NMIRacle: Multi-modal Generative Molecular Elucidation from IR and NMR Spectra

2025-12-17

AI Summary

This study addresses the end-to-end generation of molecular structures from multimodal infrared (IR) and dual-nucleus (¹H/¹³C) NMR spectra—a long-standing challenge in analytical chemistry. We propose a two-stage conditional generative framework: first, a count-aware molecular fragment encoder enables accurate structural reconstruction; second, a joint spectral encoder maps IR and both NMR modalities into a unified conditional embedding to guide generation. Our approach introduces, for the first time, count-aware fragment representations and a multimodal spectral joint conditioning mechanism, enabling fully automated structure elucidation without expert-defined rules or candidate structure libraries. Built upon a synergistic architecture integrating deep generative models, conditional variational autoencoders, and pretrained generators, our method achieves a 12.6% absolute improvement in top-1 accuracy on standard elucidation benchmarks and demonstrates strong robustness for highly complex molecules.

Technology Category

• Machine Learning: Multimodal Learning
• Natural Language Processing: Generation
• Computer Vision: Multi-modal Vision

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Molecular structure elucidation from spectroscopic data is a long-standing challenge in Chemistry, traditionally requiring expert interpretation. We introduce NMIRacle, a two-stage generative framework that builds upon recent paradigms in AI-driven spectroscopy with minimal assumptions. In the first stage, NMIRacle learns to reconstruct molecular structures from count-aware fragment encodings, which capture both fragment identities and their occurrences. In the second stage, a spectral encoder maps input spectroscopic measurements (IR, 1H-NMR, 13C-NMR) into a latent embedding that conditions the pre-trained generator. This formulation bridges fragment-level chemical modeling with spectral evidence, yielding accurate molecular predictions. Empirical results show that NMIRacle outperforms existing baselines on molecular elucidation, while maintaining robust performance across increasing levels of molecular complexity.

Problem

• Automates molecular structure elucidation from spectroscopic data
• Generates molecular structures using IR and NMR spectral inputs
• Improves accuracy over existing methods for complex molecules

Innovation

• Two-stage generative framework for molecular structure elucidation
• Fragment encodings capture identities and occurrences for reconstruction
• Spectral encoder conditions generator with IR and NMR embeddings