Continual graph learning routinely finds its role in a variety of real-world applications where the graph data with different tasks come sequentially. Despite the success of prior works, it still faces great challenges. On the one hand, existing methods work with the zero-curvature Euclidean space, and largely ignore the fact that curvature varies over the coming graph sequence. On the other hand, continual learners in the literature rely on abundant labels, but labeling graph in practice is particularly hard especially for the continuously emerging graphs on-the-fly. To address the aforementioned challenges, we propose to explore a challenging yet practical problem, the self-supervised continual graph learning in adaptive Riemannian spaces. In this paper, we propose a novel self-supervised Riemannian Graph Continual Learner (RieGrace). In RieGrace, we first design an Adaptive Riemannian GCN (AdaRGCN), a unified GCN coupled with a neural curvature adapter, so that Riemannian space is shaped by the learnt curvature adaptive to each graph. Then, we present a Label-free Lorentz Distillation approach, in which we create teacher-student AdaRGCN for the graph sequence. The student successively performs intra-distillation from itself and inter-distillation from the teacher so as to consolidate knowledge without catastrophic forgetting. In particular, we propose a theoretically grounded Generalized Lorentz Projection for the contrastive distillation in Riemannian space. Extensive experiments on the benchmark datasets show the superiority of RieGrace, and additionally, we investigate on how curvature changes over the graph sequence.
翻译:连续的图形学习通常会发现它在一系列真实世界应用中的角色, 在这些应用中, 以不同任务相继生成的图形数据。 尽管先前的工作取得了成功, 但它仍然面临着巨大的挑战 。 一方面, 现有方法在零曲线的 Eucliidean 空间中发挥作用, 并在很大程度上忽视了曲线与即将到来的图形序列不同这一事实。 另一方面, 文献中的持续学习者依赖丰富的标签, 但在实践中标签图形特别困难, 特别是对于连续不断出现的直流直流图形来说。 为了应对上述挑战, 我们提议探索一个具有挑战性但实际的问题, 在适应性里曼空间空间的自我监督持续图形学习。 在本文中, 我们提出一个新的自我监督的里曼尼古力连续学习者( RieGrace) 。 在里格瑞, 我们首先设计一个适应性里格曼GCN (AdargrgCN), 一个统一的GCN, 加上一个不断不断出现的直径的直径直径直径调整的图形。 因此, 里曼空间的基空间由对每一图的经深曲曲调调整形成如何调整。 然后, 我们提出一个不偏直的里格连续的里格序列, 的里格的里格, 向的里格列的排序, 向自我演演演演,,, 向着, 向着, 向上, 向内校内演演进进进进进, 向, 向, 向, 方向向, 方向向, 向, 向, 方向向, 向, 方向向, 方向向, 我们向上, 方向, 向, 我们向, 向, 向, 向, 方向, 向, 向上, 向, 向, 我们向, 我们向, 向, 向, 向, 我们向, 直向, 直向, 直向, 直向, 直向, 直向, 直向, 直向, 向, 我们向, 直向, 直向, 直向, 直向, 直向, 直向, 直向, 直向, 直向, 直向, 向, 向, 向, 向, 向, 直向, 向,