Post-Hoc Robustness Enhancement in Graph Neural Networks with Conditional Random Fields

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Abbahaddou, Yassine, Ennadir, Sofiane, Lutzeyer, Johannes F., Malliaros, Fragkiskos D., Vazirgiannis, Michalis
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917831576125440
author Abbahaddou, Yassine
Ennadir, Sofiane
Lutzeyer, Johannes F.
Malliaros, Fragkiskos D.
Vazirgiannis, Michalis
author_facet Abbahaddou, Yassine
Ennadir, Sofiane
Lutzeyer, Johannes F.
Malliaros, Fragkiskos D.
Vazirgiannis, Michalis
contents Graph Neural Networks (GNNs), which are nowadays the benchmark approach in graph representation learning, have been shown to be vulnerable to adversarial attacks, raising concerns about their real-world applicability. While existing defense techniques primarily concentrate on the training phase of GNNs, involving adjustments to message passing architectures or pre-processing methods, there is a noticeable gap in methods focusing on increasing robustness during inference. In this context, this study introduces RobustCRF, a post-hoc approach aiming to enhance the robustness of GNNs at the inference stage. Our proposed method, founded on statistical relational learning using a Conditional Random Field, is model-agnostic and does not require prior knowledge about the underlying model architecture. We validate the efficacy of this approach across various models, leveraging benchmark node classification datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2411_05399
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Post-Hoc Robustness Enhancement in Graph Neural Networks with Conditional Random Fields
Abbahaddou, Yassine
Ennadir, Sofiane
Lutzeyer, Johannes F.
Malliaros, Fragkiskos D.
Vazirgiannis, Michalis
Machine Learning
Social and Information Networks
Applications
Graph Neural Networks (GNNs), which are nowadays the benchmark approach in graph representation learning, have been shown to be vulnerable to adversarial attacks, raising concerns about their real-world applicability. While existing defense techniques primarily concentrate on the training phase of GNNs, involving adjustments to message passing architectures or pre-processing methods, there is a noticeable gap in methods focusing on increasing robustness during inference. In this context, this study introduces RobustCRF, a post-hoc approach aiming to enhance the robustness of GNNs at the inference stage. Our proposed method, founded on statistical relational learning using a Conditional Random Field, is model-agnostic and does not require prior knowledge about the underlying model architecture. We validate the efficacy of this approach across various models, leveraging benchmark node classification datasets.
title Post-Hoc Robustness Enhancement in Graph Neural Networks with Conditional Random Fields
topic Machine Learning
Social and Information Networks
Applications
url https://arxiv.org/abs/2411.05399