A State-of-the-Art Automatic Speech Recognition and Conversational Platform to Enable Socially Assistive Robots for Persons with Alzheimer's Disease and Related Dementias

点击翻译按钮获取中文摘要

ABSTRACT This commercialization readiness pilot (CRP) by care.coach corporation (Millbrae, CA) requests funds for late- stage research and development and technical assistance for a conversational technology driven by artificial intelligence (AI) that was developed with and for older adults, especially those aging with Alzheimer’s disease and related dementias (ADRD). Past work has resulted in a socially assistive robot (SAR) and virtual health assistant overseen by humans that improves health outcomes and reduces loneliness in older adults at a lower cost of care. Customer loyalty is world-class and far exceeds net promotor scores (NPS) for well-known conversational agents, attesting to the product’s commercial potential. Moreover, active Phase II funding (5R44AG0620-14-03) has resulted in an automated speech recognition (ASR) engine that is more accurate in understanding truly natural speech by older adults, including those with ADRD, than major conversional AI platforms. This is significant given the inherent challenges that the speech of older adults, especially of those with ADRD, poses to conversational AI due to normal age-related neurological changes or brain pathology associated with ADRD. Leveraging these advances and innovations, care.coach is preparing for rapid growth and has attracted considerable interest from insurers, corporate development arms of large insurers, and traditional investors. To successfully scale, we propose to re-platform our current system on top of an enterprise-grade, integrated knowledge solution enabled by KIE, an opensource platform powered by KIE’s Drools/jBPM rules and process engines. The KIE platform has revolutionized enterprise operations, and has been adopted in healthcare to drive quality and efficiency with applications ranging from the adjudication of health insurance claims to the implementation of clinical decision support systems. In addition to business process automation, the cost of goods sold (COGS) will come down with full conversational automation in allowable use cases by training and using a large language model. Both capabilities will be developed in the proposed CRP using state-of-the-art technical approaches, and will be evaluated for performance and acceptance in a replication study with ADRD patients in partnership with a health plan and medical providers that serve a large, diverse elderly population. This technical assistance and late-stage development will deliver a state-of-the-art, consumer-facing conversa- tional agent and virtual health assistant driven by a technology platform with industrial production and automa- tion methods that ensure consistent and scalable digital health service delivery and manufacturing, and that meet the highest security and consumer satisfaction standards. The overall goal is to improve the health and wellbeing of millions Americans within 5 years, including those aging with ADRD.