Deep learning for molecular generation and optimization - a review of the state of the art

By Daniel C. Elton, Z. Boukouvalas, M. Fuge, Peter W. Chung. Molecular Systems Design & Engineering, 2019

355 citations recorded by the source corpus.

We review a recent groundswell of work which uses deep learning techniques to generate and optimize molecules.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design (2022, background): “Although many recent efforts have been put to review the literature of machine learning models for molecule design, most of them only focus on part of the literature in terms of generative methods and tasks [8], [9], [10], [11], [12], [13], [14].”
Machine Learning on Graphs: A Model and Comprehensive Taxonomy (2020, methodology): “A review of recent methods can be found in [36].”
Sample-Efficient Optimization in the Latent Space of Deep Generative Models via Weighted Retraining (2020, background): “Recently, machine learning has shown promising results in many problems that can be framed as optimization, such as conditional image [57, 40] and text [44] generation, molecular and materials design [11, 47], and neural architecture search [10].”
A Review of Deep Learning Methods for Antibodies (2020, methodology): “For example, deep learning applications in small molecule design benefit from commonly used benchmarking datasets such as ZINC and ChEMBL and image-based deep learning methods are often evaluated on MNIST or ImageNet [69,70].”
Inverse design of 3d molecular structures with conditional generative neural networks (2021, methodology): “These methods typically represent molecules as graphs or SMILES strings [30, 31], which lack information about the threedimensional structure of a molecule.”