Landmark hierarchies in context

Abstract

We are interested in the generation of distinguishing place or route descriptions for urban environments. Such descriptions require a hierarchical model of the discourse, the elements of the city. We postulate that cognitive hierarchies, as used in human communication, can be sufficiently reflected in machine-generated hierarchies. In this paper we (a) propose a computational model for the generation of a hierarchy of one of these elements of the city—landmarks—and (b) demonstrate that a set of filter rules applied on this hierarchy derives distinguishing route descriptions from spatial context.