Graph Condensation Method
Graph condensation aims to reduce the size of large graphs while preserving their essential properties for efficient graph neural network (GNN) training. Recent research focuses on improving the accuracy and speed of condensation, exploring methods that align gradients, feature distributions, and even the spectral properties of the original and condensed graphs, with some moving towards structure-free condensation. These advancements are significant because they enable the application of GNNs to massive datasets that were previously computationally intractable, impacting various fields requiring graph analysis.
Papers
Navigating Complexity: Toward Lossless Graph Condensation via Expanding Window Matching
Yuchen Zhang, Tianle Zhang, Kai Wang, Ziyao Guo, Yuxuan Liang, Xavier Bresson, Wei Jin, Yang You
Two Trades is not Baffled: Condensing Graph via Crafting Rational Gradient Matching
Tianle Zhang, Yuchen Zhang, Kun Wang, Kai Wang, Beining Yang, Kaipeng Zhang, Wenqi Shao, Ping Liu, Joey Tianyi Zhou, Yang You