PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Relatedness curves for acquiring paraphrases
Georgiana Dinu and Grzegorz Chrupala
In: ACL workshop GEMS 2010, Uppsala, Sweden(2010).

Abstract

In this paper we investigate methods for computing similarity of two phrases based on their relatedness scores across all ranks k in a SVD approximation of a phrase/term co-occurrence matrix. We confirm the major observations made in previous work and our preliminary experiments indicate that these methods can lead to reliable similarity scores which in turn can be used for the task of paraphrasing.

EPrint Type:Conference or Workshop Item (Talk)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Natural Language Processing
ID Code:8851
Deposited By:Grzegorz Chrupala
Deposited On:21 February 2012