Data Fusion and Contrastive Alignment for Unconstrained IR Molecular Structure Elucidation

2026-07-28

AI Summary

This work addresses the limitation of existing infrared (IR) spectroscopy–based molecular structure elucidation methods, which rely on predefined molecular formulas and struggle to predict complete structures under unconstrained conditions. To overcome this, the authors propose an enhanced encoder–decoder Transformer architecture featuring a non-additive aggregation mechanism grounded in linear order statistics and the Choquet integral. The model further incorporates a Mixture-of-Experts decoder and a contrastive alignment loss to improve its capacity to infer arbitrary molecular structures directly from raw IR spectra. Experimental results demonstrate that the proposed method achieves a Top-K prediction accuracy more than 10 percentage points higher than IR-only baselines, confirming that IR spectra inherently encode rich and discriminative chemical information sufficient for de novo molecular reconstruction.

Abstract

Automated molecular structure elucidation from infrared (IR) spectroscopy data has seen significant advancements in recent years, but its broad applicability is limited by a reliance on pre-determined chemical formulas provided as auxiliary model inputs. This limits model predictions to isomer identification rather than full molecular structure prediction. Although transformer models have been shown to identify molecular isomers with high accuracy, their reliability for unconstrained structure elucidation is comparatively low and poorly understood. In this work, we propose and evaluate key modifications to the traditional encoder-decoder transformer. To better address the vast chemical space of the unconstrained problem, we implement a novel Mixture-of-Experts (MoE) decoder module that utilizes non-additive aggregation via linear-order statistics and the Choquet integral. We further modify the transformer to utilize these non-additive operators when aggregating spectral representations as well. Together with an auxiliary contrastive alignment loss term, these enhancements improve Top-K prediction accuracy by over 10 percentage points compared to baseline IR-only models. Through sub-structure fragment analysis of molecular predictions, we further confirm that infrared spectra encode the vast majority of relevant chemical information, implying that the higher performance of isomer-ranking models is largely due to underrepresented or overlapping absorption bands for molecules in the explored chemical space. Ultimately, by demonstrating the efficacy of automated molecular structure elucidation from measured IR spectra, this work serves to significantly broaden the utility of AI in analytical chemistry.

Problem

• molecular structure elucidation
• infrared spectroscopy
• unconstrained prediction
• isomer identification
• chemical space

Innovation

• Mixture-of-Experts
• non-additive aggregation
• Choquet integral
• contrastive alignment
• unconstrained structure elucidation