Cross-Lingual Entity Query from Large-Scale Knowledge Graphs

A knowledge graph is a structured knowledge system which contains a huge amount of entities and relations. It plays an important role in the field of named entity query. DBpedia, YAGO and other English knowledge graphs provide open access to huge amounts of high-quality named entities. However, Chinese knowledge graphs are still in the development stage, and contain fewer entities. The relations between entities are not rich. A natural question is: how to use mature English knowledge graphs to query Chinese named entities, and to obtain rich relation networks. In this paper, we propose a Chinese entity query system based on English knowledge graphs. For entities we build up links between Chinese entities and English knowledge graphs. The basic idea is to build a cross-lingual entity linking model, RSVM, between Chinese and English Wikipedia. RSVM is used to build cross-lingual links between Chinese entities and English knowledge graphs. The experiments show that our approach can achieve a high precision of 82.3 % for the task of finding cross-lingual entities on a test dataset. Our experiments for the sub task of finding missing cross-lingual links show that our approach has a precision of 89.42 % with a recall of 80.47 %.