Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2601.22610 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917233449500672 |
|---|---|
| author | Liu, Xiayu Lu, Zhengyi Liao, Yunhong Fan, Chan Li, Hou-biao |
| author_facet | Liu, Xiayu Lu, Zhengyi Liao, Yunhong Fan, Chan Li, Hou-biao |
| contents | Accurate molecular property prediction requires integrating complementary information from molecular structure and chemical semantics. In this work, we propose LGM-CL, a local-global multimodal contrastive learning framework that jointly models molecular graphs and textual representations derived from SMILES and chemistry-aware augmented texts. Local functional group information and global molecular topology are captured using AttentiveFP and Graph Transformer encoders, respectively, and aligned through self-supervised contrastive learning. In addition, chemically enriched textual descriptions are contrasted with original SMILES to incorporate physicochemical semantics in a task-agnostic manner. During fine-tuning, molecular fingerprints are further integrated via Dual Cross-attention multimodal fusion. Extensive experiments on MoleculeNet benchmarks demonstrate that LGM-CL achieves consistent and competitive performance across both classification and regression tasks, validating the effectiveness of unified local-global and multimodal representation learning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_22610 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Local-Global Multimodal Contrastive Learning for Molecular Property Prediction Liu, Xiayu Lu, Zhengyi Liao, Yunhong Fan, Chan Li, Hou-biao Machine Learning Artificial Intelligence Accurate molecular property prediction requires integrating complementary information from molecular structure and chemical semantics. In this work, we propose LGM-CL, a local-global multimodal contrastive learning framework that jointly models molecular graphs and textual representations derived from SMILES and chemistry-aware augmented texts. Local functional group information and global molecular topology are captured using AttentiveFP and Graph Transformer encoders, respectively, and aligned through self-supervised contrastive learning. In addition, chemically enriched textual descriptions are contrasted with original SMILES to incorporate physicochemical semantics in a task-agnostic manner. During fine-tuning, molecular fingerprints are further integrated via Dual Cross-attention multimodal fusion. Extensive experiments on MoleculeNet benchmarks demonstrate that LGM-CL achieves consistent and competitive performance across both classification and regression tasks, validating the effectiveness of unified local-global and multimodal representation learning. |
| title | Local-Global Multimodal Contrastive Learning for Molecular Property Prediction |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2601.22610 |