View Full Document

Neural Nets for Adaptive Filtering and Adaptive Pattern Recognition



View the full content.
View Full Document
View Full Document

18 views

Unformatted text preview:

Neural Nets for Adaptive Filtering and Adaptive Pattern Recognition Bernard Widrow Stanford University Rodney Winter United States Air Force he fields of adaptive signal processing and adaptive neural networks have been developing independently but have the adaptive linear combiner ALC in common With its inputs connected to a tapped delay line the ALC becomes a key component of an adaptive filter With its output connected to a quantizer the ALC becomes an adaptive threshold element or adaptive neuron Adaptive filters have enjoyed great commercial success in the signal processing field All high speed modems now use adaptive equalization filters Longdistance telephone and satellite communications links are being equipped with adaptive echo cancelers to filter out echo allowing simultaneous two way communications Other applications include noise canceling and signal prediction Adaptive threshold elements on the other hand are the building blocks of neural networks Today neural nets are the focus of widespread research interest Areas of investigation include pattern recognition and trainable logic Neural network systems have not yet had the commercial impact of adaptive filtering The commonality of the ALC to adaptive signal processing and adaptive neural networks suggests the two fields have much to share with each other This article describes practical applications of the ALC in signal processing and pattern recognition March 1988 A new multilayer adaptation algorithm that descrambles output and reproduces original patterns is advancing the practicality of neuralnetwork patternrecognition systems The adaptive linear combiner The ALC shown in Figure 1 is the basic building block for most adaptive systems The output is a linear combination of the many input signals The weighting coefficients comprise a weight vector The input signals comprise an input signal vector The output signal is the inner product or dot product of the input signal vector with 00l8 9162 88 0300



Access the best Study Guides, Lecture Notes and Practice Exams

Loading Unlocking...
Login

Join to view Neural Nets for Adaptive Filtering and Adaptive Pattern Recognition and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Neural Nets for Adaptive Filtering and Adaptive Pattern Recognition and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?