CacheDyG：解耦时间传播以实现高效动态图学习
CacheDyG: Decoupling Temporal Propagation for Efficient Dynamic Graph Learning
查看机构详情
- College of Computer and Information Science, Southwest University（西南大学计算机与信息科学学院）
机构由 AI 辅助整理，请以论文原文为准。
浏览论文内容
中文总结 AI 辅助
针对动态图学习中的高训练和参数开销，提出CacheDyG框架，通过缓存解耦时间传播，仅更新轻量组件，在五个基准上以更少参数和更低耗时取得更优性能。
中文摘要 AI 辅助
动态图被广泛用于建模现实世界应用中随时间演化的关系系统。动态图神经网络为捕获此类数据中的结构依赖和时间动态提供了有效框架。然而，它们通常将时间图传播与每次优化迭代交织在一起，并常常为每个节点-时间对维护大型可训练表示。这种设计反复重新计算基本不变的历史结构，导致大量的训练和参数开销。为解决这一关键问题，我们提出了CacheDyG，一种用于高效动态图学习的缓存精炼框架。具体而言，它通过构建一个按时间排序的时间依赖缓存，将时间传播从常规参数更新中解耦，该缓存以非可训练缓冲区存储图感知的节点-时间表示。在标准训练迭代中，CacheDyG从缓存中读取，仅更新轻量级缓存精炼器、自适应残差门和链接预测器。选择性缓存刷新进一步使缓存表示与监督目标保持一致，同时避免逐迭代的稀疏传播。在五个动态图基准上的实验表明，CacheDyG采用显著更少的可训练参数和更低的运行时间，获得了比基线更具竞争力的预测性能。这些结果表明，基于缓存的解耦为可扩展动态图学习提供了有效原则。
英文摘要
Dynamic graphs are widely used to model time-evolving relational systems in real-world applications. Dynamic graph neural networks provide an effective framework for capturing both structural dependencies and temporal dynamics in such data. However, they typically intertwine temporal graph propagation with every optimization epoch and often maintain large trainable representations for each node-time pair. This design repeatedly recomputes largely unchanged historical structures, leading to substantial training and parameter overhead. To address this critical issue, we propose CacheDyG, a Cache-refine framework for efficient Dynamic Graph learning. Specifically, it decouples temporal propagation from routine parameter updates by constructing a time-ordered temporal dependency cache that stores graph-aware node-time representations in non-trainable buffers. During standard training epochs, CacheDyG reads from the cache and updates only a lightweight cache refiner, an adaptive residual gate, and the link predictor. Selective cache refresh further keeps cached representations aligned with the supervised objective while avoiding epoch-wise sparse propagation. Experiments on five dynamic graph benchmarks show that CacheDyG adopts substantially fewer trainable parameters and lower runtime to obtain more competitive predictive performance than baselines. These results demonstrate that cache-based decoupling provides an effective principle for scalable dynamic graph learning.
