Journal of the Korean Institute of Intelligent Systems (한국지능시스템학회논문지)

Korean Institute of Intelligent Systems (한국지능시스템학회)
권호 표지
DOI QR코드

DOI QR Code

CMOS neuron activation function

CMOS 뉴런의 활성화 함수

  • 강민제 (제주대학교 전기.전자공학부) ;
  • 김호찬 (제주대학교 전기.전자공학부) ;
  • 송왕철 (제주대학교 정보공학과) ;
  • 이상준 (제주대학교 정보공학과)
  • Published : 2006.10.25
Download PDF
⟨ Previous Next ⟩

Abstract

We have proposed the methods how to control the slope of CMOS inverter's characteristic and how to shift it in y axis. We control the MOS transistor threshold voltage for these methods. By observing that two transistors are in saturation region at the center of the CMOS inverter's characteristic, we have presented how to make the characteristic for one pole neuron. The circuit level simulation is used for verifying the proposed method. PSpice(OrCAD Co.) is used for circuit level simulation.

CMOS 인버터 특성곡선의 기울기를 조절하는 방법과 y축으로 이동할 수 있는 방법을 제안하였다. 기울기의 변경과 y축으로 이동은 트랜지스터의 문턱 값을 조절하는 방법을 사용하였다. 그리고 특성곡선의 중심에서는 두 트랜지스터 모두 포화영역에 머물러 있음에 착안하여, 단극성 뉴런의 특성곡선을 만드는 방법을 제안하였다. 제안된 방법은 회로레벨의 시뮬레이션을 통해 검증하였으며, 회로레벨의 시뮬레이션은 OrCAD사의 PSpice(Professional Simulation Program with Integrated Circuit Emphasis)를 사용하였다.

Keywords

References

  1. Hopfield, J. J., and D. W. Tank., 'Computing with Neural Circuits: A Model,', Science, vol 233, pp.625-633, 1986 https://doi.org/10.1126/science.3755256
  2. T. Higuchi and M. Kameyarna. 'Multiple-Valued digital Processing System'. Syokoda, Japan, 1989
  3. Zurada, J. M.: Introduction to Artificial Neural Systems. 1st edn. West, St. Paul MN. (1992)
  4. Anthony N. Michel, Jay A. Farrel and Wolfgang Porod, 1989, 'Qualitative Analysis of Neural Networks.' IEEE Trans. Circuits Syst., vol 36, pp. 229-243, February 1989 https://doi.org/10.1109/31.20200
  5. G. Bogason, 'Generation of a Neuron Transfer Function and its Derivatives', Electron. Lett., vol.29, pp.1867-1869, 1993 https://doi.org/10.1049/el:19931243
  6. A. J. Annema, 'Hardware Realization of a Neuron Transfer Function and its Derivatives', Electron. Lett., vol.30, pp.576-577, 1994 https://doi.org/10.1049/el:19940375
  7. J,Shen, K. Tanno, and O. Ishizuka. 'Design and Analysis of down literal circuit using neuron-MOS transistors'. In Proc. 12th MVL Research Society of Japan, pp. 1-9, 1999
  8. Amit K Gupta, and Navakanta Bhat, 'Hardware Realization of a Digitally Controllable Neuron Activation Function and its Derivative for Extremely Low Power Application,' IEICE Trans. Fundamentals. vol. E82-A, pp.1582-1587, August 1999
  9. Massimo Conti, 'Analog CMOS Implementation of Approximate Identity Neural Networks,' IEICE Trans. Fundamentals. vol. E80 A, pp.427-433, February 1997
  10. Motoi Inaba, Koichi Tanno, 'Analog Inverter with Neuron-MOS Transistors and Its Application,' IEICE Trans. Fundamentals. vol. E85-A, pp.366-372, February 2002
  11. Bo Liu; Frenzel, J.F., 'A CMOS neuron for VLSI circuit implementation of pulsed neural networks,' IECON 02 vol. 4, pp.3182-3185, Nov. 2002 https://doi.org/10.1109/IECON.2002.1182906