Wikipedia Annotated Data

Abbas Ghaddar, Philippe Langlais

* This work was done in collaboration with the NLP
team of Huawei's Noah's Ark Lab in Montreal

We gratefully acknowledge the support of NVIDIA Corporation
with the donation of the Titan Xp GPU used for this research.


Main Concept Entities in Wikipedia

This resource is the full English Wikipedia dump of April 2013, where all mentions coreferring to the main concept are automatically extracted using the classifier described in this article, along with information we extracted from Wikipedia and Freebase.

WiNER: Coarse Named Entities in Wikipedia

This resource is the full English Wikipedia dump of April 2013, where all mentions are automatically annotated with coarse named entity types (PER, LOC, ORG and MISC) as described in this article. The baseline classifier proposed in the article can be found on this GitHub page.

WiFiNE: Transforming Wikipedia into a Large-Scale Fine-Grained Entity Type Corpus

This resource is the full English Wikipedia dump of April 2013, where all mentions are automatically annotated with fine-grained entity type following Figer (113 types) and Gillick (89 types) schemes as described in this article.

Robust Lexical Features for Improved Neural Network Named-Entity Recognition

The code used in the paper can be found on this GitHub page.

Contextualized Word Representations from Distant Supervision with and for NER

The code used in the paper can be found on this GitHub page.

* Context-aware Adversarial Training for Name Regularity Bias in Named Entity Recognition

The data and code used in the paper can be found at this link.

 

Download most of the resources here.