通过系统实验提升图神经网络集成的多样性，显著增强性能。
Training Diverse Graph Experts for Ensembles: A Systematic Empirical Study
- 设计20种多样化策略，让多个图神经网络各具不同学习模式。
- 在14个数据集上构建超200个集成模型，验证多样性能有效提效。
- 为图神经网络专家集成提供可落地的设计指南，适合模型优化者参考。
图神经网络（GNN）已成为处理关系型数据的重要工具，但单一GNN的性能常受限于真实图数据的异质性。近年来，混合专家（MoE）框架的发展表明，将多个具有不同泛化特性的GNN组合起来可显著提升性能。本文首次对GNN集成中的专家级多样化技术进行了系统的实证研究。我们在14个节点分类基准上评估了20种多样化策略——包括随机重初始化、超参数调优、架构差异、方向性建模及训练数据划分——构建并分析了超过200种集成变体。全面评估涵盖每种方法在专家多样性、互补性与集成性能方面的表现。研究还揭示了训练高度多样化专家的内在机制。这些发现为图数据上的有效MoE框架设计提供了可操作的指导。代码已公开于 https://github.com/Hydrapse/bench-gnn-diversification。
原文摘要 · Abstract (English)
Graph Neural Networks (GNNs) have become essential tools for learning on relational data, yet the performance of a single GNN is often limited by the heterogeneity present in real-world graphs. Recent advances in Mixture-of-Experts (MoE) frameworks demonstrate that assembling multiple, explicitly diverse GNNs with distinct generalization patterns can significantly improve performance. In this work, we present the first systematic empirical study of expert-level diversification techniques for GNN ensembles. Evaluating 20 diversification strategies -- including random re-initialization, hyperparameter tuning, architectural variation, directionality modeling, and training data partitioning -- across 14 node classification benchmarks, we construct and analyze over 200 ensemble variants. Our comprehensive evaluation examines each technique in terms of expert diversity, complementarity, and ensemble performance. We also uncovers mechanistic insights into training maximally diverse experts. These findings provide actionable guidance for expert training and the design of effective MoE frameworks on graph data. Our code is available at https://github.com/Hydrapse/bench-gnn-diversification.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。