Inferring Transition-Dipoles from Two-Dimensional Infrared Spectroscopy with Deep Learning
Abstract
Two-dimensional infrared (2D IR) spectroscopy is a powerful nonlinear technique for probing vibrational couplings, anharmonicities, coherence, and molecular dynamics. Among the microscopic parameters encoded in 2D IR spectra, transition dipoles are fundamental quantities that characterize molecular structure and govern light-matter interactions. However, reconstructing these quantities from spectroscopic data remains a challenging inverse problem. Here, we introduce a physics-informed convolutional neural network (CNN), Spec2DipoleNet, trained on 2D IR spectra generated via the non-Hermitian Hamiltonian (NHH) method. This multi-task CNN reconstructs the transition dipole matrices under physically-motivated masking constraints. On the validation set, the model achieves a root-mean-square error (RMSE) of 0.085, an R2 of 0.97, and a normalized cross-correlation of 0.98. It reliably recovers matrix elements associated with allowed transitions, while larger deviations are observed for nominally forbidden transitions. Our results suggest that transition dipole operators can be quantitatively inferred from 2D IR spectra, highlighting the potential of physics-informed deep learning for extracting microscopic quantum parameters from multidimensional spectroscopic data.
Related
- Machine-learning-enhanced Optical Photothermal infrared (O-PTIR) spectroscopy for rapid simultaneous identification of multiple metabolites to monitor biotransformations: proof of concept on Brewer's spent grain (BSG)
- Ultraviolet to Infrared Spectroscopy of the Type Ibn SN 2023tsz Suggests a Lower-mass Progenitor
- A dipole-forbidden conduction-band transition and the mid-infrared absorption of n-type Mg2Si, Mg2Ge and Mg2Sn: all-electron quasiparticle self-consistent GW calculations
- Ultra-broadband quantum infrared spectroscopy extending to 20 {\textmu}m
- Undetected Photon Spectroscopy in the Long-wave and Far Infrared
BibTeX
@misc{zhang2026inferring,
title = {{Inferring Transition-Dipoles from Two-Dimensional Infrared Spectroscopy with Deep Learning}},
author = {Hao-Yue Zhang and Rong-Hang Chen and Qing Ai},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.19972},
eprint = {2610.19972},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.19972},
}
RIS
TY - UNPB
TI - Inferring Transition-Dipoles from Two-Dimensional Infrared Spectroscopy with Deep Learning
AU - Hao-Yue Zhang
AU - Rong-Hang Chen
AU - Qing Ai
PY - 2026
DA - 2026/10/14
PB - arXiv
DO - 10.48550/arxiv.2610.19972
UR - https://arxiv.org/abs/2610.19972
AB - Two-dimensional infrared (2D IR) spectroscopy is a powerful nonlinear technique for probing vibrational couplings, anharmonicities, coherence, and molecular dynamics. Among the microscopic parameters encoded in 2D IR spectra, transition dipoles are fundamental quantities that characterize molecular structure and govern light-matter interactions. However, reconstructing these quantities from spectroscopic data remains a challenging inverse problem. Here, we introduce a physics-informed convolutional neural network (CNN), Spec2DipoleNet, trained on 2D IR spectra generated via the non-Hermitian Hamiltonian (NHH) method. This multi-task CNN reconstructs the transition dipole matrices under physically-motivated masking constraints. On the validation set, the model achieves a root-mean-square error (RMSE) of 0.085, an R^2 of 0.97, and a normalized cross-correlation of 0.98. It reliably recovers matrix elements associated with allowed transitions, while larger deviations are observed for nominally forbidden transitions. Our results suggest that transition dipole operators can be quantitatively inferred from 2D IR spectra, highlighting the potential of physics-informed deep learning for extracting microscopic quantum parameters from multidimensional spectroscopic data.
ER -