Sbir phase II: evaluative artificial speech intelligence and autism screener (EASI-as)

This Cooperative Agreement award from the National Science Foundation's Technology, Innovation, and Partnerships program (CFDA 47.084) provides $1 million to Itherapy, LLC of Martinez, California to develop an evaluative artificial speech intelligence and autism screener system called EASI-AS. The system aims to improve treatment for children with communication impairments by providing specialists and parents with accurate, efficient reports on speech and language evaluations along with recommendations. Key objectives include improving machine learning modules for language proficiency assessment, recommendation generation, child voice and accent recognition, and intonation detection. The award also supports developing an autism screener using eye tracking technology and a web/mobile platform for data collection from children. The two-year project period runs from May 1, 2022 to April 30, 2024.

Abstract

The broader impact of this small business innovation research (SBIR) phase II project will improve treatment for children with communication impairments. The project will develop a system that provides specialists and parents with reports that have accurate, efficient, and easy-to-understand speech and language evaluations with recommendations and suggestions, for better understanding the children?S communication challenges and how to solve them. IT will bridge the GAP between a clinical assessment and support at home, impacting clinical outcomes by providing a more accurate assessment of child language developmental needs. Consequently, this will result in more personalized education and treatment, particularly for diverse communities. This small business innovation research (SBIR) phase II project addresses pain points surrounding speech and language assessment with children with communication impairments. The child?S language use must be captured in a natural setting to obtain an accurate analysis. The solution is a flexible, self-customizing system using a trained AI approach to evaluate speech and language skills, then provide analytical feedback in key communication areas. To evaluate a child?S communication skills, the system collects a language sample based on natural adult-child social language interactions. Research objectives include: (1) improvement and development of machine-learning modules, specifically the language proficiency assessment, recommendation engine, recognition of child voices, recognition of accents and dialects, and intonation detection; (2) an autism screener development, to be used as a pre-diagnostic tool with eye-tracking technology; and (3) web and mobile application development for data collection by the child. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.