Representation of fragmentary and multilayered knowledge: a semiformal metatheory as an interactive metalogic program

Abstract: "Metalogic programming is an important technique for the three interrelated topics 'knowledge representation', 'knowledge processing', and 'knowledge assimilation'. Its expressiveness enables a logically pure structure preserving representation of multila...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hamfelt, Andreas (VerfasserIn), Hansson, Åke (VerfasserIn)
Format: Buch
Sprache:English
Veröffentlicht: Uppsala, Sweden 1991
Schriftenreihe:Uppsala Programming Methodology and Artificial Intelligence Laboratory: UPMAIL technical report 68
Schlagworte:
Zusammenfassung:Abstract: "Metalogic programming is an important technique for the three interrelated topics 'knowledge representation', 'knowledge processing', and 'knowledge assimilation'. Its expressiveness enables a logically pure structure preserving representation of multilayered, fragmentary knowledge which, since not fully formalisable, requires the assimilation of externally supplied knowledge. We have gained new methodological insight contrasting multilayered and imprecise theories to exact theories, by splitting the formalisation problem into three distinct theories: the informal theory to be formalised, the formal theory, and the informal or, as in our case, semiformal metatheory discussing the two former
Our example is hierarchical legal knowledge where rules proposed for inclusion in a theory of legally acceptable rules at a level i must be accepted by (meta)rules of legal interpretation in a theory of legally acceptable metarules at level i+1 which in their turn must be accepted by metametarules at level i+2, ets. Moreover, legal knowledge is fragmentary in the sense that it contains only schematic descriptions of the rules at the respective levels and from these schemata case specific rules must be dynamically generated for each particular adjudication. Legal reasoning between two adjacent levels exhibits a clear informal counterpart to upward reflection while, on the contrary, downward reflection is disallowed. Our study is the first illustrating the use of reflection in a representation of realistic knowledge
Furthermore, metalogic programming with reflection allows us to easily cope with changes in the dynamic domain knowledge since it enables a modular representation which is structure preserving both horizontally and vertically. We present a semiformal metatheory expressing legal knowledge as a Horn clause metaprogram, implemented in Prolog, which interactively constructs and presents metaproofs (proof terms) allowing the user to assess and accept or reject the conclusions arrived at.
Beschreibung:45 S.

Es ist kein Print-Exemplar vorhanden.

Fernleihe Bestellen Achtung: Nicht im THWS-Bestand!