Increasing access to data sources on the Internet offers expanding opportunities for equipping intelligent applications with the content they require whether broad in scope or rich in detail. Although typically originating within the web in a semi-structured form, with the use of inference-based translation and analysis mechanisms such content can be transformed into useful information and ultimately into actionable knowledge. Service-Oriented Architecture (SOA) offers a platform for accessing the web as invocable resources and effectively incorporating multiple sources of data and capabilities on the Internet into enterprise applications. Adding inference capabilities to SOA-based applications not only aids in the translation of data into information thus increasing visibility into the sea of content that is the web, but also provides a powerful mechanism for performing the domain-centric decision making that is the heart of intelligent applications. The Web Ontology Language (OWL) offers the medium and the tools necessary to represent models of business activities as well as support native inference across related semantic concepts. In this paper the authors present an architecture for combining OWL with a SOA-based paradigm to enhance traditional web applications with powerful inference capabilities. Commensurate with a service-oriented theme, specific techniques are presented for representing the translation activity itself as a service. The paper concludes with a discussion of two distinct types of inference: one internal to the OWL model and the other externalized into intelligent agents that operate across OWL-based concepts.


Software Engineering



URL: https://digitalcommons.calpoly.edu/cadrc/95