How Robust Are Graph Neural Networks to Structural Noise?
- Georgia Institute of Technology, Atlanta, GA (United States); Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
Graph neural networks (GNNs) are an emerging model for learning graph embeddings and making predictions on graph structured data. However, robustness of graph neural networks is not yet well-understood. In this work, we focus on node structural identity predictions, where a representative GNN model is able to achieve near-perfect accuracy. We also show that the same GNN model is not robust to addition of structural noise, through a controlled dataset and set of experiments. Finally, we show that under the right conditions, graph-augmented training is capable of significantly improving robustness to structural noise.
- Research Organization:
- Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States); Georgia Institute of Technology, Atlanta, GA (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
- DOE Contract Number:
- AC04-94AL85000; NA0003525
- OSTI ID:
- 1592845
- Report Number(s):
- SAND--2020-0092R; 681942
- Country of Publication:
- United States
- Language:
- English
Similar Records
Track Seeding and Labelling with Embedded-space Graph Neural Networks
Enhancing Network Anomaly Detection Using Graph Neural Networks
Enhancing Network Anomaly Detection Using Graph Neural Networks
Conference
·
Tue Jun 30 00:00:00 EDT 2020
·
OSTI ID:1656641
Enhancing Network Anomaly Detection Using Graph Neural Networks
Conference
·
Tue Jun 11 00:00:00 EDT 2024
· 2024 22nd Mediterranean Communication and Computer Networking Conference (MedComNet)
·
OSTI ID:2426922
Enhancing Network Anomaly Detection Using Graph Neural Networks
Conference
·
Tue Jun 11 00:00:00 EDT 2024
·
OSTI ID:3003233