asahi417's picture
fix readme
806b513
metadata
language:
  - en
license:
  - other
multilinguality:
  - monolingual
size_categories:
  - n<1K
pretty_name: SemEval2012 relational similarity dataset

Dataset Card for "relbert/semeval2012_relational_similarity"

Dataset Description

Dataset Summary

Relational similarity dataset from SemEval2012 task 2, compiled to fine-tune RelBERT model. The dataset contains a list of positive and negative word pair from 89 pre-defined relations. The relation types are constructed on top of following 10 parent relation types.

{
    1: "Class Inclusion",  # Hypernym
    2: "Part-Whole",  # Meronym, Substance Meronym
    3: "Similar",  # Synonym, Co-hypornym
    4: "Contrast",  # Antonym
    5: "Attribute",  # Attribute, Event
    6: "Non Attribute",
    7: "Case Relation",
    8: "Cause-Purpose",
    9: "Space-Time",
    10: "Representation"
}

Each of the parent relation is further grouped into child relation types where the definition can be found here.

Dataset Structure

Data Instances

An example of train looks as follows.

{
  'relation_type': '8d',
  'positives': [ [ "breathe", "live" ], [ "study", "learn" ], [ "speak", "communicate" ], ... ]
  'negatives': [ [ "starving", "hungry" ], [ "clean", "bathe" ], [ "hungry", "starving" ], ... ] 
}

Data Splits

train validation
79 79

Citation Information

@inproceedings{jurgens-etal-2012-semeval,
    title = "{S}em{E}val-2012 Task 2: Measuring Degrees of Relational Similarity",
    author = "Jurgens, David  and
      Mohammad, Saif  and
      Turney, Peter  and
      Holyoak, Keith",
    booktitle = "*{SEM} 2012: The First Joint Conference on Lexical and Computational Semantics {--} Volume 1: Proceedings of the main conference and the shared task, and Volume 2: Proceedings of the Sixth International Workshop on Semantic Evaluation ({S}em{E}val 2012)",
    month = "7-8 " # jun,
    year = "2012",
    address = "Montr{\'e}al, Canada",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/S12-1047",
    pages = "356--364",
}