FITTER: Vocabulary-Agnostic Cross-Domain Inference on Temporal Knowledge Graphs

2026-08-11

AI Summary

Existing methods struggle to generalize to out-of-domain temporal knowledge graphs containing entirely new entities, relations, and timestamps. This work proposes the first vocabulary-agnostic, fully inductive structural model that enables cross-graph link prediction without retraining and exhibits temporal translation invariance. By leveraging interaction-pattern-based relation representations, relative temporal encoding, and a temporally aware message-passing mechanism, the model effectively captures dynamic relational structures across diverse temporal contexts. Evaluated on six benchmarks spanning varied domains, granularities, and temporal extents, the approach significantly outperforms current inductive baselines, demonstrating strong generalization capability and effectiveness in cross-domain temporal reasoning.

Abstract

Temporal knowledge graphs are central to many uses of the Semantic Web, but existing completion methods assume the entities, relation names, and timestamps to be reasoned about are already known at training time, restricting each model to a single graph and vocabulary. We propose FITTER, the first fully-inductive structural model for temporal knowledge graph link prediction that supports cross-domain transfer: the inference graph may contain entirely unseen entities, relation names, and timestamps drawn from a different domain. FITTER represents each predicate by its interaction patterns with others and time through encodings of relative rather than absolute ordering; message-passing fuses local and global temporal context to produce vocabulary-agnostic embeddings. We prove the temporal encoding is time-shift invariant and evaluate FITTER on cross-domain, cross-graph transfer over six temporal knowledge graph benchmarks of diverse domains, granularities, and time spans. FITTER consistently outperforms inductive baselines without retraining, indicating that vocabulary-agnostic structural learning is a viable foundation for inference over the heterogeneous knowledge graphs of the Semantic Web.

Problem

• temporal knowledge graphs
• cross-domain inference
• vocabulary-agnostic
• link prediction
• inductive reasoning

Innovation

• temporal knowledge graph
• inductive reasoning
• vocabulary-agnostic
• cross-domain transfer
• time-shift invariance