Knowledge Graph Embedding Bi-Vector Models for Symmetric Relation
This work tackles a specific issue in knowledge graph reasoning for AI applications, offering an incremental improvement in handling symmetric relations.
The paper addresses the problem of symmetric relations in knowledge graph embedding models tending to zero vectors during training, which impairs link prediction tasks, and proposes bi-vector models that represent symmetric relations as vector pairs to improve performance, with experiments on benchmark datasets showing effectiveness and superiority over baselines.
Knowledge graph embedding (KGE) models have been proposed to improve the performance of knowledge graph reasoning. However, there is a general phenomenon in most of KGEs, as the training progresses, the symmetric relations tend to zero vector, if the symmetric triples ratio is high enough in the dataset. This phenomenon causes subsequent tasks, e.g. link prediction etc., of symmetric relations to fail. The root cause of the problem is that KGEs do not utilize the semantic information of symmetric relations. We propose KGE bi-vector models, which represent the symmetric relations as vector pair, significantly increasing the processing capability of the symmetry relations. We generate the benchmark datasets based on FB15k and WN18 by completing the symmetric relation triples to verify models. The experiment results of our models clearly affirm the effectiveness and superiority of our models against baseline.