01 / Research program
Artificial intelligence
We study how intelligent systems can become useful capabilities: dependable in practice, grounded in real environments, and designed to extend what people can do.
From capability to consequence.
Progress in AI matters when it changes what people can understand, create, or accomplish. Our work connects advances in learning and reasoning to concrete problems, with attention to the full system around a model: data, tools, evaluation, and the people who use it.
Questions we pursue
- How can systems act reliably when conditions are unfamiliar?
- What does meaningful evaluation look like beyond a benchmark?
- How can AI support discovery across science and engineering?