Towards Brain-inspired Interconnects and Circuits

Research Article

Reduced Interconnects in Neural Networks Using a Time Multiplexed Architecture Based on Quantum Devices

Download90 downloads
  • @INPROCEEDINGS{10.1007/978-3-642-04850-0_32,
        author={Peter Kelly and Fergal Tuffy and Valeriu Beiu and Liam McDaid},
        title={Reduced Interconnects in Neural Networks Using a Time Multiplexed Architecture Based on Quantum Devices},
        proceedings={Towards Brain-inspired Interconnects and Circuits},
        proceedings_a={TBIC},
        year={2012},
        month={5},
        keywords={Quantum device resonant tunneling device (RTD) time multiplexed architecture (TMA) neural network (NN) interconnects},
        doi={10.1007/978-3-642-04850-0_32}
    }
    
  • Peter Kelly
    Fergal Tuffy
    Valeriu Beiu
    Liam McDaid
    Year: 2012
    Reduced Interconnects in Neural Networks Using a Time Multiplexed Architecture Based on Quantum Devices
    TBIC
    Springer
    DOI: 10.1007/978-3-642-04850-0_32
Peter Kelly1,*, Fergal Tuffy1,*, Valeriu Beiu2,*, Liam McDaid1,*
  • 1: University of Ulster
  • 2: United Arab Emirates University
*Contact email: pm.kelly@ulster.ac.uk, f.tuffy@ulster.ac.uk, vbeiu@uaeu.ac.ae, lj.mcdaid@ulster.ac.uk

Abstract

The interconnection problem associated with large scale hardware-based neural networks is well known. A time multiplexed neural network architecture using silicon based quantum devices with MOS/CMOS devices is described and shows significant increased functional density compared to conventional devices.