Minimally-Supervised Morphological Segmentation using Adaptor Grammars

Transactions of the Association for Computational Linguistics, 1 (2013) 255–266. Action Editor: Kristina Toutanova. Submitted 11/2012; Published 5/2013. c⃝2013 Association for Computational Linguistics. Minimally-Supervised Morphological Segmentation using Adaptor Grammars Kairit Sirts Institute of Cybernetics Tallinn University of Technology sirts@phon.ioc.ee Sharon Goldwater School of Informatics The University of Edinburgh sgwater@inf.ed.ac.uk Abstract This paper explores the use of Adaptor Gram- mars, a nonparametric Bayesian modelling framework, for minimally supervised morpho- logical segmentation. We compare three train- ing methods: unsupervised training, semi- supervised training, and a novel model selec- tion method. In the model selection method, we train unsupervised Adaptor Grammars us- ing an over-articulated metagrammar, then use a small labelled data set to select which poten- tial morph boundaries identified by the meta- grammar should be returned in the final output. We evaluate on five languages and show that semi-supervised training provides a boost over unsupervised training, while the model selec- tion method yields the best average results over all languages and is competitive with state-of- the-art semi-supervised systems. Moreover, this method provides the potential to tune per- formance according to different evaluation met- rics or downstream tasks. 1 Introduction Research into unsupervised learning of morphology has a long history, starting with the work of Harris (1951). While early research was mostly motivated by linguistic interests, more recent work in NLP often aims to reduce data sparsity in morphologically rich languages for tasks such as automatic speech recogni- tion, statistical machine translation, or automatic text generation. For these applications, however, com- pletely unsupervised systems may not be ideal if even a small amount of segmented training data is available. In this paper, we explore the use of Adap- tor Grammars (Johnson et al., 2007) for minimally supervised morphological segmentation. Adaptor Grammars (AGs) are a nonparametric Bayesian modelling framework that can learn latent tree structures over an input corpus of strings. For example, they can be used to define a morpholog- ical grammar where each word consists of zero or more prefixes, a stem, and zero or more suffixes; the actual forms of these morphs (and the segmentation of words into morphs) are learned from the data. In this general approach AGs are similar to many other unsupervised morphological segmentation systems, such as Linguistica (Goldsmith, 2001) and the Mor- fessor family (Creutz and Lagus, 2007). A major difference, however, is that the morphological gram- mar is specified as an input to the program, rather than hard-coded, which allows different grammars to be explored easily. For the task of segmenting utterances into words, for example, Johnson and col- leagues have experimented with grammars encoding different kinds of sub-word and super-word structure (e.g., syllables and collocations), showing that the best grammars far outperform other systems on the same corpora (Johnson, 2008a; Johnson and Goldwa- ter, 2009; Johnson and Demuth, 2010). These word segmentation papers demonstrated both the power of the AG approach and the syner- gistic behavior that occurs when learning multiple levels of structure simultaneously. However, the best- performing grammars were selected using the …

Content truncated to 500 words