D.J. Buehrer and L.-R. Chien (Taiwan)
Class Algebra, Artificial Intelligence, Reasoning, Fuzzy, State Space.
Class algebra extends the current definitions of RDF, RDF S, and OWL DL semantics [1,2] to include other traditional AI knowledge and inference mechanisms, such as planning using state-space operators. There are basically two kinds of inferences: the first phase adds a complete set of entailments for a given state of the knowledge base. A second "state-space" phase adds new states, based on agent actions on a given state of the knowledge base. Each operation of an agent is accompanied by a set of additions and deletions of triples, which may involve new object identifiers (oids). The "frame axiom" is implemented similarly to other inference mechanisms (e.g. symmetry, transitivity, cardinality axioms) in OWL, and can implicitly calculate which triples can be copied from previous states without being affected by the addition and deletion triples of the operators. Fuzzy subsumption is used both in forming the IS-A hierarchy and during the state-space search. As well as outlining the semantics of class algebra, we will also discuss some implementation optimizations for storing the results of such entailed knowledge.
Important Links:
Go Back
