Self-Supervised Temporal Graph learning with Temporal and Structural Intensity Alignment
Meng Liu, Ke Liang, Yawei Zhao, Wenxuan Tu, Sihang Zhou, Xinbiao Gan, Xinwang Liu, Kunlun He
TL;DR
S2T tackles temporal graph learning by explicitly aligning temporal information modeled via a Hawkes process with high-order structural information captured through GNNs, supplemented by a global representation to assist long-tail nodes. The framework jointly optimizes a task loss for link prediction, an alignment loss between two intensity vectors, and a global loss, achieving improved performance over state-of-the-art on multiple datasets with robustness to noise and incomplete data. The key contributions include the dual-intensity formulation, a global structural augmentation, and a learnable FiLM-based global parameter to weight intensity dimensions. The approach offers a scalable, inductive-capable solution that advances temporal-structural fusion in dynamic graphs, with practical impact for forecasting interactions in real-world networks.
Abstract
Temporal graph learning aims to generate high-quality representations for graph-based tasks with dynamic information, which has recently garnered increasing attention. In contrast to static graphs, temporal graphs are typically organized as node interaction sequences over continuous time rather than an adjacency matrix. Most temporal graph learning methods model current interactions by incorporating historical neighborhood. However, such methods only consider first-order temporal information while disregarding crucial high-order structural information, resulting in suboptimal performance. To address this issue, we propose a self-supervised method called S2T for temporal graph learning, which extracts both temporal and structural information to learn more informative node representations. Notably, the initial node representations combine first-order temporal and high-order structural information differently to calculate two conditional intensities. An alignment loss is then introduced to optimize the node representations, narrowing the gap between the two intensities and making them more informative. Concretely, in addition to modeling temporal information using historical neighbor sequences, we further consider structural knowledge at both local and global levels. At the local level, we generate structural intensity by aggregating features from high-order neighbor sequences. At the global level, a global representation is generated based on all nodes to adjust the structural intensity according to the active statuses on different nodes. Extensive experiments demonstrate that the proposed model S2T achieves at most 10.13% performance improvement compared with the state-of-the-art competitors on several datasets.