Realization of the conscience mechanism in CMOS implementation of winner-takes-all self-organizing neural networks

Authors:
Rafał Długosz;Tomasz Talaśka;Witold Pedrycz;Ryszard Wojtyna
Affiliations:
Institute of Microtechnology, Swiss Federal Institute of Technology, Lausanne, Neuchâtel, Switzerland;Faculty of Telecommunication and Electrical Engineering, University of Technology and Life Sciences, Bydgoszcz, Poland;Department of Electrical and Computer Engineering, University of Alberta, Edmonton, AB, Canada and Systems Research Institute, Polish Academy of Science, Warsaw, Poland;Faculty of Telecommunication and Electrical Engineering, University of Technology and Life Sciences, Bydgoszcz, Poland and Academy of Information Technology, Institute of Computer Science, Ł& ...
Venue:
IEEE Transactions on Neural Networks
Year:
2010

Citing 13
Cited 2

Analog VLSI and neural systems

Analog VLSI and neural systems
Competitive learning algorithms for vector quantization

Neural Networks
Analog VLSI Circuits for Competitive Learning Networks

Analog Integrated Circuits and Signal Processing - Special issue: cellular neural networks and analog VLSI
Self-Organizing Maps and Learning Vector Quantization forFeature Sequences

Neural Processing Letters
VLSI-Compatible Immplementations for Artificial Neural Networks

VLSI-Compatible Immplementations for Artificial Neural Networks
Learning on Silicon: Adaptive VLSI Neural Systems

Learning on Silicon: Adaptive VLSI Neural Systems
Self-Organizing Maps

Self-Organizing Maps
Neural Network Initialization

IWANN '96 Proceedings of the International Workshop on Artificial Neural Networks: From Natural to Artificial Neural Computation
Comparison between analog and digital neural network implementations for range-finding applications

IEEE Transactions on Neural Networks
Low power current-mode binary-tree asynchronous Min/Max circuit

Microelectronics Journal
Frequency-based multilayer neural network with on-chip learning and enhanced neuron characteristics

IEEE Transactions on Neural Networks
New adaptive color quantization method based on self-organizing maps

IEEE Transactions on Neural Networks
A comparison between habituation and conscience mechanism in self-organizing maps

IEEE Transactions on Neural Networks

A programmable triangular neighborhood function for a Kohonen self-organizing map implemented on chip

Neural Networks
Low-voltage CMOS current-mode exponential circuit with 70dB output dynamic range

Microelectronics Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a complementary metal-oxide-semiconductor (CMOS) implementation of a conscience mechanism used to improve the effectiveness of learning in the winnertakes-all (WTA) artificial neural networks (ANNs) realized at the transistor level. This mechanism makes it possible to eliminate the effect of the so-called "dead neurons," which do not take part in the learning phase competition. These neurons usually have a detrimental effect on the network performance, increasing the quantization error. The proposed mechanism comes as part of the analog implementation of the WTA neural networks (NNs) designed for applications to ultralow power portable diagnostic devices for online analysis of ECGbiomedical signals. The study presents Matlab simulations of the network's model, discusses postlayout circuit level simulations and includes results of measurement completed for the physical realization of the circuit.