arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2510.07035cs.LGcs.AI
收藏

统一分子预训练:灵活2D与3D模态的单模态及配对模态整合

Unified Molecule Pre-training with Flexible 2D and 3D Modalities: Single and Paired Modality Integration

Computational Bioscience Research Center, King Abdullah University of Science and Technology(国王阿卜杜勒·阿齐兹科技大学计算生物科学研究中心);Institute for Infocomm Research, A*STAR(信息与通信研究机构,A*STAR);School of Computing and Information Systems, Singapore Management University(新加坡管理大学计算机与信息系统学院)

Tengwei Song, Min Wu, Yuan Fang

AI总结:

针对现有分子预训练需配对2D/3D数据的局限,提出FlexMol框架,通过分离模型、参数共享和缺失模态特征生成,支持单模态输入,实现统一表示学习,并在分子性质预测任务上取得优越性能。

摘要:

分子表示学习在推动药物发现和材料设计等应用中发挥着关键作用。现有工作利用分子信息的2D和3D模态进行预训练,旨在捕获全面的结构和几何洞察。然而,这些方法需要配对的2D和3D分子数据来有效训练模型并防止其坍缩为单一模态,这在某些模态不可用或生成成本高昂的场景中造成了限制。为克服这一局限,我们提出了FlexMol,一个灵活的分子预训练框架,它学习统一的分子表示,同时支持单模态输入。具体而言,受视觉-语言模型中统一结构的启发,我们的方法对2D和3D分子数据采用分离的模型,利用参数共享来提高计算效率,并使用解码器为缺失模态生成特征。这实现了多阶段连续学习过程,在训练期间两种模态协同贡献,同时确保在推理时仅有一种模态可用时的鲁棒性。大量实验表明,FlexMol在广泛的分子性质预测任务中取得了优越性能,我们还通过实验证明了其在不完整数据上的有效性。我们的代码和数据可在https://github.com/tewiSong/FlexMol获取。

英文摘要:

Molecular representation learning plays a crucial role in advancing applications such as drug discovery and material design. Existing work leverages 2D and 3D modalities of molecular information for pre-training, aiming to capture comprehensive structural and geometric insights. However, these methods require paired 2D and 3D molecular data to train the model effectively and prevent it from collapsing into a single modality, posing limitations in scenarios where a certain modality is unavailable or computationally expensive to generate. To overcome this limitation, we propose FlexMol, a flexible molecule pre-training framework that learns unified molecular representations while supporting single-modality input. Specifically, inspired by the unified structure in vision-language models, our approach employs separate models for 2D and 3D molecular data, leverages parameter sharing to improve computational efficiency, and utilizes a decoder to generate features for the missing modality. This enables a multistage continuous learning process where both modalities contribute collaboratively during training, while ensuring robustness when only one modality is available during inference. Extensive experiments demonstrate that FlexMol achieves superior performance across a wide range of molecular property prediction tasks, and we also empirically demonstrate its effectiveness with incomplete data. Our code and data are available at https://github.com/tewiSong/FlexMol.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑