Extended Metacognition for Artificially Intelligent Systems (AIS): Artificial Locus of Control and Cognitive Economy

摘要

Theories into human learning and cognition have led to much research into new methods and structures for Artificial Intelligence (AI) and Artificially Intelligent Systems (AIS) to learn and reason like humans. As we move toward completely autonomous AIS, the ability to provide metacognitive capabilities becomes important [Crowder and Friess 2011b] in order for the AIS to deal with entirely new situations within the environment it may find itself (e.g., deep space, deep undersea). Presented here are theories and methodologies for Constructivist Learning (CL) processes that provide the methodologies to allow completely autonomous AIS to understand, evaluate, and evolve its "Locus of Control [Watts 2003]." Presented will be the a discussion of how the use of AI learning systems, like Occam [Crowder and Carbone 2011a] and PAC learning can be combined with Cognitive Economy concepts to provide this constructivist learning process to allow a Locus of Control evolution within the AIS. The goal here is to provide the AIS with a fully autonomous, cognitive framework that would be required for autonomous environmental interaction, evolution, and control. In addition, provided are the mathematical constructs, based in Banach Spaces and Lebesque's work in Bounded Variability, that will provide the basis for Cognitive Economy structures in Artificially Intelligent Systems (AIS), allowing the AIS to operate in a "Bounded Rationality" mode, similar to humans, that will allow the autonomous system to function in new, unforeseen, and challenging environments it may find itself in. Natural intelligence filters out irrelevant information (either raw sensory perception information or higher-level conception information), and categorizes the problem representations to allow for maximum information processing with the least cognitive effort. This work is based on the use of Intelligent Software Agents (ISAs) [Crowder 2010a] which will represent the world (its tasks, goals, and information) in terms of the reward values associated with different actions when those features of its abilities are active.