OntoTKGE: Ontology-Enhanced Temporal Knowledge Graph Extrapolation
Summary: arXiv:2604.05468v1 Announce Type: new
Abstract: Temporal knowledge graph (TKG) extrapolation is an important task that aims to predict future facts through historical interaction information within KG snapshots. A key challenge for most existing TKG extrapolation models is handling entities with sparse historical interaction. The ontological knowledge is beneficial for alleviating this sparsity issue by enabling these entities to inherit behavioral patterns from other entities with the same concept, which is ignored by previous studies.
In this paper, we propose a novel encoder-decoder framework OntoTKGE that leverages the ontological knowledge from the ontology-view KG (i.e., a KG modeling hierarchical relations among abstract concepts as well as the connections between concepts and entities) to guide the TKG extrapolation model’s learning process through the effective integration of the ontological and temporal knowledge, thereby enhancing entity embeddings. OntoTKGE is flexible enough to adapt to many TKG extrapolation models. Extensive experiments on four data sets demonstrate that OntoTKGE not only significantly improves the performance of many TKG extrapolation models but also surpasses many SOTA baseline methods.
Introduction
Temporal knowledge graphs (TKGs) are crucial in understanding and predicting temporal events and relationships in a dynamic environment. As such, TKG extrapolation plays a vital role in various applications, such as social network analysis, recommendation systems, and event forecasting.
Challenges in TKG Extrapolation
Despite the advancements in TKGs, several challenges remain, particularly in relation to:
- Sparse Historical Interactions: Many entities lack sufficient historical data, making it difficult for models to learn effective patterns.
- Entity Inference: Existing models often fail to recognize that entities sharing similar characteristics can inherit behavioral traits from one another.
- Integration of Knowledge Types: There is a need for effective methodologies to integrate ontological and temporal knowledge seamlessly.
Proposed Solution: OntoTKGE
OntoTKGE addresses these challenges by introducing a hybrid framework that combines both ontological and temporal knowledge. Key features of OntoTKGE include:
- Ontology-View Knowledge Graph: This component models hierarchical relationships among abstract concepts and their connections to specific entities, providing a rich source of information.
- Enhanced Learning Process: By integrating ontological knowledge, OntoTKGE helps in improving the learning process of TKG extrapolation models.
- Flexibility and Adaptability: The framework can be easily adapted to a variety of existing TKG models, enhancing their performance.
Experimental Results
To validate the efficacy of OntoTKGE, extensive experiments were conducted across four diverse datasets. The results demonstrated a significant improvement in the performance metrics of various TKG extrapolation models, showcasing:
- Enhanced predictive accuracy.
- Robustness in handling sparse data situations.
- Superior performance compared to state-of-the-art baseline methods.
Conclusion
OntoTKGE represents a significant advancement in the field of temporal knowledge graph extrapolation. By effectively leveraging ontological knowledge, it not only overcomes the limitations of existing models but also opens up new avenues for future research and application in temporal knowledge graphs.