Computational Models of Development in Infants and Machines
Research Prospectus
A common paradigm in artificial intelligence takes the position that
intelligent behavior is derived from lots of domain-specific knowledge.
However, our efforts at engineering knowledge work only in narrow contexts and
produce, therefore, very special purpose artifacts. In contrast, traditions in
the behavioral sciences appreciate the interactive aspects of knowledge and
intelligence, but provide relatively little insight into mechanisms. Recent
paradigms emerging in Philosophy, Psychology, Linguistics and Robotics offer
a plausible framework for modeling the process of sensorimotor
development. Our goal is a computational account of development
that helps us design intelligent machines and explains observations of
human infants.
We posit mechanisms for organizing experience as behavioral policies, and for
representing sensorimotor experience. These "grounded" representations form
the basis of a conceptual structure that influences all subsequent cognitive
development.