Enrich cross-lingual entity links for online wikis via multi-modal semantic matching

作者：

Highlights：

• This paper investigated the task of enriching cross-lingual links for online wikis, which is a significant step and a pretty good starting point for cross-lingual knowledge graph construction.

• we propose two end-to-end neural matching models for matching entity descriptions and images, and then jointly train them with handcraft features. To the best of our knowledge, it is the first to utilize multi-modal information to enrich cross-lingual entity links.

• Three datasets CEMZH−ENEasy, CEMZH−ENChallenge and CEMFR−ENEasy with different languages and difficulties were created, and our approach gets the best performance compared with other baseline approaches.

摘要

•This paper investigated the task of enriching cross-lingual links for online wikis, which is a significant step and a pretty good starting point for cross-lingual knowledge graph construction.•we propose two end-to-end neural matching models for matching entity descriptions and images, and then jointly train them with handcraft features. To the best of our knowledge, it is the first to utilize multi-modal information to enrich cross-lingual entity links.•Three datasets CEMZH−ENEasy, CEMZH−ENChallenge and CEMFR−ENEasy with different languages and difficulties were created, and our approach gets the best performance compared with other baseline approaches.

论文关键词：Cross-lingual entity matching,Multi-modal semantic matching,Entity description matching,Entity image matching

论文评审过程：Received 21 August 2019, Revised 13 March 2020, Accepted 16 April 2020, Available online 15 May 2020, Version of Record 15 May 2020.

论文官网地址：https://doi.org/10.1016/j.ipm.2020.102271