Towards Temporal Knowledge Graph Embeddings with Arbitrary Time Precision

Acknowledging the dynamic nature of knowledge graphs, the problem of learning temporal knowledge graph embeddings has recently gained attention. Essentially, the goal is to learn vector representation for the nodes and edges of a knowledge graph taking time into account. These representations must preserve certain properties of the original graph, so as to allow not only classification or clustering tasks, as for classical graph embeddings, but also approximate time-dependent query answering or link predictions over knowledge graphs. For instance, "who was the leader of Germany in 1994?'' or "when was Bonn the capital of Germany?'' Several existing work in the area adapt existing knowledge graph embedding models, adding a time dimension, usually restricting to one time granularity, like years or days, or treating time as fixed labels. However, this is not adequate for many facts of life, for instance historical and sensory data. In this work, we introduce and evaluate an approach that gracefully adjusts to time validity of virtually any granularity. Our model is robust to non-contiguous validity periods. It is generic enough to adapt to many existing non-temporal models and its size (number of parameters) does not depend on the size of the graph (number of entities and relations).