Several mismatch conditions can be modeled as an additive bias. This bias is considered independent of the observation vectors, although this approximation is not always accurate. In this paper the dependence of the bias on the observation vectors is taken into consideration in the context of compensating the GSM coding distortion in speech recognition. However, the results presented here can e...

High-throughput sequencing makes possible to process samples containing multiple genomic sequences and then estimate their frequencies or even assemble them. The maximum likelihood estimation of frequencies of the sequences based on observed reads can be efficiently performed using expectationmaximization (EM) method assuming that we know sequences present in the sample. Frequently, such knowle...

Recent developments in inference algorithms based on stochastic Expectationmaximization or stochastic cellular automata (SCA) have made it possible to employ a variety of randomized data structures that are unavailable to the dominant inference methods in the Bayesian toolkit, including collapsed Gibbs sampling and stochastic variational inference (SVI). Equipped with this recent capability, we...

In this paper we address the problem of estimating the parameters of a Gaussian mixture model. Although the EM (ExpectationMaximization) algorithm yields the maximum-likelihood solution it has many problems: (i) it requires a careful initialization of the parameters; (ii) the optimal number of kernels in the mixture may be unknown beforehand. We propose a criterion based on the entropy of the p...

Representing the image to be inpainted in an appropriate sparse representation dictionary, and combining elements from Bayesian statistics and modern harmonic analysis, we introduce an expectationmaximization (EM) algorithm for image inpainting and interpolation. From a statistical point of view, the inpainting/interpolation can be viewed as an estimation problem with missing data. Towards this...

We study properties and parameter estimation of finite-state homogeneous continuous-time bivariate Markov chains. Only one of the two processes of the bivariate Markov chain is observable. The general form of the bivariate Markov chain studied here makes no assumptions on the structure of the generator of the chain, and hence, neither the underlying process nor the observable process is necessa...

We propose an unsupervised, probabilistic method for learning visual feature hierarchies. Starting from local, low-level features computed at interest point locations, the method combines these primitives into high-level abstractions. Our appearance-based learning method uses local statistical analysis between features and ExpectationMaximization (EM) to identify and code spatial correlations. ...

We introduce the truncated Gaussian graphical model (TGGM) as a novel framework for designing statistical models for nonlinear learning. A TGGM is a Gaussian graphical model (GGM) with a subset of variables truncated to be nonnegative. The truncated variables are assumed latent and integrated out to induce a marginal model. We show that the variables in the marginal model are non-Gaussian distr...

This paper addresses the problem of detecting anomalous interactions or traffic within a very large network using a limited number of unlabeled observations. In particular, consider n recorded interactions among p nodes, where p may be very large relative to n. A novel method based on using a hypergraph representation of the data is proposed to deal with this very high-dimensional, “big p, smal...

Log analysis is a method to identify intrusions at the host or network level by scrutinizing the log events recorded by the operating systems, applications, and devices. Most work contemplates a single type of log for analysis, leading to an unclear picture of the situation and difficulty in deciding the existence of an intrusion. Moreover, most existing detection methods are knowledge-dependen...