• DocumentCode
    500773
  • Title

    How to make computers that work like the brain

  • Author

    George, Dileep

  • Author_Institution
    Numenta Inc., Redwood City, CA, USA
  • fYear
    2009
  • fDate
    26-31 July 2009
  • Firstpage
    420
  • Lastpage
    423
  • Abstract
    By using neuroanatomy and neurophysiology as a set of constraints, we believe that we have started to uncover how the brain uses hierarchy and time to create a model of the world, and to recognize novel patterns as part of that model. Hierarchically organized memory is fundamentally different than the linear memory used in current computers, and therefore offers the potential for new computer architectures. Today, we are exploring and advancing this technology by using traditional computer architectures (benefited by multiple CPU cores) to emulate the hierarchical structure of the neocortex. Exploiting the hierarchical temporal structure of the neocortex to build intelligent machines could open up many opportunities to rethink how integrated circuits and systems can play a leading role.
  • Keywords
    artificial intelligence; brain; integrated circuits; learning (artificial intelligence); memory architecture; neurophysiology; brain; computer architectures; hierarchically organized memory; integrated circuits; intelligent machines; neocortex; neuroanatomy; neurophysiology; Cities and towns; brain; hierarchical learning; neocortex; spatio-temporal learning;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Design Automation Conference, 2009. DAC '09. 46th ACM/IEEE
  • Conference_Location
    San Francisco, CA
  • ISSN
    0738-100X
  • Print_ISBN
    978-1-6055-8497-3
  • Type

    conf

  • Filename
    5227027