doi: 10.4304/jsw.2.4.12-23
New Functions for Unsupervised Asymmetrical Paraphrase Detection
- 1 HULTIG, University of Beira Intetrior, Covilh˜a, Portugal
- 2 LIACC, University of Porto, Porto, Portugal
Abstract
Monolingual text-to-text generation is an emergingresearch area in Natural Language Processing. Onereason for the interest in such generation systems is thepossibility to automatically learn text-to-text generationstrategies from aligned monolingual corpora. In this context,paraphrase detection can be seen as the task of aligningsentences that convey the same information but yet arewritten in different forms, thereby building a training setof rewriting examples. In this paper, we propose a newtype of mathematical functions for unsupervised detectionof paraphrases, and test it over a set of standard paraphrasecorpora. The results are promising as they outperform stateof-the-art functions developed for similar tasks. We considertwo types of paraphrases - symmetrical and asymmetricalentailed - and show that although our proposed functionswere conceived and oriented toward the asymmetrical detection,they perform rather well for symmetrical sentencepairs identification.
Keywords
- Paraphrasing
- Paraphrase Identification
How to Cite
Cordeiro Jo˜ao, Dias Ga¨el, and Brazdil Pavel, "New Functions for Unsupervised Asymmetrical Paraphrase Detection," Journal of Software, vol. 2, no. 4, pp. 12-23, 2007. https://doi.org/ 10.4304/jsw.2.4.12-23
Copyright & License
Copyright © 2007 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).