Search

A new blind equalization approach that aims to force the probability density function (PDF) at the equalizer output to match the known constellation PDF is presented. The proposed method relies on the Parzen window method to estimate the data PDF and is implemented by a stochastic gradient descent algorithm.

The suitability of a class of Parzen-window-based entropy estimates, namely Renyi's entropy, as a criterion for blind deconvolution of linear channels are investigated. The results indicate that this nonparametric entropy estimation approach outperforms the standard Bell-Sejnowski and normalized kurtosis algorithms in blind deconvolution.

A new method for example-dependent cost (EDC) classification is proposed. The method constitutes an extension of a recently introduced training algorithm for neural networks. The surrogate cost function is an estimate of the Bayesian risk, where the estimates of the conditional probabilities for each class are defined in terms of a 1-D Parzen window estimator of the output of (discriminative) neural networks. This probability density is modeled with the objective of allowing an easy minimization of a sampled version of the Bayes risk. The conditional probabilities included in the definition of the risk are not explicitly estimated, but the risk is minimized by a gradient-descent algorithm. The proposed method has been evaluated using linear classifiers and neural networks, with both shallow (a single hidden layer) and deep (multiple hidden layers) architectures. The experimental results show the potential and flexibility of the proposed method, which can handle EDC classification under imbalanced data situations that commonly appear in this kind of problems. [ABSTRACT FROM AUTHOR]

We consider a distributed detection system formed by a large number of local detectors and a data fusion center that performs a Neyman–Pearson fusion of the binary quantizations of the sensor observations. In the analyzed two-stage detection system the local decisions are taken with no kind of cooperation among the devices and they are transmitted to the fusion center over an error free parallel access channel. In addition, the sensors are randomly deployed along a straight line, and the corresponding sensor spacings are drawn independently from a common probability density function (pdf). For both hypothesis, \mbi H\bf 0 and \mbi H\bf 1, depending on the correlation structure of the observed phenomenon the local decisions might be dependent. In the case of being dependent, their correlation structure is modelled with a one-dimensional Markov random field with nearest neighbor dependency and binary state space. Under this scenario, we first derive a closed-form error exponent for the Neyman–Pearson fusion of the local decisions when the involved data fusion center only knows the distribution of the sensor spacings. Second, based on a single parameter that captures the mean correlation strength among the local decisions, some analytical properties of the error exponent are investigated. Finally, we develop a physical model for the conditional probabilities of the Markov random fields that might be present under each hypothesis. Using this model we characterize the error exponent for two well-known models of the sensor spacing: i) equispaced sensors with failures, and ii) exponentially spaced sensors with failures. [ABSTRACT FROM AUTHOR]

Provides information on a study in which a smoothed piecewise-linear model is used as a tool in the device modeling field. Introduction to metal semiconductor field effect transistors; Information on the model training; Conclusions.