ギガワード

  • 説明:

約 400 万の記事からなる Gigaword の記事ペアのコーパスでの見出し生成。 https://github.com/microsoft/unilm/が提供する 'org_data' を使用します。これはhttps://github.com/harvardnlp/sent-summaryと同じですが、形式がより優れています。

2 つの機能があります。 - ドキュメント: 記事。 - 要約: 見出し。

スプリット
'test' 1,951人
'train' 3,803,957
'validation' 189,651
  • 機能構造:
FeaturesDict({
    'document': Text(shape=(), dtype=string),
    'summary': Text(shape=(), dtype=string),
})
  • 機能のドキュメント:
特徴クラスDtype説明
特徴辞書
資料文章ストリング
まとめ文章ストリング
  • 引用
@article{graff2003english,
  title={English gigaword},
  author={Graff, David and Kong, Junbo and Chen, Ke and Maeda, Kazuaki},
  journal={Linguistic Data Consortium, Philadelphia},
  volume={4},
  number={1},
  pages={34},
  year={2003}
}

@article{Rush_2015,
   title={A Neural Attention Model for Abstractive Sentence Summarization},
   url={http://dx.doi.org/10.18653/v1/D15-1044},
   DOI={10.18653/v1/d15-1044},
   journal={Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing},
   publisher={Association for Computational Linguistics},
   author={Rush, Alexander M. and Chopra, Sumit and Weston, Jason},
   year={2015}
}