PPKE: Knowledge Representation Learning by Path-based Pre-training
This work provides a feasible way to integrate more graph contextual information into knowledge representation learning models, which is significant for researchers working on knowledge graph embeddings and their applications.
This paper addresses the limitation of traditional knowledge representation learning (KRL) methods that neglect multi-step relationships and graph contextual information in knowledge graphs. The authors propose PPKE, a Path-based Pre-training model, which achieves state-of-the-art results on several benchmark datasets for link prediction and relation prediction tasks.
Entities may have complex interactions in a knowledge graph (KG), such as multi-step relationships, which can be viewed as graph contextual information of the entities. Traditional knowledge representation learning (KRL) methods usually treat a single triple as a training unit, and neglect most of the graph contextual information exists in the topological structure of KGs. In this study, we propose a Path-based Pre-training model to learn Knowledge Embeddings, called PPKE, which aims to integrate more graph contextual information between entities into the KRL model. Experiments demonstrate that our model achieves state-of-the-art results on several benchmark datasets for link prediction and relation prediction tasks, indicating that our model provides a feasible way to take advantage of graph contextual information in KGs.