Association Rules Enhanced Knowledge Graph Attention Network

摘要

Embedding knowledge graphs into continuous vector spaces has recently attracted increasing interest in knowledge base completion. However, in most existing embedding methods, only fact triplets are utilized, and logical rules have not been thoroughly studied for the knowledge base completion task. To overcome the problem, we propose an association rules enhanced knowledge graph attention network (AR-KGAN). In this paper, triplets and logical rules are jointly modeled in the proposed unified framework to achieve more predictive entity and relation embeddings. Association rules and corresponding correlation degrees between them can be automatically obtained according to our designed mining algorithm. The major component of AR-KGAN is an encoder of an effective neighborhood aggregator, which addresses the problems by aggregating neighbors with both association rules based and graph based attention weights. The decoder enables AR-KGAN to be translational between entities and relations while keeping the superior link prediction performance. Then, the global loss is minimized over both atomic and complex formulas to achieve the embedding task. In this manner, we learn embeddings compatible with triplets and association rules, which are certainly more predictive for knowledge acquisition and inference. The results show that the proposed AR-KGAN model achieves significant and consistent improvements over state-of-the-art methods on three benchmark datasets.