Big Tech joins forces to improve speech recognition technology for those with disabilities
The Speech Accessibility Project will utilize machine learning and artificial intelligence to expand the range of speech patterns that speech recognition systems can recognize. Microsoft, Meta, Apple, Google, and Amazon are all involved.
The Speech Accessibility Project is a collaboration between the University of Illinois at Urbana-Champaign (UIUC), Amazon, Apple, Google, Meta, Microsoft, and non-profit partners. The project's goal is to enhance speech recognition technology so that everyone, including those with disabilities and speakers with a variety of speech patterns, can benefit from it.
According to Mark Hasegawa-Johnson, an electrical and computer engineering professor at the University of Illinois at Chicago who is leading the project, "the ability to communicate and operate things using speech is vital for anyone working with technology or the digital economy today." Everyone, including those with disabilities, should have access to speech interfaces.
Currently, a variety of speech patterns, including those frequently connected with accents, are not recognized by many speech recognition systems, such as voice assistants and translation tools.
As a result, members of these groups frequently are unable to benefit from what speech recognition software has to offer. The Speech Accessibility Project's objective is to help technology firms increase the inclusivity of their voice recognition software through the use of artificial intelligence and machine learning.
We've assembled a particularly interdisciplinary team with expertise in linguistics, speech, AI, security, and privacy to help us meet this significant challenge because this task requires a lot of infrastructure, ideally the kind that can be supported by leading technology companies, said Hasegawa-Johnson.
By enlisting compensated volunteers to provide recorded samples, the Speech Accessibility Project will gather speech samples from persons reflecting a variety of speech patterns.
No comments: