Starting from maximizing information ow through a nonlinear neuron Bell and Sejnowski derived adaptation equations for blind deconvolution using an FIR lter [1]. In this paper we will derive a simpler form of the adaptation and we will apply it to more complex lter structures, such as recursive lters. As an application, we study blind echo cancellation for speech signals. We will also present a...

I consider a model used in many communication systems where the sum of n stochastic sinusoidal signals, each with general random amplitudes as well as phase angles is present. The exact probability distribution of the resulting signal’s amplitude is of particular interest in many important applications, however the general problem has remained open. I use probabilistic methods to prove new form...

This paper proposes a novel cluster modeling method for intercluster distance measurement within the framework of agglomerative hierarchical speaker clustering, namely, incremental Gaussian mixture cluster modeling. This method uses a single Gaussian distribution to model each initial cluster, but represents any newly merged cluster using a distribution whose pdf is the weighted sum of the pdf’...

We derive the expression of an optimum non-Gaussian radar detector from the non-Gaussian spherically invariant random process (SIRP) clutter model and a bayesian estimator of the SIRP characteristic density. SIRP modelizes non-Gaussian process as a complex Gaussian process whose variance, the so-called texture, is itself a positive random variable (r.v.). After performing a bayesian estimation ...

The experiments of Leptos et al. [1] show that the displacements of small particles affected by swimming microorganisms achieve a non-Gaussian distribution, which nevertheless scales diffusively. We use a simple model where the particles undergo repeated ‘kicks’ due to the swimmers to explain the shape of the distribution. Leptos et al. study the microscopic algae Chlamydomonas reinhardthii. Th...

In this paper the design of semi-continuous segmental probability models (SCSPMs) in large vocabulary continuous speech recognition is presented. The tied Gaussian densities are trained using data from all states of all utterances while the mixture weights are estimated using data from the state being trained individually. The SCSPMs tie all the densities of all states from all Speech Recogniti...

The properties of the probability distribution function of the cosmological continuous density field are studied. We present further developments and compare dynamically motivated methods to derive the PDF. One of them is based on the Zel’dovich approximation (ZA). We extend this method for arbitrary initial conditions, regardless of whether they are Gaussian or not. The other approach is based...

The problem of acoustic modeling for continuous speech recognition is addressed. To deal with coarticulation effects and interspeaker variability, an extension of the Mixture Stochastic Trajectory Model (MSTM) is proposed. MSTM is a segment-based model using phonemes as speech units. In MSTM, the observations of a phoneme are modeled by a set of stochastic trajectories. The trajectories are mod...

To many, high-frequency (HF) radio communications is obsolete in this age of longdistance satellite communications and undersea optical fiber. Yet despite this, the HF band is used by defense agencies for backup communications and spectrum surveillance, and is monitored by spectrum management organizations to enforce licensing. Such activity usually requires systems capable of locating distant ...