UBC entity recognition and disambiguation at ERD 2014

Ander Barrena, Eneko Agirre, Aitor Soroa

Abstract: This paper describes the system developed at the University of the Basque Country (UBC) for the Entity Recognition and Disambiguation Challenge (ERD 2014). We developed a single system for both long and short tracks. We implemented a very basic mention detection component and complement it with a strong disambiguation step, based on Personalized PageRank algorithm. The result and confidence of the disambiguation step is used to decide whether a mention has to be linked, that is, we only link mentions if the disambiguation algorithm is confident enough. This simple method obtained good results in the ERD challenge, reaching the top 10 for both tracks.