Intelligent assistive communication and the web as a social medium

Augmentative and alternative communication (AAC) describes strategies and technologies commonly used by people for whom spoken interaction is extremely difficult or infeasible. AAC includes facial expressions, sign languages, and picture cards, as well as high-tech devices. Most AAC systems present users with hierarchical arrays of symbols that are sequentially selected to construct messages. Although speed of message construction and ease of use are critical, these systems are often slow and require fatiguing physical movements. Our work challenges three main assumptions common to AAC that influence design decisions and place increased demands on the user rather than the system. We leverage natural language processing and machine learning to design intelligent communication interfaces that shift the cognitive and physical burden from the user to the system to allow for faster, less fatiguing communication. This work has implications for both face-to-face and Web-based communication. AAC users are increasingly turning to the Web for social interaction, and W4A provides a unique opportunity to explore how to leverage intelligent interfaces to lower the barriers to active participation in the World Wide Web.