Dissertations, Theses, and Capstone Projects

Date of Degree

9-2023

Document Type

Dissertation

Degree Name

Ph.D.

Program

Computer Science

Advisor

Lei Xie

Committee Members

Zhigang Zhu

Liang Zhao

Yuwei Yang

Subject Categories

Artificial Intelligence and Robotics | Computer Sciences | Structural Biology

Keywords

Deep Learning, Machine Learning, Graph Neural Networks, Drug Discovery

Abstract

Graph Neural Networks (GNNs) are widely recognized for their potential in learning from graph-structured data and solving complex problems. However, optimal performance and applicability of GNNs have been an open-ended challenge. This dissertation presents a series of substantial advances addressing this problem. First, we investigate attention-based GNNs, revealing a critical shortcoming: their ignorance of cardinality information that impacts their discriminative power. To rectify this, we propose Cardinality Preserved Attention (CPA) models that can be applied to any attention-based GNNs, which exhibit a marked improvement in performance. Next, we introduce the Directional Node Pair (DNP) descriptor and the Robust Molecular Graph Convolutional Network (RoM-GCN). These are robust tools designed to encode both chemical and geometric information, ensuring an accurate representation of 3D molecules. Finally, we develop Physics-Aware Multiplex Graph Neural Network (PAMNet), a universal, physics-informed framework that models 3D molecular systems with high accuracy and efficiency. This innovation is particularly effective in a variety of molecular tasks, outperforming existing baselines. Collectively, these advances underscore the need for continued exploration of bespoke optimization strategies to fully realize the potential of GNNs across different application domains.

Share

COinS