AI for Public Health Initiative
Research Projects Using AI Tools and Analysis
In public health research, AI provides powerful tools to help gather digital data from multiple sources, and to analyze disparate and large datasets. We have already begun to use AI for research projects, and we’re training our students to understand the different ways that AI can be deployed to enhance and target research that will help us shape a healthier world.
AI-Enhanced Infectious Disease Epidemiology
Through three complementary projects funded by the Global Flu View Spark program, researchers are fusing participatory surveillance, wastewater analytics, hospitalization records, weather signals, and county-level engagement to model influenza-like-illness from the national down to the neighborhood scale. In close partnership with local health departments, our students turn these multimodal forecasts into “data-to-action” briefs that inform the needs of public health interventions. These efforts are funded by the Skoll Foundation.
Researchers
Paulina Colombo, MPH
Royani Saha
Kelly Lee
Onicio B. Leal Neto, PhD, MPH
Multimodal AI to Contain MDROs in Nursing Homes
Supported by the Arizona Department of Health Services (ADHS), this project links electronic health records, wearable sensor streams, and realistic synthetic datasets to map microbial transmission pathways and pinpoint high-risk touchpoints inside long-term-care facilities. The goal is to generate precision intervention bundles that cut multidrug-resistant organism (MDRO) incidence without increasing staff burden. These efforts are funded by the Arizona Department of Health Services (NIA-Grant).
Researchers
Onicio B. Leal Neto, PhD, MPH
Kate Ellingsson, PhD
Paulina Colombo, MPH
Ciro Cattuto, PhD (ISI Foundation, Italy)
Rodrigo Paiva, PhD (UPE, Brazil)
Privacy-Preserving Graph Neural Networks & Federated Learning
Our multi-disciplinary team is customizing graph neural networks to model contact structures in school settings while keeping identifiable student data decentralized via federated learning. This approach enables statewide risk forecasting, scenario testing (e.g., mask policies, ventilation upgrades), and equitable resource allocation, all without exporting raw data beyond district firewalls.
Researchers
Onicio B. Leal Neto, PhD, MPH
Amanda Wilson, PhD (MEZCOPH)
AI Against Sexual Violence Online
Leveraging cross-platform text mining and behavior-sequence modeling, we build archetypes of rape-supportive discourse that transcend individual social networks. The resulting classifiers help moderators and public-health agencies spot coordinated harassment campaigns, tailor prevention messaging, and evaluate the impact of platform policy changes on victim safety.
Researchers
Maisun, Ansary, MPH
Iman Hakim, MBBCh, PhD, MPH
Mary Koss, PhD
Onicio B Leal Neto, PhD, MPH