WeiChow/DyGraphs

该数据集是为论文CrossLink准备的动态图数据。CrossLink学习特定下游图的演化模式，并进行模式特定的链接预测。它采用了一种称为条件链接生成的技术，该技术结合了演化和结构建模，以执行演化特定的链接预测。这种条件链接生成是通过一个transformer-decoder架构实现的，支持高效的并行训练和推理。CrossLink在跨领域的广泛动态图上进行了训练，涵盖了600万条动态边。在八个未训练图上的广泛实验表明，CrossLink在跨域链接预测中实现了最先进的性能。与相同设置下的先进基线相比，CrossLink在八个图上的平均精度平均提高了11.40%。令人印象深刻的是，它在六个未训练图上超越了8个先进基线的完全监督性能。

The CrossLink dataset is used to learn the evolution pattern of a specific downstream graph and make pattern-specific link predictions. It employs a technique called conditioned link generation, which integrates both evolution and structure modeling to perform evolution-specific link prediction. The dataset contains 6 million dynamic edges across diverse domains. Experimental results show that CrossLink achieves state-of-the-art performance in cross-domain link prediction, with an average precision improvement of 11.40%. The dataset is in CSV format, including source node, target node, interaction time, and other information.