Graphs, Entities, and Step Mixture
This addresses a key problem in graph learning for researchers and practitioners, offering improved performance in both transductive and inductive scenarios, though it appears incremental as it builds on existing methods.
The paper tackles the oversmoothing and poor generalization issues in graph neural networks by proposing GESM, which uses a mixture of random walk steps and attention mechanisms, achieving state-of-the-art or comparable results on eight benchmark datasets.
Existing approaches for graph neural networks commonly suffer from the oversmoothing issue, regardless of how neighborhoods are aggregated. Most methods also focus on transductive scenarios for fixed graphs, leading to poor generalization for unseen graphs. To address these issues, we propose a new graph neural network that considers both edge-based neighborhood relationships and node-based entity features, i.e. Graph Entities with Step Mixture via random walk (GESM). GESM employs a mixture of various steps through random walk to alleviate the oversmoothing problem, attention to dynamically reflect interrelations depending on node information, and structure-based regularization to enhance embedding representation. With intensive experiments, we show that the proposed GESM achieves state-of-the-art or comparable performances on eight benchmark graph datasets comprising transductive and inductive learning tasks. Furthermore, we empirically demonstrate the significance of considering global information.