My research in this domain involves using big data analytics to augment critical care service provider, specifically, critical care nurses' perceptual and reasoning capabilities, by solving two kind of "timing issues" in detecting physiological deterioration among continuously monitored patients in critical care units. The first stream of work address the alarm fatigue problem that causes delay in recognition of critical states. A second stream of work aims to provide proactive alarm before the vital sign deterioration become overt: this will give medical team extra time for preparedness and improve patient's safety and wellness. My ultimate vision in this line of research is to create "peace of mind" for critical care team, patient and their family by equipping them with an extra pair of vigilant "virtual eye", which is enabled by powerful computational techniques with massive amount of data.
My research in this domain focuses on augmenting human teachers’ perceptual and reasoning capability in providing personalized and adaptive support to students in off-line coaching scenarios. Effective coaching of problem solving at a young age needs to resolve an instance of the “assistance dilemma”, i.e. making real time decisions on which kind of support to provide (cognitive, metacognitive or social), at what times, in order to maximize students' exposure to “productive struggles” while being mindful of unproductive ones. Though this type of ideal coaching is possible at a one-on-one level, it is often not practical given the large student-to-teacher ratios in a typical classroom. I envision an intelligent monitoring agent that could sense students’ cognitive and affective states, infer latent assistance needs, and alert the teacher to direct their coaching resources. Wherever possible, this agent could also act as the busy teacher’s assistant to provide just-in-time metacognitive or social emotional supports to improve engagement and foster student perseverance behaviors. Part of my dissertation work is the first step toward this vision: developing proof-of-concept for the perceptual module of the system (i.e. detection of cognitive/affective states of students) and exploring the dynamic interplay between those states and interventions.