Join TensorFlow at Google I/O, May 11-12 Register now

gap

  • Description:

GAP is a gender-balanced dataset containing 8,908 coreference-labeled pairs of (ambiguous pronoun, antecedent name), sampled from Wikipedia and released by Google AI Language for the evaluation of coreference resolution in practical applications.

Split Examples
'test' 2,000
'train' 2,000
'validation' 454
  • Feature structure:
FeaturesDict({
    'A': Text(shape=(), dtype=tf.string),
    'A-coref': tf.bool,
    'A-offset': tf.int32,
    'B': Text(shape=(), dtype=tf.string),
    'B-coref': tf.bool,
    'B-offset': tf.int32,
    'ID': Text(shape=(), dtype=tf.string),
    'Pronoun': Text(shape=(), dtype=tf.string),
    'Pronoun-offset': tf.int32,
    'Text': Text(shape=(), dtype=tf.string),
    'URL': Text(shape=(), dtype=tf.string),
})
  • Feature documentation:
Feature Class Shape Dtype Description
FeaturesDict
A Text tf.string
A-coref Tensor tf.bool
A-offset Tensor tf.int32
B Text tf.string
B-coref Tensor tf.bool
B-offset Tensor tf.int32
ID Text tf.string
Pronoun Text tf.string
Pronoun-offset Tensor tf.int32
Text Text tf.string
URL Text tf.string
  • Citation:
@article{DBLP:journals/corr/abs-1810-05201,
  author    = {Kellie Webster and
               Marta Recasens and
               Vera Axelrod and
               Jason Baldridge},
  title     = {Mind the {GAP:} {A} Balanced Corpus of Gendered Ambiguous Pronouns},
  journal   = {CoRR},
  volume    = {abs/1810.05201},
  year      = {2018},
  url       = {http://arxiv.org/abs/1810.05201},
  archivePrefix = {arXiv},
  eprint    = {1810.05201},
  timestamp = {Tue, 30 Oct 2018 20:39:56 +0100},
  biburl    = {https://dblp.org/rec/bib/journals/corr/abs-1810-05201},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}