Simple Truncated SVD based Model for Node Classification on Heterophilic Graphs

TL;DR

This paper introduces a simple TSVD-based model for node classification on heterophilic graphs, achieving significant performance improvements over existing methods by leveraging topological structure and node features.

Contribution

It proposes a novel, non-aggregation approach using Truncated SVD for heterophilic graph node classification, challenging traditional aggregation-based methods.

Findings

Up to 30% performance improvement over state-of-the-art methods

TSVD approach effectively captures heterophilic graph structures

Explores alternative methods beyond aggregation for heterophilic graphs

Abstract

Graph Neural Networks (GNNs) have shown excellent performance on graphs that exhibit strong homophily with respect to the node labels i.e. connected nodes have same labels. However, they perform poorly on heterophilic graphs. Recent approaches have typically modified aggregation schemes, designed adaptive graph filters, etc. to address this limitation. In spite of this, the performance on heterophilic graphs can still be poor. We propose a simple alternative method that exploits Truncated Singular Value Decomposition (TSVD) of topological structure and node features. Our approach achieves up to ~30% improvement in performance over state-of-the-art methods on heterophilic graphs. This work is an early investigation into methods that differ from aggregation based approaches. Our experimental results suggest that it might be important to explore other alternatives to aggregation methods for heterophilic setting.