Presentation 1: Integrating LLMs and Other AI Technologies in an Intelligent Cognitive Assistant for Word Retrieval Support for Older Adults with ADRD
Speaker: Archna Bhatia, PhD, Research Scientist, Institute for Human and Machine Cognition
Abstract: Occasional inability to retrieve words accompanies aging, even for healthy people, and especially for those with Alzheimer's Disease and Related Dementias (ADRD). This phenomenon tends to thrust people into social isolation, resulting in an increased rate of decline. In this webinar, we will discuss the development of an Intelligent Personalized Cognitive Assistant, MyWords!, an AI-enhanced application that provides real-time word retrieval support and personalized training for individuals with ADRD. The development of the MyWords! application, as part of a pilot project, involved participatory design and usability testing to develop the application collaboratively, informing the user interface as well as the processing components of the application. The natural language processing components currently involve technologies, including keyword matching, graph matching, parsing, and generative LLMs to handle diverse linguistic challenges. In this session, we will present the application interface and processing components, discuss our efforts, challenges, and solutions in involving LLMs in the application, and provide insights from ongoing application testing for future developments.
Presentation 2: Generating Compassion
Speaker: Richard Curtis, VP Product Development, Rippl Care
Abstract: Join us for an insightful webinar on the integration of advanced AI technologies into the Kinto platform, a support service for family caregivers of individuals living with Alzheimer's Disease and Related Dementias (ADRD). This talk will delve into the findings from a recent pilot project aimed at enhancing caregiver support through data-driven insights and personalized recommendations. We will discuss the project's objectives: data extraction and anonymization, strain assessment, and resource recommendation using the GPT-4 language model. Our initial attempts with large language models for data anonymization revealed significant challenges, leading to the adoption of off-the-shelf tools for effective de-identification. The webinar will cover the development of AI-driven strain assessments and a personalized resource recommendation system, highlighting the feasibility and areas for improvement identified through subjective evaluations by Kinto coaches. Join us to explore how AI technologies are transforming caregiver support services and providing a foundation for future advancements in the caregiving sector. Kinto was recently acquired by Rippl Care, where we look forward to building on this foundational work. This webinar offers valuable insights for researchers, practitioners, and anyone interested in the intersection of AI and caregiving.
Dr. Archna Bhatia: She is a Research Scientist at the Institute for Human and Machine Cognition. Her research is focused on (i) speech and natural language processing for health and other applications such as cybersecurity, information extraction and cognitive modeling, and (ii) modeling linguistic phenomena to improve semantic parsing and natural language understanding.
Rich Curtis: He is an experienced software product manager who has helped guide successful products at a wide range of companies, from computer game development, mobile consumer apps, and large, enterprise software projects. At Kinto, he led product development efforts, updating the platform to be more flexible, scalable, and smarter through the application of generative AI. He now works as the VP of Product Development at Rippl Care.