KBP37 is a revision of MIML-RE annotation dataset, provided by Gabor Angeli et al. (2014). They use both the 2010 and 2013 KBP official document collections, as well as a July 2013 dump of Wikipedia as the text corpus for annotation. There are 33811 sentences been annotated. Zhang and Wang made several refinements:

They add direction to the relation names, e.g. 'per:employee_of' is split into 'per:employee of(e1,e2)' and 'per:employee of(e2,e1)'. They also replace 'org:parents' with 'org:subsidiaries' and replace 'org:member of’ with 'org:member`' (by their reverse directions).
They discard low frequency relations such that both directions of each relation occur more than 100 times in the dataset.

KBP37 contains 18 directional relations and an additional 'no_relation' relation, resulting in 37 relation classes.

Note:

There is a formatted version that you can load with datasets.load_dataset('kbp37', name='kbp37_formatted'). This version is tokenized with str.split() and provides entities as offsets instead of being enclosed by xml tags. It discards some examples, however, that are invalid in the original dataset and lead to entity offset errors, e.g. example train/1276.

Supported Tasks and Leaderboards

More Information Needed

Languages

The language data in KBP37 is in English (BCP-47 en)

Dataset Structure

Data Instances

kbp37

Size of downloaded dataset files: 5.11 MB
Size of the generated dataset: 4.7 MB An example of 'train' looks as follows:

{
  "id": "0",
  "sentence": "<e1> Thom Yorke </e1> of <e2> Radiohead </e2> has included the + for many of his signature distortion sounds using a variety of guitars to achieve various tonal options .",
  "relation": 27
}

kbp37_formatted

Size of downloaded dataset files: 5.11 MB
Size of the generated dataset: 6.58 MB An example of 'train' looks as follows:

{
  "id": "1",
  "token": ["Leland", "High", "School", "is", "a", "public", "high", "school", "located", "in", "the", "Almaden", "Valley", "in", "San", "Jose", "California", "USA", "in", "the", "San", "Jose", "Unified", "School", "District", "."],
  "e1_start": 0,
  "e1_end": 3,
  "e2_start": 14,
  "e2_end": 16,
  "relation": 3
}

Data Fields

kbp37

id: the instance id of this sentence, a string feature.
sentence: the sentence, a string features.
relation: the relation label of this instance, an int classification label.

{"no_relation": 0, "org:alternate_names(e1,e2)": 1, "org:alternate_names(e2,e1)": 2, "org:city_of_headquarters(e1,e2)": 3, "org:city_of_headquarters(e2,e1)": 4, "org:country_of_headquarters(e1,e2)": 5, "org:country_of_headquarters(e2,e1)": 6, "org:founded(e1,e2)": 7, "org:founded(e2,e1)": 8, "org:founded_by(e1,e2)": 9, "org:founded_by(e2,e1)": 10, "org:members(e1,e2)": 11, "org:members(e2,e1)": 12, "org:stateorprovince_of_headquarters(e1,e2)": 13, "org:stateorprovince_of_headquarters(e2,e1)": 14, "org:subsidiaries(e1,e2)": 15, "org:subsidiaries(e2,e1)": 16, "org:top_members/employees(e1,e2)": 17, "org:top_members/employees(e2,e1)": 18, "per:alternate_names(e1,e2)": 19, "per:alternate_names(e2,e1)": 20, "per:cities_of_residence(e1,e2)": 21, "per:cities_of_residence(e2,e1)": 22, "per:countries_of_residence(e1,e2)": 23, "per:countries_of_residence(e2,e1)": 24, "per:country_of_birth(e1,e2)": 25, "per:country_of_birth(e2,e1)": 26, "per:employee_of(e1,e2)": 27, "per:employee_of(e2,e1)": 28, "per:origin(e1,e2)": 29, "per:origin(e2,e1)": 30, "per:spouse(e1,e2)": 31, "per:spouse(e2,e1)": 32, "per:stateorprovinces_of_residence(e1,e2)": 33, "per:stateorprovinces_of_residence(e2,e1)": 34, "per:title(e1,e2)": 35, "per:title(e2,e1)": 36}

kbp37_formatted

id: the instance id of this sentence, a string feature.
token: the list of tokens of this sentence, using str.split(), a list of string features.
e1_start: the 0-based index of the start token of the first argument', an int feature.
e1_end: the 0-based index of the end token of the first argument, exclusive, an int feature.
e2_start: the 0-based index of the start token of the second argument, an int feature.
e2_end: the 0-based index of the end token of the second argument, exclusive, an int feature.
relation: the relation label of this instance, an int classification label (same as 'kbp37'').

Data Splits

	Train	Dev	Test
kbp37	15917	1724	3405
kbp37_formatted	15807	1714	3379

Dataset Creation

Curation Rationale

More Information Needed

Source Data

Initial Data Collection and Normalization

More Information Needed

Who are the source language producers?

More Information Needed

Annotations

Considerations for Using the Data

Social Impact of Dataset

More Information Needed

Discussion of Biases

More Information Needed

Other Known Limitations

More Information Needed

Additional Information

Dataset Curators

More Information Needed

Licensing Information

More Information Needed

Citation Information

@article{DBLP:journals/corr/ZhangW15a,
  author    = {Dongxu Zhang and
               Dong Wang},
  title     = {Relation Classification via Recurrent Neural Network},
  journal   = {CoRR},
  volume    = {abs/1508.01006},
  year      = {2015},
  url       = {http://arxiv.org/abs/1508.01006},
  eprinttype = {arXiv},
  eprint    = {1508.01006},
  timestamp = {Fri, 04 Nov 2022 18:37:50 +0100},
  biburl    = {https://dblp.org/rec/journals/corr/ZhangW15a.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}