Modeling Epistemic Knowledge about Users

Representing what an intelligent agent knows or believes can be an important feature for an entity that acts in a collaborative environment. Epistemic logic programs are a clear way to reason about what an intelligent agent knows or believes, but the complexity of the reasoning is exponential. The paper presents a new kind of logic programs, the logic programs with two types of negation as failure operators. An answer set semantics is defined for these programs and it is proved that this semantics allows representing epistemic knowledge. Moreover, a polynomial approximation for this semantics is constructed and the approximation maintains the epistemic knowledge representation possibilities. The paper concludes with examples of knowledge representation resolved using the new proposed type of logic programs. They show how the logic programs can be integrated in a human computer interface. Key–Words: Knowledge representation, epistemic logic programs, answer set semantics, well founded semantics, human computer interface.