GC4NC: A Benchmark Framework for Graph Condensation on Node Classification with New Insights

2024-06-24

AI Summary

Graph compression (GC) for node classification has long suffered from the absence of a unified evaluation framework, hindering fair method comparison and principled design optimization. To address this, we propose GC4NC—the first dedicated benchmark for GC in node classification—featuring a six-dimensional evaluation framework encompassing performance, efficiency, privacy preservation, denoising capability, neural architecture search (NAS) support, and cross-domain transferability. We introduce the first multi-dimensional evaluation paradigm, uncovering two core design principles: structural preservation and task alignment, while empirically exposing substantial discrepancies between conventional reconstruction metrics and downstream classification accuracy. Leveraging a hybrid protocol combining graph reconstruction, label propagation, and meta-learning, we conduct over 2,000 experiments across 12 benchmark datasets, systematically characterizing the compression–accuracy trade-off. Our framework enables 5.3× NAS acceleration and supports cross-graph generalization and noise-robust training.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Data Compression
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

Graph condensation (GC) is an emerging technique designed to learn a significantly smaller graph that retains the essential information of the original graph. This condensed graph has shown promise in accelerating graph neural networks while preserving performance comparable to those achieved with the original, larger graphs. Additionally, this technique facilitates downstream applications like neural architecture search and deepens our understanding of redundancies in large graphs. Despite the rapid development of GC methods, particularly for node classification, a unified evaluation framework is still lacking to systematically compare different GC methods or clarify key design choices for improving their effectiveness. To bridge these gaps, we introduce extbf{GC4NC}, a comprehensive framework for evaluating diverse GC methods on node classification across multiple dimensions including performance, efficiency, privacy preservation, denoising ability, NAS effectiveness, and transferability. Our systematic evaluation offers novel insights into how condensed graphs behave and the critical design choices that drive their success. These findings pave the way for future advancements in GC methods, enhancing both performance and expanding their real-world applications. Our code is available at https://github.com/Emory-Melody/GraphSlim/tree/main/benchmark.

Problem

• Lack unified framework to evaluate graph condensation methods for node classification
• Need systematic comparison of GC methods across multiple performance dimensions
• Require clarification of key design choices for effective graph condensation

Innovation

• Develops comprehensive benchmark framework GC4NC
• Evaluates graph condensation methods across multiple dimensions
• Provides insights on condensed graph behavior and design choices