Event Date
Secure and Privacy-Preserving AI/ML for Health Applications:
AI & machine learning techniques have demonstrated great promises in tackling various predictive and decision control problems in a variety of application domains, including healthcare. In collaboration with UC Davis Medical Center, Alzheimer Disease Center, and the MIND Institute, my research group has developed the data analytic pipeline for AI-assisted predictive analytics and critical patientcare, histopathology image analysis for deep phenotyping of human brain, and video-based autistic behavior detection. Using different case studies, this talk will give an overview of technology behind our AI/ML-powered solutions, as well as the challenges (e.g., security, privacy, fairness) and opportunities associated with their practical deployments. In addition to accuracy, our team strives to develop computationally efficient AI/ML models with minimal reliance on labeled data in our attempt to make the AI/ML workflow accessible to all researchers and deliver affordable AI/ML-powered health services to under-resourced populations. We are also interested in designing secure and privacy-preserving AI/ML workflow for delivering digital health services indifferent operating scenarios.