A Large-Scale Dataset for Molecular Structure-Language Description via a Rule-Regularized Method

TL;DR

This paper introduces an automated framework for creating a large-scale dataset of molecular structure descriptions aligned with natural language, facilitating chemical reasoning with language models.

Contribution

It presents a rule-based annotation method that generates high-quality, structure-preserving molecular descriptions at scale, enabling better molecule-language alignment.

Findings

Achieved 98.6% description precision on validation subset.

Curated a dataset of approximately 163,000 molecule-description pairs.

Demonstrated the framework's utility for chemical tasks relying on structural descriptions.

Abstract

Molecular function is largely determined by structure. Accurately aligning molecular structure with natural language is therefore essential for enabling large language models (LLMs) to reason about downstream chemical tasks. However, the substantial cost of human annotation makes it infeasible to construct large-scale, high-quality datasets of structure-grounded descriptions. In this work, we propose a fully automated annotation framework for generating precise molecular descriptions that preserve complete structural details at scale. Our approach builds upon and extends a rule-based chemical nomenclature parser to interpret IUPAC names and construct enriched, structural XML metadata that explicitly encodes molecular structure. This metadata is then used to guide LLMs in producing accurate natural-language descriptions. Using this framework, we curate a large-scale dataset of approximately $163$k molecule--description pairs. A rigorous validation protocol combining LLM-based and expert human evaluation on a subset of $2,000$ molecules demonstrates a high description precision of $98.6$%. The proposed annotation framework is readily beneficial to broader chemical tasks that rely on structural descriptions, with the resulting dataset providing a reliable foundation for molecule--language alignment. The source code and dataset are hosted at https://github.com/TheLuoFengLab/MolLangData and https://huggingface.co/datasets/ChemFM/MolLangData, respectively.