Semantic Web Dog Food

Presented at: 7th International Semantic Web Conference (ISWC2008)

by Aris van Dijk, Huib Aldewereld, Virginia Dignum

Reidentification has been recognised as the most central job of cognition. In this paper, we motivate that concepts as abilities to reidentify, rather than classifications, should be the basis of an agent's conceptuology. Most concepts are not classes; class definitions are artificial, often context-dependent, and don't use inductive knowledge. We will present the basic concepts of CROC, a Representational Ontology for Concepts. Artificial agents can have concepts through language representations alone. Language-like representations, based on lexical concepts, plus reasoning, will be able to solve the interoperability problem to a large extent. By using these concepts, agents can interoperate without need for shared ontologies and with freedom for own conceptions.

Keywords: Semantic Web, agents, concepts, ontology, representation

Resource URI on the dog food server: http://data.semanticweb.org/conference/iswc/2008/paper/poster_demo/10

Explore this resource elsewhere:

• on Snorql
• on VisiNav
• on Sindice