Graph Convolutional Networks (GCNs) for Molecular Property Prediction in Drug Development
Loading...
Date
2020-05
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
The Ohio State University
Abstract
Molecular property prediction is key to drug development. The rising of deep learning techniques provides new possibilities to learn the molecular properties directly from chemical data. In particular, graph convolutional networks have been introduced into the field and made significant enhancements compared to traditional methods. The first part of this paper serves as a study to explore and evaluate this emerging method while the second part demonstrates that graph convolution networks can be further improved by incorporating attention mechanism, another influential deep learning idea.
Description
Keywords
Deep Learning, Neural Networks, Molecular Property Prediction, Drug Development