Research Overview
During the course of development, human infants gather information about the external world without the benefit of an extensive base of knowledge that adults automatically bring to bear on perceptual, motor, cognitive, and language tasks. What mechanisms allow infants to acquire this initial level of information and how does that information guide subsequent learning? Clearly, most learning that occurs in infancy, and a substantial amount of learning in adulthood, is performed without instruction—it is implicit and based on an analysis of the distributional properties of environmental stimulation.

For over two decades, my research has been directed at exploring and understanding these implicit learning mechanisms, which are typically referred to as “statistical learning”. Although initially studied in the task of word segmentation from fluent speech, statistical learning has been extended to other domains, such as musical tones, phonetic categories, sequences of visual shapes, sequences of motor responses, and combinations of objects (or object parts) in complex visual scenes. An important goal of these studies is to reveal the computational constraints that enable statistical learning to be tractable given the complexity of the input and the infinite number of statistical computations that are possible over any set of inputs. Initial computational models of statistical learning focused on bi-gram statistics and conditional probabilities, but more recent work has broadened to include Bayesian ideal learning models. Empirical studies of statistical learning have also evolved to explore order effects in learning multiple structures and to understand how statistical patterns trigger the formation of categories.