Formal semantics of meta-level architectures: Dynamic control of reasoning

Abstract

Meta-level architectures for dynamic control of reasoning processes are quite powerful. In the literature many applications in reasoning systems modelling complex tasks are described, usually in a procedural manner. In this paper we present a semantic framework based on temporal partial logic to describe the dynamics of reasoning behaviour. Using these models the semantics of the behaviour of the whole (meta-level) reasoning system can be described by a set of (intended) temporal models.

A temporal-interactivist perspective on the dynamics of mental states

“…Work from this perspective has addressed the dynamics of reasoning about the control of reasoning (Treur, 2002), and defeasible reasoning processes from a temporal perspective: Treur, 1995, 1998;Engelfriet et al, 2001;Marek et al, 2001); see also (Meyer and Treur, 2001). This work fully concentrates on the internal interaction and dynamics of mental states during a (defeasible) reasoning process; interaction with the external world is not addressed.…”

Formal semantics of meta-level architectures: Temporal epistemic reflection

“…6, where it is pointed out how temporal models can provide an adequate semantics of meta-level architectures for dynamic control, and Ref. 8, where similar ideas have been worked out to obtain a linear time temporal semantics for default logic.…”

“…In Ref. 6, we introduced formal semantics for meta-level architectures of the first kind based on temporal models. It may be considered quite natural that for such a dynamic type of reasoning system the temporal element of the reasoning should be made explicit in the formal semantics.…”

Adaptive Networks at the Crossroad of Artificial Intelligence and Formal, Biological, Medical, and Social Sciences

“…By addressing adaptiveness through self-models within the network based on the notion of network reification, the modeling approach also belongs to another longstanding area within AI, metalevel architectures, and metaprogramming [19][20][21][22][23][24][25]. The use of self-models provides a modeling technique by which adaptation can be modeled easily and also iteratively.…”