Hardware implementation of artificial neural networks for arbitrary boolean functions with generalised threshold gate circuits

  • Authors:
  • Maciej Nikodem

  • Affiliations:
  • Wrocław University of Technology, Institute of Computers, Control and Robotics, Wrocław, Poland

  • Venue:
  • MICAI'10 Proceedings of the 9th Mexican international conference on Artificial intelligence conference on Advances in soft computing: Part II
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper describes nanocircuits that draw on negative differential resistance and are capable of implementing complex threshold functions in a single gate structure. Due to nanometer dimensions, high operational frequencies and low power consumption these devices can be used for efficient hardware realisation of artificial neural networks (ANNs). We present state of the art in development of such circuit and focus on Generalised Threshold Gates (GTGs) that are capable of implementing arbitrary Boolean functions in a single gate structure. Algorithm for implementing Boolean functions outputs circuits with predefined weights and thresholds. This enables to construct application specific ANNs and eliminates the requirement for network learning when this kind of gates are used.