Modelling Multi-relations for Convolutional-based Knowledge Graph Embedding

TL;DR

This paper introduces ConvMR, a convolutional multi-relational embedding model for knowledge graphs that captures semantic connections among multiple relations and employs an attention mechanism to weight relations, improving embedding quality.

Contribution

ConvMR is the first model to encode multi-relations into a unified vector with an attention-based encoder, enhancing semantic connection preservation in knowledge graph embeddings.

Findings

Improved mean rank on FB15k-237 and WN18RR datasets.

Effective handling of less frequent entities.

Consistent performance improvements over existing models.

Abstract

Representation learning of knowledge graphs aims to embed entities and relations into low-dimensional vectors. Most existing works only consider the direct relations or paths between an entity pair. It is considered that such approaches disconnect the semantic connection of multi-relations between an entity pair, and we propose a convolutional and multi-relational representation learning model, ConvMR. The proposed ConvMR model addresses the multi-relation issue in two aspects: (1) Encoding the multi-relations between an entity pair into a unified vector that maintains the semantic connection. (2) Since not all relations are necessary while joining multi-relations, we propose an attention-based relation encoder to automatically assign weights to different relations based on semantic hierarchy. Experimental results on two popular datasets, FB15k-237 and WN18RR, achieved consistent improvements on the mean rank. We also found that ConvMR is efficient to deal with less frequent entities.