Is a Semantic Web Agent a Knowledge-Savvy Agent?

Abstract

The issue of knowledge sharing has permeated the field of distributed AI and in particular, its successor, multiagent systems. Through the years, many research and engineering efforts have tackled the problem of encoding and sharing knowledge without the need for a single, centralized knowledge base. However, the emergence of modern computing paradigms such as distributed, open systems have highlighted the importance of sharing distributed and heterogeneous knowledge at a larger scale—possibly at the scale of the Internet. The very characteristics that define the Semantic Web—that is, dynamic, distributed, incomplete, and uncertain knowledge—suggest the need for autonomy in distributed software systems. Semantic Web research promises more than mere management of ontologies and data through the definition of machine-understandable languages. The openness and decentralization introduced by multiagent systems and service-oriented architectures give rise to new knowledge management models, for which we can’t make a priori assumptions about the type of interaction an agent or a service may be engaged in, and likewise about the message protocols and vocabulary used. We therefore discuss the problem of knowledge management for open multi-agent systems, and highlight a number of challenges relating to the exchange and evolution of knowledge in open environments, which pertinent to both the Semantic Web and Multi Agent System communities alike.

More information

Identifiers

Catalogue record

Export record