Temporal Graph Thumbnail: Robust Representation Learning with Global Evolutionary Skeleton

Weining Shi, Zhisen Wen, Qinggang Zhang, Chentao Zhang, Zhihong Zhang

Keywords: Dynamic Graph Neural Network, Global Evolution, Von Neumann Entropy, Robust Representation Learning

TL;DR: We introduce Temporal Graph Thumbnail (TGT), a robust representation method that captures temporal regularities via von Neumann graph entropy and feature mutual information to guide learning on noisy, rapidly evolving temporal graphs.

Abstract: Temporal graphs are commonly employed as conceptual models for capturing time-evolving interactions in real-world systems. Representation learning on such non-Euclidean data typically depends on aggregating information from neighbors, and the presence of temporal dynamics further complicates this process. However, neighbors often contain noisy information in practice, making the unreliable propagation of knowledge and may even lead to the model failure. Although existing methods employ adaptive spatiotemporal neighbor sampling strategies or temporal dependency modeling frameworks to enhance model robustness, their constrained sampling scope limits handling of severe noise and long-term dependencies. This limitation can be attributed to a fundamental cause: neglecting global evolution inherently overlooks the temporal regularities encoded in continuous dynamics. To address this, we propose the **T**emporal **G**raph **T**humbnail (**TGT**), encapsulating a temporal graph’s global evolutionary skeleton as a thumbnail to characterize temporal regularities and enhance model robustness. Specifically, we model the thumbnail by leveraging von Neumann graph entropy and node mutual information to extract essential evolutionary skeleton from the raw temporal graph, and subsequently use it to guide optimization for model learning. In addition to rigorous theoretical derivation, extensive experiments demonstrate that TGT achieves superior capability and robustness compared to baselines, particularly in rapidly evolving and noisy environments. The code is available at https://anonymous.4open.science/r/TGT-BDF2.

Primary Area: learning on graphs and other geometries & topologies

Submission Number: 1887