A commonsense knowledge base supported multi-agent architecture
Expert Systems with Applications: An International Journal
| Hi-index | 0.00 |
The context of a statement or assertion can be critical to a usefulunderstanding of its meaning. In huge knowledge bases such as theCYC project, capturing context explicitly encounters significantproblems. Research at the Intelligent Agent Lab at The Universityof Melbourne has been investigating lightweight approaches toincorporating context in the construction of knowledge-basedinformation agents. The prototypical example of a successful agentbuilt in the lab is SportsFinder. SportsFinder extracts sportingmatch results from a large variety of web sites without any formalknowledge representation. This is a task for which understandingthe context of knowledge on the web pages is critical, as theindividual elements that make u a page of sports results are veryambiguous without consideration of the surrounding information.This paper advocates incorporating context via a taskspecification, without resort to general purpose knowledgetechniques.