Preprint
Aug 2026
MolEmb: Multimodal Large Language Models Can Be Strong Molecular Embedding Models
This work introduces \textbf{MolEmb}, a lightweight framework that adapts MLLMs by aligning molecular profiles with textual descriptions in a shared embedding space using a bidirectional contrastive objective, and finds that context-aware molecular embedding is primarily a data property of the supervision.
Xinjian Zhao, Xiangru Jian, Yaoyao Xu et al.
· 0 citations