arXiv:2606.31722v1 Announce Type: new Abstract: Automatic speech recognition (ASR) systems often perform poorly in dysarthric speech, limiting their usefulness to affected speakers in everyday communication. This paper presents a personalized ASR system for a dysarthric speaker, built by adapting a foundation ASR model to speaker-specific data. Using the TEQST tool, we collected 92 hours of read speech and later added 8.8 hours of user corrections gathered through a deployed mobile application. Starting from Whisper, fine-tuning reduced word error rate to 15.8% with only 1.4 hours of adaptatio
Advances in foundation models and fine-tuning techniques are making it possible to address long-standing challenges in accessibility for specialized populations.
This development demonstrates a practical application of AI in improving quality of life for individuals with disabilities, expanding the utility of ASR systems beyond general populations.
The ability to personalize ASR with relatively small datasets of speaker-specific data makes AI-powered communication tools more accessible and effective for dysarthric individuals.
- · Dysarthric individuals
- · Assistive technology companies
- · Healthcare providers
- · AI model developers
- · ASR systems lacking adaptability
- · Generalized communication solutions
Improved communication for individuals with dysarthria, leading to greater independence and social inclusion.
Increased demand for personalized AI solutions in healthcare and accessibility, spurring further research and product development.
The development of more inclusive digital interfaces and broader societal benefits from AI tailored to diverse user needs.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL