Generating Syntactic Paraphrases - Archive ouverte HAL Access content directly
Conference Papers Year : 2018

Generating Syntactic Paraphrases

Abstract

We study the automatic generation of syntactic paraphrases using four different models for generation: data-to-text generation, textto-text generation, text reduction and text expansion, We derive training data for each of these tasks from the WebNLG dataset and we show (i) that conditioning generation on syntactic constraints effectively permits the generation of syntactically distinct paraphrases for the same input and (ii) that exploiting different types of input (data, text or data+text) further increases the number of distinct paraphrases that can be generated for a given input.
Fichier principal
Vignette du fichier
D18-1113.pdf (265.3 Ko) Télécharger le fichier
Origin Explicit agreement for this submission

Dates and versions

hal-03046028 , version 1 (08-12-2020)

Identifiers

Cite

Émilie Colin, Claire Gardent. Generating Syntactic Paraphrases. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Oct 2018, Brussels, Belgium. pp.937-943, ⟨10.18653/v1/D18-1113⟩. ⟨hal-03046028⟩
53 View
61 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More