Graph Classification via Network Usable Information: From Representation Evaluation to Structure Selection

2026-07-03

AI Summary

This work addresses the limited interpretability and difficulty in evaluating representation quality inherent in traditional graph neural networks that rely on black-box embeddings. The authors propose NetinfoGC, a novel framework that introduces the Network Utility Information (NUI) paradigm to graph classification for the first time. By integrating classical structural descriptors through a propagation mechanism, NetinfoGC constructs permutation-invariant, training-free graph representations. The effectiveness of these representations is evaluated via clustering consistency, and sparse group LASSO is employed to automatically select salient features. Extensive experiments demonstrate that NUI-based classical centrality measures achieve performance on par with or superior to learned representations across multiple benchmarks. Moreover, NUI estimates exhibit strong correlation with downstream classification accuracy, substantially enhancing both model interpretability and computational efficiency.

Abstract

We propose NetinfoGC, a framework for graph classification that extends the Network Usable Information (NUI) paradigm to graph-level learning. Unlike conventional graph neural network approaches that rely on end-to-end training of black-box embeddings, NetinfoGC constructs a family of permutation-invariant graph representations derived from propagation-based mechanisms and classical structural descriptors, including graph centrality measures. To evaluate representation quality, we introduce a training-free NUI estimation procedure based on clustering consistency with ground-truth labels, providing a proxy for task-relevant information without supervised learning. We further exploit the same representations using sparse-group LASSO regularization, enabling automatic selection of informative structural descriptors while suppressing redundant ones. Experiments on benchmark datasets show that classical centrality measures are highly competitive with learned propagation-based representations, and in several cases yield superior performance. Moreover, we observe a strong correlation between estimated NUI and downstream classification accuracy, validating NUI as an effective measure of representation utility. Overall, NetinfoGC provides a unified and interpretable framework for evaluating and exploiting graph representations without requiring end-to-end neural training.

Problem

• graph classification
• Network Usable Information
• representation evaluation
• structure selection
• graph representation

Innovation

• Network Usable Information
• Graph Classification
• Permutation-invariant Representation
• Centrality Measures
• Sparse-group LASSO