Simulated annealing and Boltzmann machines: a stochastic approach to combinatorial optimization and neural computing
| Hi-index | 0.00 |
The discrete wavelet transform is a multi-scale signal representation. Here it is shown that coarse form is equivalent to the approximated output and fine part to the differentially fed output or hyper planes. The wave let coefficients represent differential feedback planes. In this paper an overview of multiresolution analysis in differentially fed neural networks is given. It has been shown that the network closely resembles multiresolution of wavelet transform.