While Graph Neural Networks (GNNs) have made significant strides in diverse areas, they are hindered by a theoretical constraint known as the 1-Weisfeiler-Lehman test. Even though latest advancements in higher-order GNNs can overcome this boundary, they typically center around certain graph components like cliques or cycles. However, our investigation goes a different route. We put emphasis on paths, which are inherent in every graph. We are able to construct a more general topological perspective and form a bridge to certain established theories about other topological domains. Interestingly, without any assumptions on graph sub-structures, our approach surpasses earlier techniques in this field, achieving state-of-the-art performance on several benchmarks.

翻译：尽管图神经网络（GNN）在多个领域取得了显著进展，但其仍受制于一种被称为1-Weisfeiler-Lehman检验的理论约束。即使高阶GNN的最新进展能够突破这一界限，这些方法通常聚焦于特定的图组件（如团或环）。然而，我们的研究另辟蹊径——重点聚焦于每条图中天然存在的路径。我们得以构建更通用的拓扑视角，并与拓扑域其他既定理论建立桥梁。值得关注的是，我们的方法无需对图子结构作任何假设，即可超越该领域先前技术，在多个基准测试中达到当前最优性能。