RTMol: Rethinking Molecule-text Alignment in a Round-trip View

TL;DR

RTMol introduces a bidirectional, self-supervised framework for aligning molecular sequences with text, improving consistency and accuracy in molecule-text tasks without relying on paired datasets.

Contribution

The paper presents RTMol, a novel round-trip learning approach that unifies molecular captioning and text-to-molecule generation, addressing limitations of existing methods.

Findings

Improves bidirectional alignment performance by up to 47%.

Enables unsupervised training for molecular captioning.

Introduces novel round-trip evaluation metrics.

Abstract

Aligning molecular sequence representations (e.g., SMILES notations) with textual descriptions is critical for applications spanning drug discovery, materials design, and automated chemical literature analysis. Existing methodologies typically treat molecular captioning (molecule-to-text) and text-based molecular design (text-to-molecule) as separate tasks, relying on supervised fine-tuning or contrastive learning pipelines. These approaches face three key limitations: (i) conventional metrics like BLEU prioritize linguistic fluency over chemical accuracy, (ii) training datasets frequently contain chemically ambiguous narratives with incomplete specifications, and (iii) independent optimization of generation directions leads to bidirectional inconsistency. To address these issues, we propose RTMol, a bidirectional alignment framework that unifies molecular captioning and text-to-SMILES generation through self-supervised round-trip learning. The framework introduces novel round-trip evaluation metrics and enables unsupervised training for molecular captioning without requiring paired molecule-text corpora. Experiments demonstrate that RTMol enhances bidirectional alignment performance by up to 47% across various LLMs, establishing an effective paradigm for joint molecule-text understanding and generation.