RSCF: Relation-Semantics Consistent Filter for Entity Embedding of Knowledge Graph

Abstract

In knowledge graph embedding, leveraging relation specific entity transformation has markedly enhanced performance. However, the consistency of embedding differences before and after transformation remains unaddressed, risking the loss of valuable inductive bias in- herent in the embeddings. This inconsistency stems from two problems. First, transforma- tion representations are specified for relations in a disconnected manner, allowing dissimilar transformations and corresponding entity em- beddings for similar relations. Second, a gener- alized plug-in approach as a SFBR (Semantic Filter Based on Relations) disrupts this consis- tency through excessive concentration of entity embeddings under entity-based regularization, generating indistinguishable score distributions among relations. In this paper, we introduce a plug-in KGE method, Relation-Semantics Consistent Filter (RSCF). Its entity transfor- mation has three features for enhancing seman- tic consistency: 1) shared affine transformation of relation embeddings across all relations, 2) rooted entity transformation that adds an en- tity embedding to its change represented by the transformed vector, and 3) normalization of the change to prevent scale reduction. To amplify the advantages of consistency that pre- serve semantics on embeddings, RSCF adds relation transformation and prediction mod- ules for enhancing the semantics. In knowledge graph completion tasks with distance-based and tensor decomposition models, RSCF signifi- cantly outperforms state-of-the-art KGE meth- ods, showing robustness across all relations and their frequencies.