Project Leaders: Holyfield, MacNeil, Raley, Jakobs, Constituent Board


Problems Addressed: Accessing relevant vocabulary on existing AAC technology often requires significant learning, time and effort. Currently, AAC users are required to adapt to their system rather than their system adapting to them and their environment. It would be so much better if they could meet in the middle!
Goals of the project:
- Develop and evaluate prototypes that predict communication options based on communication partner speech, location, and co-location.
- Explore approaches to predict communication options based on other information as prioritized by AAC users.


How will we accomplish these goals?
- “Wizard of Oz” initial prototypes to observer interactions that mimic those that would occur with functioning technology. This prototypes use a “wizard” who offers options in place of AI to quickly test concepts.
- Develop required backend AI capabilities.
- Test the novel capabilities
- Collaborate with AAC users to iterate on prototypes and to explore other avenues for responsive prediction.
Updates
Additional Information