Many important questions lie in areas where data does not yet exist, is scarce, or is difficult and expensive to obtain. Progress therefore requires more than training larger models on existing observations.
Our work aims to give researchers with deep domain knowledge the context, tools, and computational methods needed to reason beyond the available data. This includes combining partial evidence, encoding physical constraints, exploring plausible hypotheses, and identifying the experiments most likely to produce useful knowledge.
The goal is not to replace scientific understanding, but to help experts extend it—creating the conditions for future breakthroughs in how we understand physical and chemical systems and, eventually, the human body.
My role in this area is to enable and empower, to scope research problems, connect disciplines, and support the people doing the scientific work. My group’s projects and publications provide the precise record of individual contributions.