# Predictive Quantum Vibrational Spectra through Active Learning 4G-NNPs

## TL;DR

This paper introduces a new method using advanced neural networks to accurately predict vibrational spectra of complex systems like water and its interface with air.

## Contribution

The novel contribution is the development of 4G-HDCNNPs integrated with quantum effects for accurate infrared spectral simulations without explicit dipole training.

## Key findings

- The framework accurately simulates infrared spectra for bulk water and air–water interfaces.
- Nonlocal charge transfer effects and nuclear quantum effects are seamlessly integrated without empirical parameters.
- The methodology is general and practical for predictive spectral simulations of complex systems.

## Abstract

Predictive simulation of vibrational spectra of complex
condensed-phase
and interface systems with thousands of degrees of freedom has long
been a challenging task of modern condensed matter theory. In this
work, fourth-generation high-dimensional committee neural network
potentials (4G-HDCNNPs) are developed using active learning and query-by-committee,
and introduced to the essential nuclear quantum effects (NQEs) as
well as conformational entropy and anharmonicities from path integral
(PI) molecular dynamics simulations. Using representative bulk water
and air–water interface test cases, we demonstrate the accuracy
of the developed framework in infrared spectral simulations. Specifically,
by seamlessly integrating nonlocal charge transfer effects from 4G-HDCNNPs
with the NQEs from PI methods, our introduced methodology yields accurate
infrared spectra using predicted charges from the 4G-HDCNNP architecture
without explicit training of dipole moments. The framework introduced
in this work is simple and general, offering a practical paradigm
for predictive spectral simulations of complex condensed phases and
interfaces, free from empirical parametrizations and ad hoc fitting.