The Imagination of Crowds: Conversational AAC Language Modeling using Crowdsourcing and Large Data Sources

Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 700–711, Edinburgh, Scotland, UK, July 27–31, 2011. c⃝2011 Association for Computational Linguistics The Imagination of Crowds: Conversational AAC Language Modeling using Crowdsourcing and Large Data Sources Keith Vertanen Department of Computer Science Princeton University vertanen@princeton.edu Per Ola Kristensson School of Computer Science University of St Andrews pok@st-andrews.ac.uk Abstract Augmented and alternative communication (AAC) devices enable users with certain com- munication disabilities to participate in every- day conversations. Such devices often rely on statistical language models to improve text entry by offering word predictions. These predictions can be improved if the language model is trained on data that closely reflects the style of the users’ intended communica- tions. Unfortunately, there is no large dataset consisting of genuine AAC messages. In this paper we demonstrate how we can crowd- source the creation of a large set of fictional AAC messages. We show that these messages model conversational AAC better than the cur- rently used datasets based on telephone con- versations or newswire text. We leverage our crowdsourced messages to intelligently select sentences from much larger sets of Twitter, blog and Usenet data. Compared to a model trained only on telephone transcripts, our best performing model reduced perplexity on three test sets of AAC-like communications by 60– 82% relative. This translated to a potential keystroke savings in a predictive keyboard in- terface of 5–11%. 1 Introduction Users with certain communication disabilities rely on augmented and alternative communication (AAC) devices to take part in everyday conversa- tions. Often these devices consist of a predictive text input method coupled with text-to-speech out- put. Unfortunately, the text entry rates provided by AAC devices are typically low, between 0.5 and 16 words-per-minute (Trnka et al., 2009). As a consequence, researchers have made nu- merous efforts to increase AAC text entry rates by employing a variety of improved language model- ing techniques. Examples of approaches include adapting the language model to recently used words (Wandmacher et al., 2008; Trnka, 2008), using syn- tactic information (Hunnicutt, 1989; Garay-Vitoria and Gonz´alez-Abascal, 1997), using semantic in- formation (Wandmacher and Antoine, 2007; Li and Hirst, 2005), and modeling topics (Lesher and Rinkus, 2002; Trnka et al., 2006). For a recent sur- vey, see Garay-Vitoria and Abascal (2006). While such language model improvement tech- niques are undoubtedly helpful, certainly they can all benefit from starting with a long-span language model trained on large amounts of closely matched data. For AAC devices this means closely modeling everyday face-to-face communications. However, a long-standing problem in the field is the lack of good data sources that adequately model such AAC communications. Due to privacy-reasons and other ethical concerns, there is no large dataset consist- ing of genuine AAC messages. Therefore, previous research has used transcripts of telephone conversa- tions or newswire text. However, these data sources are unlikely to be an ideal basis for AAC language models. In this paper we show that it is possible to signif- icantly improve conversational AAC language …

Content truncated to 500 words