Path Ranking with Attention to Type Hierarchies
This work addresses the problem of inferring missing facts in knowledge bases for AI applications, offering an incremental improvement by balancing generalization and discrimination in path patterns.
The paper tackled the knowledge base completion problem by introducing Attentive Path Ranking, which leverages entity type hierarchies to improve path pattern representation, resulting in statistically significant performance gains of 26% on WN18RR and 10% on FB15k-237 over existing methods.
The objective of the knowledge base completion problem is to infer missing information from existing facts in a knowledge base. Prior work has demonstrated the effectiveness of path-ranking based methods, which solve the problem by discovering observable patterns in knowledge graphs, consisting of nodes representing entities and edges representing relations. However, these patterns either lack accuracy because they rely solely on relations or cannot easily generalize due to the direct use of specific entity information. We introduce Attentive Path Ranking, a novel path pattern representation that leverages type hierarchies of entities to both avoid ambiguity and maintain generalization. Then, we present an end-to-end trained attention-based RNN model to discover the new path patterns from data. Experiments conducted on benchmark knowledge base completion datasets WN18RR and FB15k-237 demonstrate that the proposed model outperforms existing methods on the fact prediction task by statistically significant margins of 26% and 10%, respectively. Furthermore, quantitative and qualitative analyses show that the path patterns balance between generalization and discrimination.