Evidence integration
How can different kinds of biological evidence inform one question?
Investigate representations that connect complementary signals while accounting for missing data, inconsistent measurements, and the limits of each source.
Research / Medicine
Connect evidence.
Clarify what comes next.
Investigating computational approaches to complex biomedical questions, with context and uncertainty at the center.
Better models begin with a clear idea of the question they should answer.
Biomedical research draws on evidence collected in different settings, populations, and formats. Our direction explores how AI can help connect that evidence while preserving the distinctions that make it meaningful.
Questions guiding this frontier
Generalization
When does a model remain useful beyond its original dataset?
Study evaluation across relevant populations and settings, with attention to data leakage, bias, and changes in the conditions under which data was collected.
Interpretation
What does the model support—and what does it leave unresolved?
Explore ways to present uncertainty, identify failure cases, and connect computational findings to the next research step.
The standard of evidence
Keep the claim
close to the evidence.
Our work is framed as research. This website does not offer diagnostic tools, treatment recommendations, or claims of clinical validation.
The next question
changes everything.
Bring a scientific challenge. Let’s explore what becomes possible when disciplines work together.
Start a collaboration