When we interact with other people, we can make rich predictive inferences about their future behavior on the basis of very little data. This learning is often supported by abstractions like preferences and goals. How do we learn from patterns in behavior in the absence of these abstractions, and how do these abstractions develop?

My work on adaptive reasoning has primarily used the game of Rock, Paper, Scissors to investigate what kinds of behavioral patterns people can learn over time and how they adjust their own behavior when a sneaky opponent is exploiting them.

A recent line of work in this space has also looked at how people develop trust in others over repeated interactions, focusing on evaluations of a helpful agent's competence in physical task settings.

In many learning environments, the process of explaining new data can lead adults and children to favor more abstract, generalizable solutions. This has obvious educational import, but also raises interesting psychological questions. How does explanation lead to greater abstraction when learning?

My work in this space has looked at the impact of explanation on relational reasoning in kids, and tried to nail down whether explanation helps adults generate certain hypotheses, or makes them evaluate existing hypotheses differently.

Adults have an amazing ability to reason about abstract number in all kinds of ways. But this doesn't come for free; kids struggle to learn basic concepts of number long after they've learned how to count. What's the basis for adult number reasoning and how do kids learn it?

Work I've done looking at numerical reasoning has used computational modeling techniques to try and formalize, a) the strategies children use when reasoning about number and, b) the process by which adults estimate number in the world around them.