用随机草图特征增强图神经网络，更好捕捉长距离依赖。
Sketch-Augmented Features Improve Learning Long-Range Dependencies in Graph Neural Networks
- 引入随机草图特征注入图神经网络，提升全局信息感知能力
- 在多个真实数据集上显著优于基线模型，提升节点分类准确率
- 无需依赖图结构，适合长程依赖弱的复杂图数据
图神经网络通过迭代聚合局部邻域信息学习图结构数据。这一局部消息传递范式虽具备强大归纳偏置并利用图稀疏性，但也带来三大挑战：(i) 长程信息过度压缩，(ii) 节点表示过度平滑，(iii) 表达能力有限。本文提出将随机化的节点特征全局嵌入——称为「草图随机特征」（Sketched Random Features）注入标准GNN，使其能高效捕捉长程依赖。这些嵌入具有唯一性、距离敏感性和拓扑无关性，理论与实证均表明其可缓解上述问题。在真实世界图学习任务上的实验结果证实，该策略持续优于基线GNN，既可独立使用，也可作为图位置编码等现有技术的补充。代码已开源。
原文摘要 · Abstract (English)
Graph Neural Networks learn on graph-structured data by iteratively aggregating local neighborhood information. While this local message passing paradigm imparts a powerful inductive bias and exploits graph sparsity, it also yields three key challenges: (i) oversquashing of long-range information, (ii) oversmoothing of node representations, and (iii) limited expressive power. In this work we inject randomized global embeddings of node features, which we term \textit{Sketched Random Features}, into standard GNNs, enabling them to efficiently capture long-range dependencies. The embeddings are unique, distance-sensitive, and topology-agnostic -- properties which we analytically and empirically show alleviate the aforementioned limitations when injected into GNNs. Experimental results on real-world graph learning tasks confirm that this strategy consistently improves performance over baseline GNNs, offering both a standalone solution and a complementary enhancement to existing techniques such as graph positional encodings. Our source code is available at \href{https://github.com/ryienh/sketched-random-features}{https://github.com/ryienh/sketched-random-features}.
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。