arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2102.01011cs.NEcs.AI
收藏

分子设计的深度进化学习

Deep Evolutionary Learning for Molecular Design

Brock University(布鲁克大学);Digital Technologies Research Centre(数字技术研究中心);National Research Council Canada(加拿大国家研究委员会)

Yifeng Li, Hsu Kiang Ooi, Alain Tchagang

摘要:

在本文中,我们提出了一个深度进化学习(DEL)过程,集成了基于片段的深度生成模型和多目标演化计算的分子设计。我们的方法使(1)生成模型的潜在空间中的进化操作,而不是结构空间,为下一代进化生成新的有前途的分子结构,以及(2)使用新生成的高质量样本生成模型微调。因此,DEL实现了一个数据模型协同进化的概念,提高了样本种群和生成模型学习。在两个公开数据集上的实验表明,DEL算法得到的样本总体具有更好的性质分布,优于多目标贝叶斯优化算法得到的样本.

英文摘要:

In this paper, we propose a deep evolutionary learning (DEL) process that integrates fragment-based deep generative model and multi-objective evolutionary computation for molecular design. Our approach enables (1) evolutionary operations in the latent space of the generative model, rather than the structural space, to generate novel promising molecular structures for the next evolutionary generation, and (2) generative model fine-tuning using newly generated high-quality samples. Thus, DEL implements a data-model co-evolution concept which improves both sample population and generative model learning. Experiments on two public datasets indicate that sample population obtained by DEL exhibits improved property distributions, and dominates samples generated by multi-objective Bayesian optimization algorithms.

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。

↑