Multi-modal Molecule Structure-text Model for Text-based Retrieval and Editing

TL;DR

MoleculeSTM is a multi-modal model that jointly learns chemical structures and textual descriptions, enabling advanced text-based retrieval and editing in drug discovery with state-of-the-art generalization.

Contribution

The paper introduces MoleculeSTM, a novel multi-modal model trained on a large dataset, enabling zero-shot text-based molecule retrieval and editing.

Findings

Achieves state-of-the-art zero-shot performance on biochemical benchmarks.

Demonstrates effective structure-text retrieval and molecule editing.

Supports open vocabulary and compositionality in molecular understanding.

Abstract

There is increasing adoption of artificial intelligence in drug discovery. However, existing studies use machine learning to mainly utilize the chemical structures of molecules but ignore the vast textual knowledge available in chemistry. Incorporating textual knowledge enables us to realize new drug design objectives, adapt to text-based instructions and predict complex biological activities. Here we present a multi-modal molecule structure-text model, MoleculeSTM, by jointly learning molecules' chemical structures and textual descriptions via a contrastive learning strategy. To train MoleculeSTM, we construct a large multi-modal dataset, namely, PubChemSTM, with over 280,000 chemical structure-text pairs. To demonstrate the effectiveness and utility of MoleculeSTM, we design two challenging zero-shot tasks based on text instructions, including structure-text retrieval and molecule editing. MoleculeSTM has two main properties: open vocabulary and compositionality via natural language. In experiments, MoleculeSTM obtains the state-of-the-art generalization ability to novel biochemical concepts across various benchmarks.