Abstract:Machine learning has emerged as an invaluable tool in many research areas. In the present work, we harness this power to predict highly accurate molecular infrared spectra with unprecedented computational efficiency. To account for vibrational anharmonic and dynamical effects -- typically neglected by conventional quantum chemistry approaches -- we base our machine learning strategy on ab initio molecular dynamics simulations. While these simulations are usually extremely time consuming even for small molecules, we overcome these limitations by leveraging the power of a variety of machine learning techniques, not only accelerating simulations by several orders of magnitude, but also greatly extending the size of systems that can be treated. To this end, we develop a molecular dipole moment model based on environment dependent neural network charges and combine it with the neural network potentials of Behler and Parrinello. Contrary to the prevalent big data philosophy, we are able to obtain very accurate machine learning models for the prediction of infrared spectra based on only a few hundreds of electronic structure reference points. This is made possible through the introduction of a fully automated sampling scheme and the use of molecular forces during neural network potential training. We demonstrate the power of our machine learning approach by applying it to model the infrared spectra of a methanol molecule, n-alkanes containing up to 200 atoms and the protonated alanine tripeptide, which at the same time represents the first application of machine learning techniques to simulate the dynamics of a peptide. In all these case studies we find excellent agreement between the infrared spectra predicted via machine learning models and the respective theoretical and experimental spectra.

Infrared Spectra at Coupled Cluster Accuracy from Neural Network Representations

Neural Network Approach for Predicting Infrared Spectra from 3D Molecular Structure

Cluster-model DFT Simulations of the Infrared Spectra of Triazine-Based Molecular Crystals.

Machine Learning Molecular Dynamics for the Simulation of Infrared Spectra

Efficient Composite Infrared Spectroscopy: Combining the Doubly-Harmonic Approximation with Machine Learning Potentials

Infrared Spectra Prediction for Functional Group Region Utilizing a Machine Learning Approach with Structural Neighboring Mechanism

Predicting Infrared Spectra with Message Passing Neural Networks

Tensorial Properties via the Neuroevolution Potential Framework: Fast Simulation of Infrared and Raman Spectra

From the Automated Calculation of Potential Energy Surfaces to Accurate Infrared Spectra

Accurate and Affordable Simulation of Molecular Infrared Spectra with AIQM Models

Theoretical Study of the Infrared Spectrum of Clusters

Optical spectra in the condensed phase: Capturing anharmonic and vibronic features using dynamic and static approaches

Infrared Spectra Prediction for Diazo Groups Utilizing a Machine Learning Approach with Structural Attention Mechanism

Machine-learning prediction of infrared spectra of interstellar polycyclic aromatic hydrocarbons

Boosting the Modeling of InfraRed and Raman Spectra of Bulk Phase Chromophores with Machine Learning

Infrared spectra of neutral polycyclic aromatic hydrocarbons based on machine learning potential energy surface and dipole mapping

Infrared Vibrational Predissociation Spectroscopy of Small Molecular Clusters

A Machine-Learned "Chemical Intuition" to Overcome Spectroscopic Data Scarcity

Accelerating molecular vibrational spectra simulations with a physically informed deep learning model

Spectroscopic constants from atomic properties: a machine learning approach