Output details
11 - Computer Science and Informatics
University of Edinburgh
A Bayesian framework for word segmentation: Exploring the effects of context
<22> Originality: Models presented in this paper pioneered the use of nonparametric Bayesian methods for unsupervised learning of linguistic structure.
Significance: Becoming a standard reference for work on word segmentation (Daland and Pierrehumbert 2011, Hewlett 2011). Both new models (Neubig et al. 2010, Jones et al. 2010) and new inference methods (Pearl et al. 2010, Liang et al. 2010, Mochihashi et al. 2009) are based on this work. Goldwater received an EPSRC grant to extend the models.
Rigour: Cited in Charniak's 2011 ACL Lifetime Achievement Award speech as an example of the kind of research the community should strive for.