Probabilistic Graphical Model for Robust Graph Neural Networks against Noisy Labels

Qingqing Ge, Jianxiang Yu, Zeyuan Zhao, Xiang Li

Keywords: graph neural network, label noise, probabilistic graphical model, heterophilous graphs

TL;DR: We propose a novel probabilistic graphical model based framework PRGNN to solve the problem of label noise for GNNs.

Abstract: While robust graph neural networks (GNNs) have been widely studied for graph perturbation and attack, those for label noise have received significantly less attention. Most existing methods heavily rely on the label smoothness assumption to correct noisy labels, which adversely affects their performance on heterophilous graphs. Further, they generally perform poorly in high noise-rate scenarios. To address these problems, in this paper, we propose a novel probabilistic graphical model based framework PRGNN. Given a noisy label set and a clean label set, our goal is to maximize the likelihood of labels in the clean set. We first present PRGNN-v1, which generates clean labels based on graphs only in the Bayesian network. To further leverage the information of clean labels in the noisy label set, we put forward PRGNN-v2, which incorporates the noisy label set into the Bayesian network to generate clean labels. The generative process can then be used to predict labels for unlabeled nodes. We conduct extensive experiments to show the robustness of PRGNN on varying noise types and rates, and also on graphs with different heterophilies. In particular, we show that PRGNN can lead to inspiring performance in high noise-rate situations. The implemented code is available at https://github.com/PRGNN/PRGNN.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 4829