Graphormer-IR: Graph Transformers Can Predict Experimental IR Spectra Using Highly Specialized Attention
Bibliographic record
Abstract
Infrared (IR) spectroscopy is crucial in various chemical and forensic domains, but faster in silico methods for predicting experimental spectra are needed due to the time and accuracy limitations of ab initio methods. We employ Graphormer, a graph neural network (GNN) transformer, to predict IR spectra using only Simplified Molecular-Input Line-Entry System (SMILES) strings. Our dataset includes 53,528 high-quality spectra with elements H, C, N, O, F, Si, S, P, Cl, Br, and I in five solvent phases. When using only atomic numbers for node encodings, Graphormer-IR achieved SIS_μ test scores of 0.8449±0.0012 (n=5), surpassing the state-of-the-art Chemprop-IR (SIS_μ = 0.8409 ± 0.0014, n=5), with only 36% of the encoded information. Augmenting node embeddings with additional node-level descriptors in learned embeddings generated through a multi-layer perceptron improves scores to SIS_μ = 0.8523±0.0006, a total improvement of 19.7σ. These improved scores show how Graphormer-IR excels in capturing long-range interactions like hydrogen bonding, anharmonic peak positions in experimental spectra, and stretching frequencies of uncommon functional groups. Scaling our architecture to 210 attention heads demonstrates specialist-like behavior for distinct IR frequencies that improves model performance. Our model utilizes novel architectures, including a global node for solvent phase encoding, learned node feature embeddings, and a 1D smoothing CNN. Graphormer-IR’s innovations underscore its potency over traditional message-passing neural networks (MPNNs) due to its expressive embeddings and ability to capture long-range intra-molecular relationships.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.010 | 0.004 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".