Direct mapping

The Direct mapping or direct imaging is a method that the conversion of logical description axioms enables in logic program rules (and vice versa).

Logic program rules can be used to establish logical relationships, e.g. B. Formulate implications . Example: the rule

${\ displaystyle Ruede (X) \ leftarrow Dog (X), Man (X).}$

means that every object X that fulfills the characteristics “dog” and “male” is a “male”. This knowledge representation formalism is widespread, among other things it is used in logical programming languages such as Prolog . In recent years, however, another form of representation has gained in importance with the description logic. Among other things, it can be used to model sub-concept relationships. B. one could define "man" as a subclass of "human", since a man is a more specified person:

${\ displaystyle man \ sqsubseteq man}$

Both formalisms have their advantages and disadvantages in terms of expressiveness and decidability, which is why they also exist in parallel. The Direct mapping allows each other to convert the two representation formalisms, which can be useful if knowledge from multiple knowledge base should be shared. Among other things, this is interesting in the context of the semantic web . However, the method can only be applied to a subset of the intersection of description logic and logic programs. The power of the language defined by this intersection corresponds roughly to that of OWL Lite .

Direct mapping was first mentioned in 2003 in an article by Benjamin N. Grosof.

Web links

http://www.michaelhanselmann.de/projekte_directmapping.htm More detailed presentation of the topic