Wednesday, 21 March 2007

Logics for Representing Knowledge

What different logics are there for representing knowledge? As an example, if told that "every bird flies, but penguins don't", some possible logics for representing this are:

Default Logic
b(X): f(X)/ f(X)
p(X) -> b(X)
p(X): / ~f(X)

Logic Programming (with negation as failure)
f(X) <- b(X) & not abnormal(f(X))
abnormal(f(X)) <- p(X)
~f(X) <- p(X)

Autoepistemic Logic (where 'L' means 'believes')
b(X) & ~L(~f(X)) -> f(X)
p(X) -> ~f(X)
a -> La (given)
~a -> L~a (given)
L(a -> b) -> (La -> Lb) (given)

No comments: