Knowledge representation is a central problem because A.I. attempts to deal with much more complex and less well-structured data than most computer programs.

There are two major classes of knowledge representation methods that are used in A.I.:

• Logic: methods based on mathematical logic, i.e. first-order predicate calculus.

• Frames: methods based on networks of nodes, which represent objects or concepts, and labeled arcs, which represent relations among nodes.

The Logic and Frame methods have sometimes been viewed as competitive, and their proponents have fought. Each method has some valuable features. More recently, some have produced representation systems that combine the good features of both.

Contents    Page-10    Prev    Next    Page+10    Index