PhD Defense: Ali Kargarandehkordi

November 10, 2:00pm - 4:30pm
Mānoa Campus, https://ucsf.zoom.us/my/pwashing

Thesis Title: Personalized Machine Learning for Affective Health: Advancing Mobile Sensing and Digital Phenotyping -------------------------------------------------- Abstract: The rise of wearable devices and smartphones has enabled continuous, real-world monitoring of human physiology and behavior — opening the door to proactive, personalized healthcare. Yet, one major challenge remains: people’s physiological and emotional responses to affective states vary widely, making one-size-fits-all models unreliable. This dissertation introduces a personalized multimodal AI framework that integrates biosignals, behavioral patterns, and contextual data to predict affective and physiological outcomes in daily life. Using three complementary studies — Emognition (emotion recognition), BanAware (substance-use craving prediction), and CardioMate (stress-induced blood pressure spike prediction) — the work demonstrates how personalization improves both accuracy and interpretability across controlled and real-world environments. By combining self-supervised representation learning, attention-based temporal modeling, and interpretable feature analyses, the thesis bridges behavioral and physiological domains, showing how personalized digital phenotyping can transform mental and cardiovascular health monitoring. The findings advance the path toward adaptive, transparent, and person-specific digital medicine.


Event Sponsor
Information and Computer Sciences, Mānoa Campus

More Information
2135445586

Share by email