Treffer: An algebraic interpretation of semantic networks

Semantic networks are among the most popular knowledge reresentation techniques. They have been applied to a large spectrum of applications, including spatial tasks, such as object recognition from images. Their appeal lies in the combination of a structure that combines standard abstraction mechanisms with a simple visual representation. However, applications of semantic networks suffer from their lack of theoretical foundations. The semantics of spatial domains is often modeled with a technique whose semantics are themselves unclear. In our work on semantic interoperability of GIS, we have found this situation to be potentially harmful, but also repairable by our tools. It can be harmful by luring necessary work on application semantics into potentially muddy waters. And it is repairable by interpreting semantic networks from the point of view of algebra, i.e. another semantic modeling technique. Thus, the paper proposes an algebraic interpretation of semantic networks, showing how this perspective clarifies their own semantics and how it allows for a sound semantic modeling of spatial domains.