Knowledge Graph Embedding with Linear Representation for Link Prediction
Year: 2020
Type: preprint
Abstract: Knowledge graph (KG) embedding aims to represent entities and relations in KGs as vectors in a continuous vector space. If an embedding model can cover different types of connectivity patterns and mapping properties of relations as many as possible, it will potentially bring more benefits for applications. In this paper, we propose a novel embedding model, namely LineaRE, which is capable of modeling four connectivity patterns (symmetry, antisymmetry, inversion, and composition) and four mapping properties of relations (one-to-one, one-to-many, many-to-one, and many-to-many). Specifically, in our model, a relation is a linear function of two low-dimensional vector-presented entities with two weight vectors and a bias vector. Since the vectors are defined in a real number space and the scoring function of the model is linear, our model is simple and scalable to large KGs. Experimental results on four datasets show that the proposed LineaRE significantly outperforms existing state-of-the-art models for link prediction task.
Source: arXiv (Cornell University)
Authors: Yanhui Peng, Jing Zhang
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3020032093