• DocumentCode
    2436912
  • Title

    The discovery-exploitation model and its computational implementation

  • Author

    Bowers, Henry M., III

  • Author_Institution
    Knowledge Eng. Team, FannieMae, USA
  • Volume
    5
  • fYear
    1997
  • fDate
    12-15 Oct 1997
  • Firstpage
    4448
  • Abstract
    The author introduces the Discovery-Exploitation (DE) Model of cognition. The DE model frames cognition as an extended form of function synthesis. It consists of a parameterized mapping M and two process stages that rely upon M: a discovery stage during which the values of the parameters of M are discovered (with or without human intervention) and an exploitation stage during which the exercising of M is performed using the values obtained during the discovery stage. The exploitation stage is said to “exploit” the generalization capabilities of M by inferring acceptable output values for input values that have never been seen before by the DE Model. The use of the DE Model as a guide for constructing computational implementations of cognition is demonstrated M is instantiated as a Fuzzy Inference System (FIS), the discovery stage is performed with an Evolutionary Search (ES) strategy that discovers input-output pairs required to identify the parameters of the FIS, and the exploitation stage is performed by passing novel input values through the FIS to produce the matching output value. This implementation is applied to the problem of robotic manipulator motor control for a three-jointed arm with movement in three dimensions. The author speculates that all other models of cognition can be subsumed by the DE Model. If this is so, then the DE Model provides a common descriptive framework which would facilitate discussion, comparison and integration of different cognitive models and implementations of those models
  • Keywords
    adaptive control; cognitive systems; fuzzy control; generalisation (artificial intelligence); intelligent control; learning systems; manipulators; position control; 3D movement; Evolutionary Search; Fuzzy Inference System; cognition; descriptive framework; discovery stage; discovery-exploitation model; exploitation stage; function synthesis; human intervention; input-output pairs; intelligent control; learning control; parameterized mapping; robotic manipulator motor control; three-jointed arm; Assembly; Cognition; Computational modeling; Equations; Fuzzy systems; Nervous system; Takagi-Sugeno-Kang model; Unsupervised learning;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Systems, Man, and Cybernetics, 1997. Computational Cybernetics and Simulation., 1997 IEEE International Conference on
  • Conference_Location
    Orlando, FL
  • ISSN
    1062-922X
  • Print_ISBN
    0-7803-4053-1
  • Type

    conf

  • DOI
    10.1109/ICSMC.1997.637547
  • Filename
    637547