Towards Understanding and Reducing Graph Structural Noise for GNNs

Published: 24 Apr 2023, Last Modified: 15 Jun 2023ICML 2023 PosterEveryoneRevisions
Abstract: Graph neural networks (GNNs) have emerged as a powerful paradigm to learn from relational data mostly through applying the message passing mechanism. However, this approach may exhibit suboptimal performance when applied to graphs possessing various structural issues. In this work, we focus on understanding and alleviating the effect of graph structural noise on GNN performance. To evaluate the graph structural noise in real data, we propose edge signal-to-noise ratio (ESNR), a novel metric evaluating overall edge noise level with respect to data features or labels based on random matrix theory. We have found striking concordance between the proposed ESNR metric and the GNN performance in various simulated and real data. To reduce the effect of the noise, we propose GPS (Graph Propensity Score) graph rewiring, which estimates the edge likelihood for rewiring data graphs based on self-supervised link prediction. We provide a theoretical guarantee for GPS graph rewiring and demonstrate its efficacy by comprehensive benchmarks.
Submission Number: 6126
Loading