Graph Coarsening with Neural Networks
This work provides a data-driven approach to improve graph coarsening, which is crucial for researchers and practitioners working with large-scale graph data, offering substantial performance gains over existing methods.
The paper addresses the computational challenges of large-scale graphs by proposing a neural network-based graph coarsening method. This method significantly improves common graph coarsening techniques across various metrics, reduction ratios, graph sizes, and types, and generalizes to graphs 25 times larger than those used for training.
As large-scale graphs become increasingly more prevalent, it poses significant computational challenges to process, extract and analyze large graph data. Graph coarsening is one popular technique to reduce the size of a graph while maintaining essential properties. Despite rich graph coarsening literature, there is only limited exploration of data-driven methods in the field. In this work, we leverage the recent progress of deep learning on graphs for graph coarsening. We first propose a framework for measuring the quality of coarsening algorithm and show that depending on the goal, we need to carefully choose the Laplace operator on the coarse graph and associated projection/lift operators. Motivated by the observation that the current choice of edge weight for the coarse graph may be sub-optimal, we parametrize the weight assignment map with graph neural networks and train it to improve the coarsening quality in an unsupervised way. Through extensive experiments on both synthetic and real networks, we demonstrate that our method significantly improves common graph coarsening methods under various metrics, reduction ratios, graph sizes, and graph types. It generalizes to graphs of larger size ($25\times$ of training graphs), is adaptive to different losses (differentiable and non-differentiable), and scales to much larger graphs than previous work.