AI 中文总结
研究如何将知识图谱基础模型的归纳迁移能力扩展到时间知识图谱，提出GRATE方法，通过门控旋转注意力编码时间，集成到NBFNet模型中，构建新基准套件测试，单个联合预训练的GRATE检查点在多数设置下优于静态基础模型。
AI 中文摘要
诸如Ultra和Trix等知识图谱基础模型通过学习可推广到未见实体和关系的关系图表示来实现强大的归纳迁移。将这种可迁移性扩展到时间知识图谱（TKG）仍然具有挑战性：现有时间模型将其参数与特定数据集的实体、关系或时间戳绑定，并非为转移到具有不相交词汇表的TKG而设计。我们提出GRATE（用于时间编码的门控旋转注意力），这是一种实体侧消息函数，不添加可学习参数，通过根据与查询的时间差旋转每条边消息并应用查询条件门来选择时间相关信号，从而通过相对时间差对时间进行编码。GRATE可集成到NBFNet风格的知识图谱基础模型中，同时保留结构可迁移性。现有的TKG基准在共享的训练/测试词汇表内进行评估，无法直接测试跨数据集的时间转移；因此，我们构建了GDELTIndT和WIKIIndT，这是具有不相交实体、关系和时间戳的归纳转移基准套件，涵盖内插和外推。在这些基准和预留的预测数据集上，单个联合预训练的GRATE检查点在大多数设置下优于静态基础模型。
英文摘要
Knowledge graph foundation models such as Ultra and Trix achieve strong inductive transfer by learning relation-graph representations that generalise to unseen entities and relations. Extending this transferability to temporal knowledge graphs (TKGs) remains challenging: existing temporal models tie their parameters to dataset-specific entities, relations, or timestamps and are not designed to transfer to TKGs with disjoint vocabularies. We propose GRATE (Gated Rotary Attention for Temporal Encoding), an entity-side message function that adds no learnable parameters and encodes time through relative time differences by rotating each edge message according to its time gap to the query and applying a query-conditioned gate to select temporally relevant signals. GRATE integrates into NBFNet-style KG foundation models while preserving structural transferability. Existing TKG benchmarks evaluate within shared train/test vocabularies and cannot directly test cross-dataset temporal transfer; we therefore construct GDELTIndT and WIKIIndT, inductive transfer benchmark suites with disjoint entities, relations, and timestamps spanning both interpolation and extrapolation. Across these benchmarks and held-out forecasting datasets, a single jointly pretrained GRATE checkpoint improves over the static base model in most settings.
CommentsAccepted at the ICML 2026 Workshop on Graph Foundation Models: A New Era for Graph Machine Learning. 17 pages, 4 figures
