Research
Research
I build AI methods that learn from real-world health data, such as electronic health records, medical images, and physiological signals, and that hold up when they are used for real patients.
Fair machine learning
Models trained on electronic health records inherit the gaps and biases of how care is delivered and recorded. I develop methods that measure and reduce those biases, such as individualized polysocial risk scores that bring social determinants of health into risk prediction without disadvantaging minority patients, and I study how data continuity and missingness affect fairness.
Key papers
-
A fair individualized polysocial risk score for identifying increased social risk in type 2 diabetes
Nature Communications, 15(1), 8653, 2024 Journal
-
A scoping review of fair machine learning techniques when using real-world data
Journal of Biomedical Informatics, 151, 104622, 2024 Journal
-
Adaptation of a fair individualized polysocial risk score for hospitalization risk prediction
JAMIA Open, 9(4), ooag161, 2026 Journal
Disease subphenotyping and prediction
Patients with the same diagnosis can follow very different courses. I build graph-based and temporal deep learning models that uncover progression subphenotypes in diseases such as Alzheimer's disease and type 2 diabetes, and early-classification methods that make reliable predictions from partial signals such as ECGs.
Key papers
-
Journal of Healthcare Informatics Research, 10(2), 317–339, 2026 Journal
-
Temporally Detailed Hypergraph Neural ODE for Disease Progression Modeling
The Fourteenth International Conference on Learning Representations (ICLR), 2026 Conference
-
Snippet Policy Network for Multi-class Varied-length ECG Early Classification
IEEE Transactions on Knowledge and Data Engineering, 35(6), 6349–6361, 2022 Journal
-
Snippet Policy Network V2: Knee-guided Neuroevolution for Multi-lead ECG Early Classification
IEEE Transactions on Neural Networks and Learning Systems, 35(2), 2167–2181, 2022 Journal
Health digital twins
A health digital twin is a computational counterpart of a patient that is updated as new data arrive. I build twins from longitudinal real-world data to forecast outcomes, such as kidney function decline in type 2 diabetes, and to compare care pathways before they are tried on the patient.
Key papers
-
Diabetes, 74(Supplement 1), 399-P, 2025 Abstract
Retinal imaging
The retina is imaged routinely and reflects vascular and neural health. I work on foundation models that combine volumetric and planar eye imaging, on explanations aligned with retinal anatomy, and on multimodal systems that use routine eye care to assess dementia risk.
Key papers
-
arXiv preprint arXiv:2605.02714, 2026 Preprint
-
SAIL: Structure-Aware Interpretable Learning for Anatomy-Aligned Post-hoc Explanations in OCT
arXiv preprint arXiv:2605.02707, 2026 Preprint
-
Nature Biomedical Engineering, 2026 JournalAccepted. Preprint: arXiv:2603.21597
Funding
Current research support
My work is supported by the National Institutes of Health, including NIA, NIEHS, and NIAID.
-
2026–2031
Senior/Key Personnel
AICC: AI-Powered Coordinating Center for the AITC Program on Aging and AD/ADRD Research
NIH 1U24AG103119 · Yale University
-
2025–2029
Co-Investigator
Semantics Standards and Tools for Spatial and Contextual Exposome Data
NIH/NIEHS 5R24ES036131 · Indiana University Indianapolis
-
2024–2029
Co-Investigator
Standardizing and Harmonizing Behavioral and Social Science Research Factors in Alzheimer's Disease through Ontology-Based Approaches
NIA 5U01AG088076 · Mayo Clinic Jacksonville
-
2024–2029
Co-Investigator
iSMART: intelligent Social risk Management in AD/ADRD paTients
NIH 7R01AG089445 · Purdue University
-
2023–2028
Co-Investigator
ACTS: Developing Advanced Informatics Approaches for an Alzheimer's Disease Clinical Trial Simulation System
NIH 5R01AG084236 · Mayo Clinic Jacksonville
-
2023–2027
Co-Investigator
An End-to-End Informatics Framework to Study Multiple Chronic Conditions' Impact on Alzheimer's Disease Using Harmonized Electronic Health Records
NIH 5R01AG083039 · UTHealth Houston
-
2023–2028
Co-Investigator
AI-ADRD: Accelerating Interventions of AD/ADRD via Machine Learning Methods
NIH 4R01AG077820 · University of Pennsylvania
-
2023–2027
Co-Investigator
Artificial Intelligence and Counterfactually Actionable Response to End HIV (AI-CARE HIV)
NIH 5R01AI172875 · University of Florida
-
2023–2028
Co-Investigator
Disparities of Alzheimer's Disease Progression in Sexual and Gender Minorities
NIH 5R01AG080624 · University of Florida
-
2023–2027
Co-Investigator
Eligibility Criteria Design for Alzheimer's Trials with Real-World Data and Explainable AI
NIH 5R01AG080991 · Weill Cornell Medicine
-
2022–2027
Co-Investigator
Computational Drug Repurposing for AD/ADRD with Integrative Analysis of Real-World Data and Biomedical Knowledge
NIH/NIA 5R01AG076234 · Cornell University