SAOT: Self-Supervised Continual Graph Learning with Structure-Aware Optimal Transport

Self-supervised Continual Graph Learning (CGL) aims to successively learn from a graph sequence with different tasks without label supervision—a paradigm that has attracted widespread attention. Most existing self-supervised CGL methods rely on instance-level consistency objectives that enforce stability of individual node (or node-pair) embeddings. Due to optimizing nodes in isolation, these methods fail to maintain global relational structure, causing inter-node correspondences to progressively distort under continual learning. To this end, we propose a novel Structure-Aware Optimal Transport (SAOT) framework that explicitly captures and preserves relational structure within graph representations across sequential tasks. Specifically, SAOT leverages optimal transport theory to capture global inter-node correspondences, thereby facilitating and enhancing graph representation learning. Simultaneously, SAOT incorporates a cross-task knowledge distillation mechanism to preserve the previous structural knowledge. Extensive experiments on four CGL benchmark datasets demonstrate that SAOT outperforms existing self-supervised baselines. In particular, SAOT achieves significant performance gains, improving average accuracy by up to 5% on CoraFull-CL and over 15% on Products-CL compared with state-of-the-art methods in the Class-IL setting.

We developed SAOT to help a system that continually learn connected data in stages without manual labels. Consider a shopping network: each node is a product, and each link indicates that two products are often purchased together. Earlier approaches mainly try to keep each product’s learned representation stable as new data arrive. However, recognizing products one by one is not enough for understanding a shopping network: the system also needs to retain the purchasing patterns that connect them. Otherwise, it may remember the products themselves while gradually losing which products are commonly bought together. SAOT addresses this problem by learning the broader pattern formed by products and their links. It uses a pattern-matching technique called optimal transport to make the system’s internal understanding reflect the relationships in the original network. As the system learns new product categories, SAOT helps it add new knowledge while preserving the purchasing patterns learned earlier.