My research explores the computational basis of human cognition, considering the abstract computational problems that people face every day, and examining how people solve these problems. I am particularly interested in problems of induction, which involve making inferences about the underlying structure of observed data. Making the leap from data to structure is an essential part of the acquisition and organization of knowledge, and the question of how people do this makes contact with issues in philosophy of science, machine learning, and statistics.

Currently, I am focusing on two questions: how people reason about chance and causality, and how the statistical properties of language influence cognition. My work on chance and causality examines how people learn that causal relationships exist, what makes something seem "random", how to measure the strength of a coincidence, and how people predict the future. My work on language involves using probabilistic generative models to extract interesting structure from collections of documents, and using this structure to make predictions about cognitive tasks. For example, one such model allows us to identify the topics discussed in a set of documents, and to predict associations between words (why you think "cat" when I say "dog") from these topics. These techniques are related to models of similarity and categorization, another of my interests.

Each of these questions involves an interesting computational problem. Through modeling and experimentation, I explore how people solve this problem. Many of the computational problems that people face appear in other contexts. For example, randomness and causality both involve problems that have been studied extensively in computer science and statistics. An important part of my research is making connections between cognitive science and other computational disciplines. These connections go both ways: I use concepts from statistics, artificial intelligence, machine learning, and elsewhere in computer science in explaining human cognition, and I use the cognitive capacities of people as inspiration to develop models and tools that contribute to these other disciplines.

click here for a complete set of references and download information