e2e_cleaned

  • Descrição :

Uma versão atualizada dos dados do E2E NLG Challenge com MRs limpos. Os dados E2E contêm representação de significado baseada em ato de diálogo (MR) no domínio do restaurante e até 5 referências em linguagem natural, que é o que é preciso prever.

Dividir Exemplos
'test' 4.693
'train' 33.525
'validation' 4.299
  • Estrutura de recursos :
FeaturesDict({
    'input_text': FeaturesDict({
        'table': Sequence({
            'column_header': string,
            'content': string,
            'row_number': int16,
        }),
    }),
    'target_text': string,
})
  • Documentação do recurso:
Característica Classe Forma Tipo D Descrição
RecursosDict
Entrada de texto RecursosDict
texto_entrada/tabela Seqüência
input_text/table/column_header tensor corda
texto_entrada/tabela/conteúdo tensor corda
input_text/table/row_number tensor int16
texto_alvo tensor corda
  • Citação :
@inproceedings{dusek-etal-2019-semantic,
    title = "Semantic Noise Matters for Neural Natural Language Generation",
    author = "Du{\v{s} }ek, Ond{\v{r} }ej  and
      Howcroft, David M.  and
      Rieser, Verena",
    booktitle = "Proceedings of the 12th International Conference on Natural Language Generation",
    month = oct # "{--}" # nov,
    year = "2019",
    address = "Tokyo, Japan",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/W19-8652",
    doi = "10.18653/v1/W19-8652",
    pages = "421--426",
    abstract = "Neural natural language generation (NNLG) systems are known for their pathological outputs, i.e. generating text which is unrelated to the input specification. In this paper, we show the impact of semantic noise on state-of-the-art NNLG models which implement different semantic control mechanisms. We find that cleaned data can improve semantic correctness by up to 97{\%}, while maintaining fluency. We also find that the most common error is omitting information, rather than hallucination.",
}