The ability to communicate and operate devices through speech is crucial for anyone interacting with technology today.. However, these technologies do not benefit the millions of people around the world who have various disabilities.
The University of Illinois at Urbana-Champaign (UIUC) in the United States launched on Monday October 3 the “Speech Accessibility Project”, a multi-year research initiative with the support of Amazon, Apple, Google, Meta, Microsoft and a few not-for-profit disabled people’s organizations.
The new research initiative aims to make speech recognition technology more accessible for people with speech impairments and various disabilities. This includes, in particular, disabilities affecting speech, such as amyotrophic lateral sclerosis, Parkinson’s disease, cerebral palsy and Down syndrome (trisomy 21).
“Voice interfaces should be available to everyone, including people disabilities. This task has been difficult because it requires a lot of infrastructure, ideally the type that can be supported by major technology companies. We have therefore created a unique interdisciplinary team, experts in linguistics, speech, AI, security and privacy, to help us meet this important challenge.”said Mark Hasegawa-Johnson, professor of electrical and computer engineering at UIUC.
Over the past decade, Apple and other tech companies have been breaking new ground in the voice assistant space with tools like Siri, Amazon Alexa, Google Assistant, and more. Apple has also invested in technologies such as VoiceOver and Voice Control. More recently in 2019, Google launched Project Euphonia which aimed to make voice recognition technology more accessible by leveraging Artificial Intelligence. These projects have each experienced flaws, hence the importance of collaborating to develop automatic speech recognition systems capable of understanding everyone’s voice, regardless of the vocal model.
To achieve this goal, this speech accessibility project will collect speech samples from people representing a diversity of speech patterns. UIUC researchers will recruit paid volunteers to provide recorded voice samples and create a private, de-identified dataset that can be used to train machine learning models to better understand a wide variety of speech patterns.
The Speech Accessibility Project will initially focus on American English. According to the National Institute on Deafness and Other Communication Disorders (NIDCD), the results of this work could benefit more than 17.5 million people with these conditions in the United States alone.
Read also: South Africa: Google will establish a data center in Cape Town in 2023 to compete with Amazon