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.

Predicting Infrared Spectra with Message Passing Neural Networks

Machine Learning Spectroscopy Using a 2-Stage, Generalized Constituent Contribution Protocol

Prediction of the Infrared Absorbance Intensities and Frequencies of Hydrocarbons: A Message Passing Neural Network Approach

Prediction of the Infrared Absorbance Intensities and Frequencies of Hydrocarbons:A Message Passing Neural Network Approach

EnviroDetaNet: Pretrained E(3)-equivariant Message-Passing Neural Networks with Multi-Level Molecular Representations for Organic Molecule Spectra Prediction

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

EnviroDetaNet: Pretrained E (3)-equivariant Message-Passing Neural Networks with Multi-Level Molecular Representations for Organic Molecule Spectra Prediction

Neural Message Passing for NMR Chemical Shift Prediction

Infrared Spectral Analysis for Prediction of Functional Groups Based on Feature-Aggregated Deep Learning

Machine Learning Molecular Dynamics for the Simulation of Infrared Spectra

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

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

Neural Message Passing for Quantum Chemistry

Machine learning prediction of organic moieties from the IR spectra, enhanced by additionally using the derivative IR data

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

Graphormer-IR: Graph Transformers Predict Experimental IR Spectra Using Highly Specialized Attention

A Machine Learning Protocol for Predicting Protein Infrared Spectra.

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

Exploring Spectrum-based Molecular Descriptors for Reaction Performance Prediction.

Infrared Spectra at Coupled Cluster Accuracy from Neural Network Representations

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