Boosting LLM's Molecular Structure Elucidation with Knowledge Enhanced Tree Search Reasoning

TL;DR

This paper presents K-MSE, a knowledge-enhanced framework that improves large language models' ability to elucidate molecular structures from spectral data by integrating external chemical knowledge and specialized scoring, achieving over 20% performance gains.

Contribution

The work introduces a novel external knowledge base and a molecule-spectrum scorer to enhance LLM reasoning in molecular structure elucidation, addressing knowledge limitations and evaluation inaccuracies.

Findings

Over 20% performance improvement on GPT-4o-mini and GPT-4o.

Effective integration of external chemical knowledge base.

Improved accuracy in molecular structure predictions.

Abstract

Molecular structure elucidation involves deducing a molecule's structure from various types of spectral data, which is crucial in chemical experimental analysis. While large language models (LLMs) have shown remarkable proficiency in analyzing and reasoning through complex tasks, they still encounter substantial challenges in molecular structure elucidation. We identify that these challenges largely stem from LLMs' limited grasp of specialized chemical knowledge. In this work, we introduce a Knowledge-enhanced reasoning framework for Molecular Structure Elucidation (K-MSE), leveraging Monte Carlo Tree Search for test-time scaling as a plugin. Specifically, we construct an external molecular substructure knowledge base to extend the LLMs' coverage of the chemical structure space. Furthermore, we design a specialized molecule-spectrum scorer to act as a reward model for the reasoning process, addressing the issue of inaccurate solution evaluation in LLMs. Experimental results show that our approach significantly boosts performance, particularly gaining more than 20% improvement on both GPT-4o-mini and GPT-4o. Our code is available at https://github.com/HICAI-ZJU/K-MSE.