moLink: Modeling link representation of knowledge base
Sogol Haghani, Mohammad Reza Keyvanpour · 2017
Knowledge bases are a significant resource for a variety of natural language processing tasks. They can be visualized as directed graphs in which nodes correspond to entities and edges represent relationships. However, knowledge bases are typically incomplete, and it is beneficial to perform knowledge base completion or link prediction, in order to describe a more complete picture of them. Embedding models for knowledge base completion aim at offering a numerical representation of entities and relations by transforming them into low dimensional vector space. In this paper, we propose a novel embedding model, called “moLink”, which is leveraging the knowledge and computing link embedding. The model is designed which is obtained high-quality distributed representation of entities and relations. Empirical experiments have proved the effectiveness of the moLink on knowledge bases.