针对区块链反洗钱任务，研究了GNN模型的初始化与归一化策略。
Normalisation and Initialisation Strategies for Graph Neural Networks in Blockchain Anomaly Detection
- 不同GNN架构需匹配特定初始化和归一化方法
- GraphSAGE用Xavier初始化表现最佳，GAT需结合GraphNorm
- 成果可直接用于高不平衡数据的金融风控场景
图神经网络（GNN）通过联合学习节点特征与交易图拓扑，为金融欺诈检测提供了一种合理的方法。然而，其在真实世界反洗钱（AML）基准上的表现高度依赖于训练实践，如权重初始化和归一化策略，这些仍缺乏深入研究。本文系统评估了三种GNN架构（GCN、GAT、GraphSAGE）在Elliptic Bitcoin数据集上使用多种初始化与归一化策略的效果。实验表明，这些策略具有架构依赖性：GraphSAGE仅用Xavier初始化即达到最优性能；GAT在结合GraphNorm与Xavier初始化时表现最佳；而GCN对这些调整不敏感。研究结果为处理严重类别不平衡的AML数据集提供了具体部署建议。我们发布了包含时间划分数据集、种子运行及完整消融结果的可复现实验框架。
原文摘要 · Abstract (English)
Graph neural networks (GNNs) offer a principled approach to financial fraud detection by jointly learning from node features and transaction graph topology. However, their effectiveness on real-world anti-money laundering (AML) benchmarks depends critically on training practices such as specifically weight initialisation and normalisation that remain underexplored. We present a systematic ablation of initialisation and normalisation strategies across three GNN architectures (GCN, GAT, and GraphSAGE) on the Elliptic Bitcoin dataset. Our experiments reveal that initialisation and normalisation are architecture-dependent: GraphSAGE achieves the strongest performance with Xavier initialisation alone, GAT benefits most from combining GraphNorm with Xavier initialisation, while GCN shows limited sensitivity to these modifications. These findings offer practical, architecture-specific guidance for deploying GNNs in AML pipelines for datasets with severe class imbalance. We release a reproducible experimental framework with temporal data splits, seeded runs, and full ablation results.
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。