NASCA / README.md
José Ángel González
Update README.md
2eafb43
|
raw
history blame
2.31 kB
metadata
language: ca
tags:
  - summarization
widget:
  - text: >-
      The tower is 324 metres (1,063 ft) tall, about the same height as an
      81-storey building, and the tallest structure in Paris. Its base is
      square, measuring 125 metres (410 ft) on each side. During its
      construction, the Eiffel Tower surpassed the Washington Monument to become
      the tallest man-made structure in the world, a title it held for 41 years
      until the Chrysler Building in New York City was finished in 1930. It was
      the first structure to reach a height of 300 metres. Due to the addition
      of a broadcasting aerial at the top of the tower in 1957, it is now taller
      than the Chrysler Building by 5.2 metres (17 ft). Excluding transmitters,
      the Eiffel Tower is the second tallest free-standing structure in France
      after the Millau Viaduct.

News Abstractive Summarization for Catalan (NASCA) is a Transformer encoder-decoder model, with the same hyper-parameters than BART, to perform summarization of Catalan news articles. It is pre-trained on a combination of several self-supervised tasks that help to increase the abstractivity of the generated summaries. Four pre-training tasks have been combined: sentence permutation, text infilling, Gap Sentence Generation, and Next Segment Generation. Catalan newspapers, the Catalan subset of the OSCAR corpus and Wikipedia articles in Catalan were used for pre-training the model (9.3GB of raw text -2.5 millions of documents-).

NASCA is finetuned for the summarization task on 636.596 (document, summary) pairs from the Dataset for Automatic summarization of Catalan and Spanish newspaper Articles (DACSA).

More details about the pretraining/finetuning datasets and the models soon:

@unpublished{DACSA, author = "Vicent Ahuir, Lluís-F. Hurtado , José Ángel González and Encarna Segarra", title = "DACSA: a Dataset for Automatic summarization of Catalan and Spanish newspaper Articles", note = "Unsubmitted", }

@unpublished{NAS, author = "Vicent Ahuir, Lluís-F. Hurtado , José Ángel González and Encarna Segarra", title = "NASCA and NASES : Two monolingual pre-trained models for abstractive summarization in Catalan and Spanish", note = "Submitted to the Special Issue on Current Approaches and Applications in Natural Language Processing (Applied Sciences)", }