Human-centered Artificial Intelligence (AI) has immense potential to improve human health and well-being, by assisting humans and augmenting their capabilities in solving the most pressing clinical needs. The third wave of AI encapsulates not only technological breakthroughs but also a human-centered approach. On the one hand, these human-centered AI-assisted healthcare systems can help clinicians in disease diagnosis and treatment planning, while on the other hand, these systems can help improve doctor-patient interactions, patient well-being, and mental health.
Since human perception of the world is multimodal, developing human-centered AI-assisted healthcare systems require the fusion of multimodal machine learning with interdisciplinary domain knowledge. Technical research questions involve how to seamlessly integrate and learn from complementary multimodal data in clinically relevant ways, how to design experiments and evaluate these systems in clinically meaningful ways, and how to enable translation from the laboratories to the clinic to enable clinician-AI synergy.
The first part of this talk will highlight some of my current and postdoctoral research on developing human-centered AI-assisted systems to assist clinicians in radiology imaging-based urologic cancer detection and risk stratification. The second part of this talk will highlight some of my doctoral research on developing human-centered AI-assisted systems to understand human behavior and group dynamics. The third part of my talk will briefly outline how I intend to marry my doctoral and postdoctoral research experience to develop human-centered AI-assisted healthcare systems that not only improve disease diagnosis and treatment planning but also assist in improving overall health and well-being of patients through unobtrusive sensing and behavioral analysis. Last, but not least, the talk will touch upon my teaching philosophy and experience, and the courses I am keen on teaching to train the next generation of engineers and researchers.