Extending the Non-monotonic Reasoning Infrastructure for the Semantic Web via Well-founded Negation and incremental Support for Courteous Logic Programs,
August 1, 2005
Production Rules, Description Logic (DL) and Logic Programs(LP) are the key paradigms of knowledge representation. Production rules systems (particularly the JESS rule engine) are based on the Rete network and primarily support forward inferencing. However, they do not have proper semantics for negation.
The Web Ontology Language (OWL) is based on description logic and cannot express rules. Several extensions to OWL including the OWL Rules Language(ORL), Semantic Web Rule Language(SWRL)have been proposed in order to overcome this deficiency. But, in order to keep the reasoning decidable we have resorted to the "Description Logic Programs(DLP)" approach, which considers the intersection of description logic and logic programs and translates OWL constructs within the DLP subset to LP.
Systems based on Logic programs, especially those supporting "Well-Founded Semantics" (WFS) such as XSB have clean semantics for handling negation. However, it is harder to author rules in such systems because of lack of higher level primitives to specify priorities and conflict handling.
All of the above systems do not have the ability to invoke external procedures which can perform side-effectful actions such as sensing and effecting. In this thesis, we present SweetRules which addresses all these issues by evangelizing Well-Founded Semantics and Situated and Courteous Logic Programs.
University of Maryland, Baltimore County
Department of Computer Science and Electrical Engineering
Downloads: 2990 downloads