Towards improving predictive AAC using crowdsourced dialogues and partner context
Document Type
Conference Proceeding
Publication Date
10-2017
Department
Department of Computer Science; Center for Human-Centered Computing
Publication Title
ASSETS '17 Proceedings of the 19th International ACM SIGACCESS Conference on Computers and Accessibility
Recommended Citation
Vertanen, K.
(2017).
Towards improving predictive AAC using crowdsourced dialogues and partner context.
ASSETS '17 Proceedings of the 19th International ACM SIGACCESS Conference on Computers and Accessibility.
http://doi.org/10.1145/3132525.3134814
Retrieved from: https://digitalcommons.mtu.edu/michigantech-p/1086
Publisher's Statement
Augmentative and Alternative Communication (AAC) devices typically rely on a language model to help make predictions or disambiguate user input. We investigate how to improve predictions in two-sided conversational dialogues. We collect and share a new corpus of crowdsourced everyday dialogues. We show how language models based on recurrent neural networks outperform N-gram models on these dialogues. We demonstrate further gains are possible using text obtained from an AAC user's communication partner, even when that text is partial or contains errors.