Molecular Property Prediction With Graph Neural Networks
Caitlin Whitter, Purdue University
Machine-learning techniques are increasingly used by the computational chemistry community to predict molecular properties. Compared to traditional methods, which can be computationally prohibitive, machine-learning models can predict these properties with high accuracy and in much less time. This poster will present a Graph Neural Network (GNN) approach for predicting molecular properties. A GNN is a neural network architecture for performing machine learning on graphs. As such, this framework is highly amenable to parallelism and hence has the potential for substantial scalability in both the size and number of molecules.
Abstract Author(s): Caitlin Whitter, Yu-Hang Tang, Aydin Buluc, Alex Pothen