发表机构
Harvard University; Wuhan University; University of California, Berkeley; Tsinghua University（哈佛大学; 武汉大学; 加州大学伯克利分校; 清华大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究揭示时序链接预测忽略源节点活跃性这一关键因素，提出轻量级SNAM模型，仅用不到20个参数即可跨模型提升性能，在多个数据集上取得领先。
AI 中文摘要
一次交互包含两个部分：某人决定行动，然后选择对谁行动。时序链接预测一直专注于第二部分，而我们证明它在构造上对第一部分是盲目的：标准负样本保留真实源节点并交换目标节点，我们证明这会从最优分数中精确消除源节点的活跃性，因此以这种方式训练和评估的任何模型都不会因学习该因素而获得奖励。在更困难的历史和归纳负样本下，其源节点不同，同一因素成为主导信号。我们通过源节点活跃性建模（SNAM）来建模该因素，这是一种自激事件强度，通过精确的点过程似然拟合到历史状态中已包含的衰减交互计数；它只有不到20个参数。单独使用这些参数，从不查看目标节点，在历史负样本下，它们在五个数据集中的四个上击败了DyGFormer和TPNet。将它们添加到TPNet、TGN、DyGFormer和DSRD这四个不同设计模型的冻结分数中，无需重新训练任何内容，它们在两种设置下几乎每个骨干-数据集对上都提升了AP，最高提升25个百分点。我们的完整模型在13个数据集和三种协议上对11个基线排名第一，在百万事件流上训练一个epoch比TPNet和DyGFormer快9-100倍。我们得出结论，源节点活跃性是时序链接预测的一个盲点，并且是一个廉价且可迁移的盲点来弥补。代码可在以下网址获取：此https URL。
英文摘要
An interaction has two parts: someone decides to act, and then chooses whom to act on. Temporal link prediction has concentrated on the second, and we show that it is blind to the first by construction: a standard negative keeps the real source and swaps the destination, and we prove that this cancels the source's activity exactly from the optimal score, so no model trained and evaluated this way is ever rewarded for learning it. Under the harder historical and inductive negatives, whose sources differ, the same factor becomes the dominant signal. We model it with Source Node Activity Modeling (SNAM), a self-exciting event intensity fitted by an exact point-process likelihood to decayed interaction counts the history states already contain; it has fewer than 20 parameters. On their own, never looking at the destination, these parameters beat DyGFormer and TPNet on four of five datasets under historical negatives. Added to the frozen scores of TPNet, TGN, DyGFormer and DSRD, four models of different design, without retraining anything, they raise AP on almost every backbone-dataset pair in both settings, by up to 25 points. Our full model ranks first overall against eleven baselines on 13 datasets and three protocols, and on million-event streams trains an epoch 9-100x faster than TPNet and DyGFormer. We conclude that source activity is a blind spot of temporal link prediction, and a cheap, transferable one to close. Code is available at https://github.com/Erutaner/Your-Temporal-Link-Predictor-Is-Blind-to-Who-Is-Active.
Comments62 pages, 6 figures, 19 tables
