DocumentCode
2816135
Title
Spectral noise-shaping in integrate-and-fire neural networks
Author
Adams, Robert W.
Author_Institution
Analog Devices Semicond., Wilmington, MA, USA
fYear
1997
fDate
19-22, Oct 1997
Abstract
A theory of coordinated neuronal firing events is proposed that allows the low-noise transmission of analog signals through a network of coupled neurons. The inherent quantization noise of a group of neurons can be shaped in the frequency domain in such a way as to provide a high signal-to-noise ratio over some specified signal bandwidth. This shaping is accomplished by using neural interconnections that resemble lateral inhibition
Keywords
analogue-digital conversion; neural chips; neural nets; noise; quantisation (signal); signal processing; spectral analysis; statistical analysis; ADC; analog signals; coordinated neuronal firing events; coupled neurons network; frequency domain; high signal-to-noise ratio; integrate-and-fire neural networks; lateral inhibition; low-noise transmission; neural interconnections; quantization noise; signal bandwidth; spectral noise-shaping; Acoustic noise; Autocorrelation; Frequency; Intelligent networks; Neural networks; Neurons; Noise shaping; Quantization; Semiconductor device noise; Signal to noise ratio;
fLanguage
English
Publisher
ieee
Conference_Titel
Applications of Signal Processing to Audio and Acoustics, 1997. 1997 IEEE ASSP Workshop on
Conference_Location
New Paltz, NY
Print_ISBN
0-7803-3908-8
Type
conf
DOI
10.1109/ASPAA.1997.625619
Filename
625619
Link To Document