STSA 针对动态文本属性图的跨域边分类提出迁移学习协议
Transfer Learning for Edge Classification on Dynamic Text-Attributed Graphs
做图学习的朋友可以看这篇：作者发现现有动态图模型跨域迁移竟然打不过一个只看文本的 BoE 基线，然后提出了 STSA 来解决。
论文针对动态文本属性图（DyTAG）上的边分类任务，建立了 leave-one-domain-out（LODO）迁移学习协议。在该协议下，现有自监督动态图学习方法迁移到未见域时表现不佳，甚至低于论文提出的仅输入无序文本特征的 Bag of Events（BoE）基线。论文进一步提出 Spatio-Temporal Semantic Alignment（STSA），用时空编码器融合时间差与节点出现频率的表示，并以 Contrastive Semantic Forecasting 目标训练，将边表示锚定到由预训练语言模型初始化的多域文本潜空间。STSA 的迁移成绩超过 BoE 及论文评估的所有现有方法。
Transfer Learning for Edge Classification on Dynamic Text-Attributed Graphs
Learning transferable representations for dynamic text-attributed graphs (DyTAGs) requires models to capture underlying interaction dynamics that persist across domains. However, existing methods tend to overfit to domain-specific structural, temporal, and semantic patterns, limiting edge classification performance under distribution shifts. To expose and address this, we formally establish a leave-one-domain-out (LODO) transfer learning protocol for edge classification on DyTAGs. Under this protocol, we demonstrate that state-of-the-art self-supervised methods for dynamic graph learning perform poorly when transferred to unseen domains. Strikingly, existing methods underperform a structurally and temporally unaware Bag of Events (BoE) model we introduce, which inputs only unordered sequences of node and edge text features. Proceeding from the BoE, we propose Spatio-Temporal Semantic Alignment (STSA), which integrates a spatio-temporal encoder that fuses representations of time deltas and node occurrence frequencies into a unified manifold. STSA is trained with a Contrastive Semantic Forecasting objective, which anchors edge representations to a multi-domain textual latent space initialized by a pretrained language model, providing a robust prior that outperforms BoE and all existing methods we evaluate.
