Retrofitting Soft Rules for Knowledge Representation Learning

Abstract Recently, a significant number of studies have focused on knowledge graph completion using rule-enhanced learning techniques, supported by the mined soft rules in addition to the hard logic rules. However, due to the difficulty in determining the confidence of the soft rules without the global semantics of knowledge graph such as the semantic relatedness between relations, the knowledge representation may not be optimal, leading to degraded effectiveness in its application to knowledge graph completion tasks. To address this challenge, this paper proposes a retrofit framework that iteratively enhances the knowledge representation and confidence of soft rules. Specifically, the soft rules guide the learning of knowledge representation, and the representation, in turn, provides global semantics of the knowledge graph to optimize the confidence of soft rules. Extensive evaluation shows that our method achieves state-of-the-art results on link prediction and triple classification tasks, brought by the fine-tuned soft rules.