An iterative stochastic algorithm to perform maximum a posteriori parameter estimation of hidden Markov models is proposed. It makes the most of the statistical model by introducing an artiicial probability model based on an increasing number of the unobserved Markov chain at each iteration. Under minor regularity assumptions, we provide suucient conditions to ensure global convergence of this ...

Context-dependent modeling is a widely used technique for better phone modeling in continuous speech recognition. While different types of context-dependent models have been used, triphones have been known as the most effective ones. In this paper, a Maximum a Posteriori (MAP) estimation approach has been used to estimate the parameters of the untied triphone model set used in data-driven clust...

Bayesian methods to solve imaging inverse problems usually combine an explicit data likelihood function with a prior distribution that explicitly models expected properties of the solution. Many kinds priors have been explored in literature, from simple ones expressing local more involved exploiting image redundancy at non-local scale. In departure modelling, several recent works proposed and s...

In this paper, the problem of direction arrival estimation is addressed by employing Bayesian learning technique in sparse domain. This paper deals with inference (SBL) for both single measurement vector (SMV) and multiple (MMV) its applicability to estimate arriving signal’s at receiving antenna array; particularly considered be a uniform linear array. We also derive hyperparameter updating eq...

A Simulated Annealing method is presented for the solution of nonlinear time series estimation problems, by maximization of the a Posteriori Likelihood function. Homogeneous temperature annealing is proposed for smoothing problems and inhomogeneous temperature annealing for filtering problems. Both methods of annealing guarantee convergence to the Maximum A Posteriori Likelihood (MAP) estimate....

In this paper, we apply Constrained Maximum a Posteriori Linear Regression (CMAPLR) transformation on Universal Background Model (UBM) when characterizing each speaker with a supervector. We incorporate the covariance transformation parameters into the supervector in addition to the mean transformation parameters. Maximum Likelihood Linear Regression (MLLR) covariance transformation is adopted....

An instance crucial to most problems in signal processing is the selection of the order of a presupposed model. Examples are the determination of the putative number of signals present in white Gaussian noise or the number of noise-contaminated sources impinging on a passive sensor array. It is shown that Maximum a Posteriori Bayesian arguments, coupled with Maximum Entropy considerations, offe...

The presented paper addresses the problem of creating hidden Markov models for fast speech. The major issues discussed are robust parameter estimation and reducing within-model variations. Regarding the first issue, the use of the maximum a posteriori parameter estimation is discussed. To reduce within-model variations, a maximum likelihood based vocal tract length normalization procedure and a...