Home

Awesome

WebNLG Dataset Summary

This repository presents the evolution of the WebNLG corpus.

Each folder contains the same data in two formats: xml and json.

  1. release_v2

    It is the latest release.

    It includes release_v1 and test data (seen categories) from the WebNLG challenge.

    We split it into train/dev/test, ensuring equal representation of DBpedia categories and tripleset sizes.

    Tree shapes and types (sibling, chain, mixed) were added for each input RDF tree.

  2. release_v2_constrained

    It has the same data as release_v2.

    The split into train/dev/test is more challenging. That split ensures that a triple occurring in train/dev is not present in test (more info in the INLG 2018 paper below).

  3. release_v1

    It matches Final Release (Larger Dataset) on the challenge website.

    It doesn't include test data (seen categories) from the challenge.

    No split into train/dev/test was provided.

    Covers 15 DBpedia categories.

  4. webnlg_challenge_2017

    Contains the data used in the WebNLG Challenge 2017.

    Covers 10 DBpedia categories (the City category only partially).

Documentation

http://webnlg.loria.fr/pages/docs.html

Publications

Citing

@InProceedings{gardent2017creating,
  author = 	"Gardent, Claire
		and Shimorina, Anastasia
		and Narayan, Shashi
		and Perez-Beltrachini, Laura",
  title = 	"Creating Training Corpora for NLG Micro-Planners",
  booktitle = 	"Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)    ",
  year = 	"2017",
  publisher = 	"Association for Computational Linguistics",
  pages = 	"179--188",
  location = 	"Vancouver, Canada",
  doi = 	"10.18653/v1/P17-1017",
  url = 	"http://www.aclweb.org/anthology/P17-1017"
}

@InProceedings{shimorina2018handling,
  author = 	"Shimorina, Anastasia
		and Gardent, Claire",
  title = 	"Handling Rare Items in Data-to-Text Generation",
  booktitle = 	"Proceedings of the 11th International Conference on Natural Language Generation",
  year = 	"2018",
  publisher = 	"Association for Computational Linguistics",
  location = 	"Tilburg, The Netherlands"
}

License

CC BY-NC-SA 4.0

Contact