Addressing the Signal Grounding Problem for Autonomic Systems

Abstract

This paper is concerned with the problem of intrinsically assigning meaning to the signals responsible for autonomic responses in a system. Without an associated cognitive system, the Symbol Grounding Problem would constitute a major barrier in system adaptation and evolution. Based on an ongoing effort towards a formal and pragmatic development of self-regenerative software systems, this paper adopts concepts from Artificial Immune Systems (AIS) engineering, Information Theory and the Situation Calculus dialect of predicate logic. These are used to formalise the monitoring and control of system autonomic functions. In this way danger signals as an immune (self-healing/protecting) response and evolutionary (self-adapting) responses can be formalised into autonomic conditional and anticipatory reaction triggers. Thus any threat or potential enhancement to the system can be monitored for and the appropriate action taken to facilitate system dependability and safety.

Engineering Autonomic Systems Self-Organisation

“…In order for autonomic components to adapt the system to novel or unforeseen circumstances it is necessary for the system to have knowledge of the intrinsic meaning of observed signals [19]. This will involve the system gaining new concepts, represented as action histories leading to an observed phenomenon, from the runtime system, which are then rewarded or penalised according to the success of future executions based on these concepts.…”

“…This will involve the system gaining new concepts, represented as action histories leading to an observed phenomenon, from the runtime system, which are then rewarded or penalised according to the success of future executions based on these concepts. Further details may be found in [19] and [20]. Rather than simply providing a stimulus to action the piece of system knowledge will have a precedent and likely consequence to the system that may be grounded at some appropriate observer module.…”

Distributed redundancy and robustness in complex systems

“…Using an approach to signal grounding [26] within situation calculus, the action based semantics dictate the grounding of an emergent symbolic representation of the system by: , do a 1 , do(a, s) with SR(a, s) = SR a, do a 1 , do(a, s) where a is the sensing action for the algebraic connectivity λ 1 and a 1 is the rewiring action giving…”

Situation Calculus for HCI Design

“…As complexity increases, however, it becomes increasingly difficult, if not impossible, to implement user interfaces through the hard coding of all the requirements for adaptation to the user's situation. The work reported on here is seeking to use formal methods with logic for matching users' specific contexts (device type and capabilities, location, interaction style, etc) to the underlying functional systems: It is proposed to use an established layered observation system [1] with a formal logical model to provide both an inherent verification process for the interface and a cognitive (logic) procedure to automatically generate and update the interface at runtime. In particular the Situation Calculus will be used, as a formal calculus of situations: It offers a flexible formalism in which to reason about dynamic systems, simulate likely system evolutions in a couterfactual manner and perform system property verification entirely within the logical calculus.…”