Research
My research develops reliable and deployable AI systems for healthcare, organized around three themes. Each theme lists the projects I have led or contributed to; the full list is on the Publications page.
- Uncertainty Quantification & Trustworthy AI
- Multimodal Mobile Sensing & Wearables
- EHR & Clinical Data: Prediction and Patient Phenotyping
Uncertainty Quantification & Trustworthy AI
Methods that let clinical models recognize when they may be wrong — from modality-specific uncertainty in multimodal fusion to uncertainty–risk alignment in language models — so predictions can be trusted, or deferred to clinicians, appropriately.
Incorporating Uncertainty in Predictive Models Using Mobile Sensing and Clinical Data: A Case Study on Persistent Post-surgical Pain
Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT), Vol. 9, No. 2, Article 58, June 2025 (33 pages)
Presented at ACM UbiComp / ISWC 2025 (Helsinki, Finland).
Multimodal Mobile Sensing & Wearables
Predictive models that fuse wearable signals, ecological momentary assessments, and clinical data to forecast multi-dimensional recovery and pain outcomes for surgical patients.
Predicting Multi-dimensional Surgical Outcomes with Multi-modal Mobile Sensing: A Case Study with Patients Undergoing Lumbar Spine Surgery
Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT), Vol. 8, No. 2, Article 81, June 2024 (30 pages)
Presented at ACM UbiComp / ISWC 2024 (Melbourne, Australia).
EHR & Clinical Data: Prediction and Patient Phenotyping
Machine learning on structured EHR, survey, and population health data to discover patient subgroups through clustering and phenotyping, predict clinical risk, and support treatment decisions.
Validation of Extracorporeal Membrane Oxygenation Mortality Prediction and Severity of Illness Scores in an International COVID-19 Cohort
Artificial Organs 47(9):1490–1502, 2023.
