Learning a POS tagger for AAVE-like language

Proceedings of NAACL-HLT 2016, pages 1115–1120, San Diego, California, June 12-17, 2016. c⃝2016 Association for Computational Linguistics Learning a POS tagger for AAVE-like language∗ Anna Jørgensen University of Amsterdam Science Park 107 1098 XG Amsterdam, NL jorgensen@uva.nl Dirk Hovy University of Copenhagen Njalsgade 140 2300 Copenhagen S, DK dirk.hovy@hum.ku.dk Anders Søgaard University of Copenhagen Njalsgade 140 2300 Copenhagen S, DK soegaard@hum.ku.dk Abstract Part-of-speech (POS) taggers trained on newswire perform much worse on domains such as subtitles, lyrics, or tweets. In addition, these domains are also heterogeneous, e.g., with respect to registers and dialects. In this paper, we consider the problem of learning a POS tagger for subtitles, lyrics, and tweets associated with African-American Vernacular English (AAVE). We learn from a mixture of randomly sampled and manually annotated Twitter data and unlabeled data, which we au- tomatically and partially label using mined tag dictionaries. Our POS tagger obtains a tag- ging accuracy of 89% on subtitles, 85% on lyrics, and 83% on tweets, with up to 55% er- ror reductions over a state-of-the-art newswire POS tagger, and 15-25% error reductions over a state-of-the-art Twitter POS tagger. 1 Introduction Modern part-of-speech (POS) taggers perform well on what some consider canonical language, as found in domains such as newswire, for which sufficient manually-annotated data is available. For many do- mains, such as subtitles, lyrics, and tweets, however, labeled data is scarce, if existing, and the perfor- mance of off-the-shelf POS taggers is prohibitive of downstream applications. Furthermore, subtitles, lyrics and tweets are very heterogeneous. Subtitles span from Shakespeare to The Wire, and the lyrics of Elvis Costello are very different from those of Tupac Shakur. Twitter can ∗This work was supported by ERC Starting Grant No. 313695. be anything from teenagers discussing where to go tonight, to researchers discussed the implications of new findings. All three sources of data exhibit a very high degree of linguistic variation, some of which is due to the dialects of the speakers or authors. In this paper, we use a corpus of POS-annotated tweets recently released by CMU,1 consisting of semi-randomly sampled US tweets. We want to use this corpus to learn a POS tagger for subti- tles, lyrics, and tweets, which are typically associ- ated with African-American Vernacular English (AAVE). We believe our POS tagger can broaden the coverage of NLP tools, and serve as an impor- tant tool for large-scale sociolinguistic analyses of language use associated with AAVE (Jørgensen et al., 2015; Stewart, 2014), which relies on the accu- racy of these NLP tools. We combine several recent trends in domain adap- tation, namely word embeddings, clusters, sam- pling, and the use of type constraints. Word rep- resentations learned from representative unlabeled data, such as word clusters or embeddings, have been proven useful for increasing the accuracy of NLP tools for low-resource languages and do- mains (Owoputi et al., 2013; Aldarmaki and Diab, 2015; Gouws and Søgaard, 2015). Since similar words receive similar labels, this can give the model support for words not in the …

Content truncated to 500 words