A sample of current projects:
- Improving Transportation Safety with AI: Funded by the Pacific Northwest Transportation Consortium (PacTrans), this project leverages AI techniques to enhance the analysis of traffic accidents, thereby improving our ability to predict and prevent such incidents. By analyzing traffic data through machine learning models, we evaluate key factors such as vehicle behavior, road conditions, and driver distractions. This analysis helps us identify dangerous patterns and conditions that pose significant risks. Utilizing this information, we will develop a dynamic risk map that integrates these insights to identify high-risk areas and anticipate potential accident locations. This tool will assist all drivers, including those with specific needs, and city planners in making safer and more informed decisions regarding road use and urban development.
- AI-Driven Strategies for Diabetes and Depression Management in Alaska Elders: Partially supported by the National Resource Center for Alaska Native Elders, this project utilizes AI to explore the complex relationship between diabetes and depression in Alaska’s elder population. By leveraging explainable AI models, we aim to make the analytical process transparent, enabling a deeper understanding of how environmental and lifestyle factors interact to influence these conditions. This AI-driven approach is crucial as it allows us to detect patterns and connections that are otherwise too complex to discern through traditional methods. As we gather more data, our models have the potential to enhance the precision of health recommendations, thereby improving the management strategies and overall quality of life for this community.
- AI-Enhanced Assistive Technologies for the Hearing Impaired: This project uses AI to specifically address communication challenges faced by hearing-impaired individuals. AI significantly enhances speech-in-noise perception by using algorithms that selectively enhance speech sounds while suppressing background noise. This specialized processing is crucial for individuals who struggle to distinguish speech in noisy environments. Moreover, our AI system extends its capabilities to speech recognition technologies, which are vital for converting spoken words into text. This technology will be designed to accommodate the diverse accents and speech patterns of hearing-impaired individuals, who might have unique speech production characteristics that standard systems like voice assistants often fail to recognize or interpret correctly. Additionally, AI-driven tools are employed to simplify complex texts, directly addressing the language comprehension challenges these individuals face. This aspect of our AI solution is tailored to enhance accessibility, ensuring that written communication is as comprehensible as verbal communication, thus bridging communication gaps that typical tools cannot effectively address.