Knowledge compilation is a family of approaches for addressing the intractability of a number of artificial intelligence problems.

A propositional model is compiled in an off-line phase in order to support some queries in polynomial time. Many ways of compiling a propositional model exist.[1]

Different compiled representations have different properties. The three main properties are:

  • The compactness of the representation
  • The queries that are supported in polynomial time
  • The transformations of the representations that can be performed in polynomial time

Classes of representations

edit

Some examples of diagram classes include OBDDs, FBDDs, and non-deterministic OBDDs, as well as MDD.

Some examples of formula classes include DNF and CNF.

Examples of circuit classes include NNF, DNNF, d-DNNF, and SDD.

Knowledge compilers

edit
  • c2d: supports compilation to d-DNNF
  • d4: supports compilation to d-DNNF
  • miniC2D: supports compilation to SDD
  • KCBox: supports compilation to OBDD, OBDD[AND], and CCDD

References

edit
  1. ^ Adnan Darwiche, Pierre Marquis, "A Knowledge Compilation Map", Journal of Artificial Intelligence Research 17 (2002) 229-264