• DocumentCode
    2340401
  • Title

    Evolving creatures that can learn by imitation: apprentice behavior and its role in robot motor learning

  • Author

    Stoica, Adrian

  • Author_Institution
    Dept. of Electr. & Electron. Eng., Victoria Univ. of Technol., Melbourne, Vic., Australia
  • fYear
    1994
  • fDate
    7-9 Sept. 1994
  • Firstpage
    440
  • Lastpage
    443
  • Abstract
    The approach presented argues in favor of a methodology of evolving robotic systems that are fitted for learning from similar creatures in their environment, i.e. exhibit an apprentice behavior. An individual system that is adapted for learning from other fitted individuals performing in the same environment is proposed and it is suggested how such a system may be obtained through evolution. As a particular implementation of such a system, an anthropomorphic manipulator is considered, which is assumed to have gained this ability. Its motor development is outlined with the apprentice behavior supporting the transition from self learning of sensory coordination to teacher based learning.
  • Keywords
    intelligent control; learning (artificial intelligence); manipulators; anthropomorphic manipulator; apprentice behavior; learn by imitation; motor development; robot motor learning; robotic systems; sensory coordination; teacher based learning; Anthropomorphism; Artificial intelligence; Cognition; Cultural differences; Humans; Manipulators; Neural networks; Phylogeny; Robot sensing systems; Writing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    From Perception to Action Conference, 1994., Proceedings
  • Print_ISBN
    0-8186-6482-7
  • Type

    conf

  • DOI
    10.1109/FPA.1994.636142
  • Filename
    636142