Graph Neural Network Prediction of Infrared Spectra of Interstellar Polycyclic Aromatic Hydrocarbons
Guoqing Tang, Jiang He, Zhao Wang, Dong Qiu · 2026
Read Graph Neural Network Prediction of Infrared Spectra of Interstellar Polycyclic Aromatic Hydrocarbons by Guoqing Tang, Jiang He, Zhao Wang, Dong Qiu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-12560.txt.
Description
Polycyclic aromatic hydrocarbons are the main contributors to aromatic infrared bands seen widely in space, yet analysing those bands is hard because the PAH family spans immense structural diversity. Guoqing Tang led this work with three co-authors, building a graph neural network frame that predicts PAH absorption spectra up to 10,000 times faster than traditional quantum chemical methods. Reliable reference spectra are otherwise difficult to compute for so many structures.
Four designs were evaluated: graph convolutional networks, graph attention networks, message passing networks, and attentive fingerprint models. The attentive fingerprint model performs best overall, and training it with five spectral distance metrics as losses shows Jensen-Shannon divergence gives the most accurate, stable results. Performance peaks for PAHs with 20 to 40 carbon atoms, falling for larger molecules where training data are scarce, offering fast approximate reference spectra for small to medium PAHs.
Catalog record
| CREATOR | Guoqing Tang, Jiang He, Zhao Wang, Dong Qiu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-12560