 
-Vita
-Classes
(Blackboard
Link)
-Laboratory
-Papers
-More
|
My
work focuses on learning and cognitive development. I am particularly interested
in
high-level cognition, including problems such as categorization, commonsense
reasoning
and language acquisition. I approach these problems by building
computational
models and testing them against behavioral data. Many of these
models
rely on probabilistic inference and draw on recent ideas from statistics,
machine
learning, and artificial intelligence.
A continuing
line of work explores how probabilistic models of reasoning can be
combined
with classic approaches to knowledge representation. Real-world
inferences
often draw on sophisticated background knowledge: structured
representations
can capture some of this knowledge, and probabilistic models can
explain
how this knowledge guides inductive reasoning. I have developed
probabilistic
models of commonsense reasoning that draw on the knowledge
embedded
in structured representations, including ontologies, causal networks, tree
structures,
and logical theories. I have also developed learning algorithms that help
to
explain how these structured representations are acquired.
Much
of this work is carried out within a hierarchical Bayesian framework. Learning
of
any kind must rely on inductive constraints, and I have developed hierarchical
models
that help to explain how some of these constraints are acquired. Two of
these
models explore how constraints that guide word learning and causal reasoning
can
be acquired and used. A third model explores how humans discover which
kind
of
representation is best for a given domain: for instance, how children discover
that
social
networks are often organized into cliques, that comparative relations such
as
"longer than'' or "better than'' are transitive, and that category labels
can be
organized
into hierarchies.
last
updated 9/17/07 CK/tc |
