Processing Schedules using Distributed Ontologies on the Semantic Web

Abstract

The Semantic Web uses formal distributed ontologies for representing relationships among concepts in the real world. A structured framework such as this allows agents to peruse and reason about published knowledge without the need for scrapers, information agents, and centralized ontologies. However, in order to process any information, an agent must be familiar with the underlying ontology used to markup that information. However, no single agent can be expected to be familiar with all possible ontologies that may be available on the Semantic Web. Therefore, translation services that transform concepts defined within previously unknown ontologies into known concepts allow agents to understand the available information and hence achieve their goal. This transformation may be achieved by invoking specific markup translation services or by logical reasoning through other shared ontologies. The RETSINA Calendar Agent (RCal) is a Distributed Meeting Scheduling Agent that processes schedules marked up on the Semantic Web, and imports them into the user's Personal Information Manager. Translation services, which are used to translate unknown concepts into known concepts, which are located using a DAML-S based service discovery mechanisms. In this paper, we present RCal, and demonstrate how it extracts and uses meaningful knowledge from the semantic markup. In addition, we describe how web-service discovery mechanisms are used when new concepts are encountered.