Abraham Glasser
Gallaudet University, DC·Computer Science
Asking a smart speaker for the weather assumes you speak to it. Many Deaf people communicate primarily in American Sign Language (ASL) instead, and no mainstream digital assistant accepts sign.
- Fields
Computer Science
Psychology
- Topics
American Sign Language
Deaf users
accessibility
- Grants
- 1 known grant, through May 2027
What they work on
- Test whether a digital assistant that recognizes only single signs, using about 100 signs serving as stand-ins for common voice queries, can meet Deaf users' needs in Wizard-of-Oz experiments.
- Measure Deaf ASL users' satisfaction when querying a simulated assistant with natural ASL phrases drawn from a larger but still limited vocabulary and receiving partially correct responses.
- Vary the simulated sign recognition accuracy of an assistant to establish how accurate ASL understanding must be, and which error types are tolerable, for users to remain satisfied.
More about the research
Automatic recognition of unconstrained signing remains out of reach, which raises a question that can be answered now: how much signing does an assistant actually have to understand before Deaf users find it worth using?
Why it matters
Deaf people who sign are effectively excluded from voice-driven digital assistants that other users rely on daily. Evidence about workable vocabulary sizes and tolerable error rates could let designers ship sign-based assistants well before unconstrained ASL recognition is solved, lowering an everyday accessibility barrier and giving future recognition research user-grounded targets.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Malihe Alikhani
Northeastern University, MA·Computer Science
Build an AI tool that recognizes ASL signs used in classroom conversation and shows related scientific terms in real time through augmented reality, signing avatars and English captions.
Hae-Na Lee
Michigan State University, MI·Computer Science
Develop detection algorithms that analyze webpage layouts, interaction workflows, user actions, and accessible navigation paths to identify deceptive design patterns, including ones created by mismatches between a page's visual structure and its screen reader reading order.
Xiaojun Bi
Stony Brook University, NY·Computer Science
Create a non-visual free-form gesture input method that lets blind users enter whole words or phrases by tracing continuous gestures on the standard Qwerty touchscreen keyboard, removing the need to confirm each letter.
The school
Gallaudet University: 2 labs on PI Pages, $10M a year on research, 12 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.