• DocumentCode
    626845
  • Title

    Analog implementation of encoded neural networks

  • Author

    Larras, Benoit ; Lahuec, Cyril ; Arzel, Matthieu ; Seguin, Fabrice

  • Author_Institution
    Technopole Brest Iroise, Electron. Dept., Telecom Bretagne, Brest, France
  • fYear
    2013
  • fDate
    19-23 May 2013
  • Firstpage
    1612
  • Lastpage
    1615
  • Abstract
    Encoded neural networks mix the principles of associative memories and error-correcting decoders. Their storage capacity has been shown to be much larger than Hopfield Neural Networks´. This paper introduces an analog implementation of this new type of network. The proposed circuit has been designed for the 1V supply ST CMOS 65nm process. It consumes 1165 times less energy than a digital equivalent circuit while being 2.7 times more efficient in terms of combined speed and surface.
  • Keywords
    CMOS analogue integrated circuits; content-addressable storage; decoding; error correction codes; neural nets; Hopfield neural networks; ST CMOS process; analog implementation; associative memories; encoded neural networks; error-correcting decoders; size 65 nm; storage capacity; voltage 1 V; Analog circuits; Associative memory; Biological neural networks; CMOS integrated circuits; Digital circuits; Power demand;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Circuits and Systems (ISCAS), 2013 IEEE International Symposium on
  • Conference_Location
    Beijing
  • ISSN
    0271-4302
  • Print_ISBN
    978-1-4673-5760-9
  • Type

    conf

  • DOI
    10.1109/ISCAS.2013.6572170
  • Filename
    6572170