Research >

Overview

We are a research department with expertise spanning the full breadth of artificial intelligence (AI) and informatics methods. Our faculty work across symbolic and probabilistic AI, machine learning, causal inference, deep learning, generative AI and human-computer interaction. This methodological range supports research programs in clinical informatics, clinical research informatics, population health informatics, translational bioinformatics and imaging informatics, applied to clinical domains including anesthesiology, neurology, pediatrics, critical care, oncology and infectious diseases.

Our research is grounded in the conviction that advancing biomedical AI requires more than developing new methods. It demands equal attention to how those methods are evaluated, implemented and integrated into real-world clinical and research settings. Our faculty and trainees therefore draw on evaluation science, implementation science, data science and human-computer interaction to ensure that the tools and systems we build are not only technically sound but also safe, effective and usable in practice.

Our projects reflect this breadth. In clinical informatics, we develop AI systems that predict surgical complications, detect neurological events in real time, support medication safety and stratify patient risk across a range of conditions, all grounded in close partnerships with clinicians at UPMC. In translational bioinformatics, we apply machine learning to protein interaction networks, gene regulatory programs, single-cell and spatial omics data, and multi-omics integration to uncover molecular mechanisms of disease and identify candidates for drug repurposing. In cancer informatics, we build tools for biomarker discovery, tumor microenvironment characterization and personalized treatment prediction, working alongside investigators at UPMC Hillman Cancer Center and partner institutions. In research data infrastructure, we lead national coordinating centers and federated networks that make large-scale biomedical data accessible, interoperable and ready for rigorous scientific inquiry.