Learning a POS tagger for AAVE-like language

By Anna Katrine van Zee, Dirk Hovy, Anders Søgaard. 2016

47 citations recorded by the source corpus.

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 automatically and partially label using mined tag dictionaries. Our POS tagger obtains a tagging accuracy of 89% on subtitles, 85% on lyrics, and 83% on tweets, with up to 55% error reductions over a state-of-the-art newswire POS tagger, and 15-25% error reductions over a state-of-the-art Twitter POS tagger.

Subfields: Artificial Intelligence, Language and Linguistics, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Language and Linguistics

In Natural Language Processing

Citations in the corpus

Language (Technology) is Power: A Critical Survey of “Bias” in NLP (2020, background): “…identification systems, and dependency parsers all work less well on text containing features associated with AAE than on text without these features (Jørgensen et al., 2015, 2016; Blodgett et al., 2016, 2018), and that toxicity detection systems score tweets containing features associated with…”
Repairing the Cracked Foundation: A Survey of Obstacles in Evaluation Practices for Generated Text (2022, background): “…using large web-scraped datasets, and to Luccioni and Viviano (2021) and Dodge et al. (2021) for analyses of such large-scale web-scraped corpora and their representational, legal, consent, and PII issues. text with AAVE features (Jørgensen et al., 2015, 2016; Blodgett et al., 2016, among others).”
Demographic Dialectal Variation in Social Media: A Case Study of African-American English (2016, background): “In fact, its presence in social media is at-tracting increasing interest for natural language processing (Jørgensen et al., 2016) and sociolinguistic (Stewart, 2014; Eisenstein, 2015; Jones, 2015) research.”
Dats Wassup!!: Investigating African-American Vernacular English in Transformer-Based Text Generation (2020, methodology): “Studies regarding AAVE have analyzed tasks such as POS tagging (Jørgensen et al., 2016), detecting AAVE syntax (Stewart, 2014), voice recognition and transcription (Dorn, 2019), dependency parsing (Blodgett et al., 2016), and hate speech detection (Sap et al., 2019), but not language generation.”
Racial Disparity in Natural Language Processing: A Case Study of Social Media African-American English (2017, positive): “own in the sociolinguistics literature to be characteristic of AAE, in line with other work that has also veri•ed the correspondence of geographical AA prevalence to AAE linguistic features on Twi−er [10, 19]. „is publicly available corpus contains 59.2 million tweets. We •lter its messages to ones ”
Twitter Universal Dependency Parsing for African-American and Mainstream American English (2018, background): “…parser with morphological information stripped and its tags predicted from the ARK Twitter POS tagger (Owoputi et al., 2013), which is both tailored for Twitter and displays a smaller AAE vs MAE performance gap than traditional taggers (Jørgensen et al., 2016); we call this the ARK Tagger setting.”
The Importance of Modeling Social Factors of Language: Theory and Practice (2021, background): “Dialectal areas still play an important role, even in online communication (Hovy and Purschke, 2018), and identifying and integrating them can be vital for fairer NLP tools (Jørgensen et al., 2016; Blodgett et al., 2016; Dorn, 2019).”
Investigating Sports Commentator Bias within a Large Corpus of American Football Broadcasts (2019, methodology): “subtitles (Jørgensen et al., 2016).”
Darling or Babygirl ? Investigating Stylistic Bias in Sentiment Analysis (2018, background): “African Americans use the word missin’ more frequently while White Americans use missing more frequently (Jørgensen et al., 2016)).”
Exploring Social Bias in Chatbots using Stereotype Knowledge (2019, background): “Recently, there have been many works (Bolukbasi et al., 2016; Dixon et al., 2017; Jørgensen et al., 2016) that illustrate the vulnerability of data-driven NLP systems in unintentionally learning biases inherent in datasets.”
Towards Robust and Privacy-preserving Text Representations (2018, methodology): “Recent demographic studies have found that the author’s gender, age and race can influence tagger performance (Hovy and Søgaard, 2015; Jørgensen et al., 2016).”