提出可学习的动态邻居选择机制，提升时序图神经网络预测性能。
FLASH: Flexible Learning of Adaptive Sampling from History in Temporal Graph Neural Networks
- 设计可自适应学习的邻居采样方法，替代固定规则采样。
- 在多个数据集上显著提升链接预测准确率，优于传统采样策略。
- 适合需要高效利用历史信息的时序图建模任务。
在动态图中进行未来链接预测时，聚合历史交互信号是关键步骤。然而，引入过长的历史信息会带来高资源开销。因此，时序图神经网络（TGNNs）通常采用均匀采样或仅选择近期邻居等静态启发式方法。这些方法缺乏对图结构的适应性。本文提出FLASH，一种可学习且图自适应的邻域选择机制，能够泛化现有启发式方法。FLASH可无缝集成到TGNN中，并通过自监督排序损失端到端训练。我们提供了理论证据，说明常用启发式方法会限制TGNN性能，从而推动本设计。在多个基准数据集上的大量实验表明，配备FLASH的TGNN在各项指标上均实现持续且显著的性能提升。
原文摘要 · Abstract (English)
Aggregating temporal signals from historic interactions is a key step in future link prediction on dynamic graphs. However, incorporating long histories is resource-intensive. Hence, temporal graph neural networks (TGNNs) often rely on historical neighbors sampling heuristics such as uniform sampling or recent neighbors selection. These heuristics are static and fail to adapt to the underlying graph structure. We introduce FLASH, a learnable and graph-adaptive neighborhood selection mechanism that generalizes existing heuristics. FLASH integrates seamlessly into TGNNs and is trained end-to-end using a self-supervised ranking loss. We provide theoretical evidence that commonly used heuristics hinder TGNNs performance, motivating our design. Extensive experiments across multiple benchmarks demonstrate consistent and significant performance improvements for TGNNs equipped with FLASH.
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。