This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: An egocentric approach to machine intelligence R. Pienaar J.J. Kruger Council for Scient. and Indus. Research Dept. of Electrical and Electronic Eng. PO Box 395 University of Pretoria Pretoria, South Africa, 0001 Pretoria, South Africa Abstract Intimate to the functioning and behavior of intelligent systems is the manner in which information is represented internally. The conventional approach to intelligent system design assumes a particular bias in the manner by which this information is represented. Typically, this is char- acterized by an abstract or objective design methodol- ogy which holds that intelligence is not a function of the physical nature of the system. Such an approach suffers from several shortcomings, most notably problems relating to scaling and complexity. Recent physiological research, however, has demonstrated that physical bodily form is a fundamental building block in the organization of mam- malian cortical structures. Consequently, this article ex- plores such a biologically motivated subjective or ego- centric approach to system design, and demonstrates its utility in a simple robot arm control problem. 1 Introduction Fundamental to intelligence, be it machine or biolog- ical, is the method by which information is organized in- ternally to create internal representations. These internal representations are in turn processed by the system as it satisfies its goal conditions. The nature of these representations are fundamental to any processing system, comprising its most basic com- ponent. This article argues that most conventional ap- proaches to intelligent system design assume a particular objective bias in constructing these internal represen- tations, and are consequently plagued with scaling and complexity issues when moving from trivial to real world domains. The article will offer a novel design method with a different underlying subjective bias, presented as a possible means with which to address the potential short- comings endemic in the conventional approach. Of course, the field of robotics spans in and of itself a wide range of views and approaches, from the mono- lithic philosophy of good old fashioned artificial intelli- gence, GOFAI, through to the distributed architectures of Brookss behavior based robotics , . Such behav- ior based systems are characterized by the interaction of competing sub-systems, each responsible for a particular motor behavior. Brooks, in addressing issues surrounding scaling such robots to humanoid-level systems, defines cognitive robotics, which expand upon behavior based systems in several areas, viz bodily form, motivation, co- herence, self adaption, development, and physiological inspiration ....
View Full Document
This note was uploaded on 08/25/2011 for the course EGN 3060c taught by Professor Sukthankar,g during the Fall '08 term at University of Central Florida.
- Fall '08