Current browse context:
q-bio.QM
Change to browse by:
References & Citations
Quantitative Biology > Quantitative Methods
Title: Guiding Deep Molecular Optimization with Genetic Exploration
(Submitted on 4 Jul 2020 (v1), last revised 27 Oct 2020 (this version, v3))
Abstract: De novo molecular design attempts to search over the chemical space for molecules with the desired property. Recently, deep learning has gained considerable attention as a promising approach to solve the problem. In this paper, we propose genetic expert-guided learning (GEGL), a simple yet novel framework for training a deep neural network (DNN) to generate highly-rewarding molecules. Our main idea is to design a "genetic expert improvement" procedure, which generates high-quality targets for imitation learning of the DNN. Extensive experiments show that GEGL significantly improves over state-of-the-art methods. For example, GEGL manages to solve the penalized octanol-water partition coefficient optimization with a score of 31.40, while the best-known score in the literature is 27.22. Besides, for the GuacaMol benchmark with 20 tasks, our method achieves the highest score for 19 tasks, in comparison with state-of-the-art methods, and newly obtains the perfect score for three tasks.
Submission history
From: Sungsoo Ahn [view email][v1] Sat, 4 Jul 2020 05:01:26 GMT (4476kb,D)
[v2] Fri, 17 Jul 2020 11:23:07 GMT (4475kb,D)
[v3] Tue, 27 Oct 2020 10:49:47 GMT (10262kb,D)
Link back to: arXiv, form interface, contact.