Leo Ferres, Avi Parush, Maria Rasouli, Meaghan Beirne:
Modelling spatial dialogs: The Speaker
Role
Abstract
This paper deals with the formalization and functional modelling of the cognitive
tasks speakers seem to carry out while engaged in a particular kind of spatial
dialog (a map comparison task). Empirical data and formalization are thus
integrated to propose a possible algorithm of these tasks. Presently, our
model consists of three sub-modules: the ontological module (which establishes
the existence of an object), the alignment module (which attempts to achieve
a shared spatial representation) and the buffer module (which decides when
information is enough to provide a judgement to the task at hand). A discussion
of the workings, interaction and weaknesses of the submodules (and of modelling
dialogs in general) ensues.
|
