[论文解读] Deeply learning molecular structure-property relationships using attention- and gate-augmented graph convolutional network

Seongok Ryu, Jaechang Lim+2 more|arXiv (Cornell University)|May 28, 2018

Machine Learning in Materials Science参考文献 53被引用 80

一句话总结

该论文通过引入注意力和门控跳跃连接来增强用于分子性质预测的图卷积网络，提高特征提取和潜在空间的组织。

ABSTRACT

Molecular structure-property relationships are key to molecular engineering for materials and drug discovery. The rise of deep learning offers a new viable solution to elucidate the structure-property relationships directly from chemical data. Here we show that the performance of graph convolutional networks (GCNs) for the prediction of molecular properties can be improved by incorporating attention and gate mechanisms. The attention mechanism enables a GCN to identify atoms in different environments. The gated skip-connection further improves the GCN by updating feature maps at an appropriate rate. We demonstrate that the resulting attention- and gate-augmented GCN could extract better structural features related to a target molecular property such as solubility, polarity, synthetic accessibility and photovoltaic efficiency compared to the vanilla GCN. More interestingly, it identified two distinct parts of molecules as essential structural features for high photovoltaic efficiency, and each of them coincided with the areas of donor and acceptor orbitals for charge-transfer excitations, respectively. As a result, the new model could accurately predict molecular properties and place molecules with similar properties close to each other in a well-trained latent space, which is critical for successful molecular engineering.

研究动机与目标

• 利用 GCNs 从结构出发激发对分子性质的更好预测。
• 结合注意力以识别处于不同环境中的原子。
• 引入门控跳跃连接以在适当的速率更新特征图。
• 证明该模型比 vanilla GCNs 更有效地捕捉结构-性质关系。

提出的方法

• 在图卷积网络中扩展注意力机制，以突出显示原子环境。
• 添加门控增强的跳跃连接以控制特征图更新。
• 在分子数据上训练以预测诸如溶解度、极性、合成可及性和光伏效率等性质。
• 分析潜在空间以显示具有相似性质的分子聚类。
• 识别与高光伏效率相关的结构子部件，并将其与给体/受体轨道区域联系起来。

实验结果

研究问题

• RQ1 GCNs 中的注意力是否有助于识别与目标分子性质相关的原子环境？
• RQ2 与 vanilla GCNs 相比，门控增强的跳跃连接是否提升学习动力学和预测性能？
• RQ3 该模型是否提取出与溶解度、极性、合成可及性和光伏效率等性质相关的有意义的结构特征？
• RQ4 学习到的表示是否能将具有相似性质的分子在潜在空间中放置得彼此接近？
• RQ5 是否存在对应功能区域（例如给体/受体区域）的可解释子结构，驱动特定性质？

主要发现

• 注意力和门控增强的 GCNs 比 vanilla GCNs 更好地预测目标分子性质。
• 该模型提取的结构特征与溶解度、极性、合成可及性和光伏效率更相关。
• 它识别出对高光伏效率至关重要的两个不同分子部分，与给体和受体轨道区域对齐。