)]
Title:Neural network Gaussian processes as efficient models of potential energy surfaces for polyatomic molecules
View PDFAbstract:Gaussian processes (GP) have been previously shown to yield accurate models of potential energy surfaces (PES) for polyatomic molecules. The advantages of GP models include Bayesian uncertainty, which can be used for Bayesian optimization, and the possibility to optimize the functional form of the model kernels through compositional function search by optimizing marginal likelihood (or equivalently the Bayesian information criterion), which can be used to enhance the accuracy of models trained with a small number of training points. This is particularly important for applications, where electronic structure calculations are exceedingly expensive. However, the compositional kernel search is computationally demanding and relies on greedy algorithms, which may yield sub-optimal kernels. In the present work, we compare two strategies of increasing kernel complexity for building accurate models of PES: the compositional kernel search and kernels built by explicitly treating GP as a multi-layer Bayesian neural network (NN). We show that the resulting NNGP models can be trained at a fraction of the cost of training composite kernel models, while offering better generalization accuracy. 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.
Submission history
From: Jun Dai [view email][v1] Tue, 11 Apr 2023 22:47:53 UTC (1,329 KB)
[v2] Sun, 30 Apr 2023 01:07:29 UTC (1,194 KB)
Current browse context:
physics.chem-ph
Change to browse by:
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.