Paper Info
Title
J-NERD: Joint Named Entity Recognition and Disambiguation with Rich Linguistic Features.
Abstract
Methods for Named Entity Recognition and Disambiguation (NERD) perform NERand NED in two separate stages. Therefore, NED may be penalized withrespect to precision by NER false positives, and suffers in recall fromNER false negatives. Conversely, NED does not fully exploit informationcomputed by NER such as types of mentions. This paper presents J-NERD, anew approach to perform NER and NED jointly, by means of a probabilisticgraphical model that captures mention spans, mention types, and themapping of mentions to entities in a knowledge base. We presentexperiments with different kinds of texts from the CoNLL’03, ACE’05, andClueWeb’09-FACC1 corpora. J-NERD consistently outperforms state-of-the-artcompetitors in end-to-end NERD precision, recall, and F1.
Year
Venue
Field
2016
TACL
Entity linking,Computer science,Natural language processing,Nerd,Artificial intelligence,Probabilistic logic,Graphical model,Knowledge base,Recall,False positive paradox
DocType
Volume
Citations 
Journal
4
11
PageRank 
References 
Authors
0.49
32
3
Name
Order
Citations
PageRank
Dat Ba Nguyen11275.87
Martin Theobald2147472.06
Gerhard Weikum3127102146.01