Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

MolFM: A Multimodal Molecular Foundation Model

About

Molecular knowledge resides within three different modalities of information sources: molecular structures, biomedical documents, and knowledge bases. Effective incorporation of molecular knowledge from these modalities holds paramount significance in facilitating biomedical research. However, existing multimodal molecular foundation models exhibit limitations in capturing intricate connections between molecular structures and texts, and more importantly, none of them attempt to leverage a wealth of molecular expertise derived from knowledge graphs. In this study, we introduce MolFM, a multimodal molecular foundation model designed to facilitate joint representation learning from molecular structures, biomedical texts, and knowledge graphs. We propose cross-modal attention between atoms of molecular structures, neighbors of molecule entities and semantically related texts to facilitate cross-modal comprehension. We provide theoretical analysis that our cross-modal pre-training captures local and global molecular knowledge by minimizing the distance in the feature space between different modalities of the same molecule, as well as molecules sharing similar structures or functions. MolFM achieves state-of-the-art performance on various downstream tasks. On cross-modal retrieval, MolFM outperforms existing models with 12.13% and 5.04% absolute gains under the zero-shot and fine-tuning settings, respectively. Furthermore, qualitative analysis showcases MolFM's implicit ability to provide grounding from molecular substructures and knowledge graphs. Code and models are available on https://github.com/BioFM/OpenBioMed.

Yizhen Luo, Kai Yang, Massimo Hong, Xing Yi Liu, Zaiqing Nie• 2023

Related benchmarks

TaskDatasetResultRank
Molecular property predictionMoleculeNet BBBP (scaffold)
ROC AUC72.9
117
Molecular property predictionMoleculeNet HIV (scaffold)
ROC AUC78.8
66
Molecular property predictionBACE (test)
ROC-AUC83.9
65
Molecule Description GenerationChEBI-20 (test)
BLEU-258.5
34
ClassificationMoleculeNet BBBP (test)
ROC AUC0.729
30
Molecular Property ClassificationBACE (MoleculeNet) scaffold (test)
ROC-AUC0.839
30
Description-guided molecule designChEBI-20 2022 (test)
Exact Match Accuracy21
26
molecule property predictionHIV MoleculeNet (test)
AUROC78.8
24
Molecule Description GenerationChEBI-20 2022 (test)
BLEU-20.585
20
molecule property predictionClintox MoleculeNet (test)
AUROC0.797
15
Showing 10 of 12 rows

Other info

Code

Follow for update