Scalable Biomedical Modeling

The Scalable Biomedical Modeling (SBM) group at Oak Ridge National Laboratory develops methods and tools that make biomedical and health data usable, reliable, and impactful at scale. We are motivated by a rapidly changing scientific landscape: Advances in AI are reshaping research workflows, from hypothesis generation and experimental design to automated code execution, model interpretation, and communication of results. Our mission is to create, validate, and deploy these capabilities to accelerate data-driven discovery while safeguarding privacy and upholding scientific rigor. 

Our research spans three thrusts:

Data curation and AI readiness 

Tackling the growing complexity of distributed and sensitive datasets by building pipelines for curation, harmonization, and interoperability. We work to make biomedical data AI-ready and to lower barriers to responsible, collaborative science.

Privacy-enhancing technologies for AI

Developing and evaluating approaches such as synthetic data generation, federated learning, and adversarial privacy testing. Our aim is to ensure that sensitive biomedical and proprietary scientific data can be shared, analyzed, and modeled while preserving utility, integrity, and trust.

High-performance computing and agentic AI workflows

Harnessing leadership-class computing platforms and emerging AI agent technologies to design, train, and deploy models at scale. We are creating workflows that link simulation, learning, and decision-making, enabling more autonomous and adaptive scientific exploration.

While centered on biomedical and public health applications, our research also addresses pressing challenges in biological and environmental sciences, energy, and national security. We aim to deliver both foundational advances—new algorithms, benchmark frameworks, and privacy-preserving methods—and practical solutions deployed on DOE’s world-leading computing resources.