Dirichlet Process Mixture of Generalized Inverted Dirichlet Distributions for Positive Vector Data With Extended Variational Inference.

A Bayesian nonparametric approach for estimation of a Dirichlet process (DP) mixture of generalized inverted Dirichlet distributions [i.e., an infinite generalized inverted Dirichlet mixture model (InGIDMM)] has been proposed. The generalized inverted Dirichlet distribution has been proven to be efficient in modeling the vectors that contain only positive elements. Under the classical variational inference (VI) framework, the key challenge in the Bayesian estimation of InGIDMM is that the expectation of the joint distribution of data and variables cannot be explicitly calculated. Therefore, numerical methods are usually applied to simulate the optimal posterior distributions. With the recently proposed extended VI (EVI) framework, we introduce lower bound approximations to the original variational objective function in the VI framework such that an analytically tractable solution can be derived. Hence, the problem in numerical simulation has been overcome. By applying the DP mixture technique, an InGIDMM can automatically determine the number of mixture components from the observed data. Moreover, the DP mixture model with an infinite number of mixture components also avoids the problems of underfitting and overfitting. The performance of the proposed approach is demonstrated with both synthesized data and real-life data applications.

Variational Bayesian Learning for Dirichlet Process Mixture of Inverted Dirichlet Distributions in Non-Gaussian Image Feature Modeling.

In this paper, we develop a novel variational Bayesian learning method for the Dirichlet process (DP) mixture of the inverted Dirichlet distributions, which has been shown to be very flexible for modeling vectors with positive elements. The recently proposed extended variational inference (EVI) framework is adopted to derive an analytically tractable solution. The convergency of the proposed algorithm is theoretically guaranteed by introducing single lower bound approximation to the original objective function in the EVI framework. In principle, the proposed model can be viewed as an infinite inverted Dirichlet mixture model that allows the automatic determination of the number of mixture components from data. Therefore, the problem of predetermining the optimal number of mixing components has been overcome. Moreover, the problems of overfitting and underfitting are avoided by the Bayesian estimation approach. Compared with several recently proposed DP-related methods and conventional applied methods, the good performance and effectiveness of the proposed method have been demonstrated with both synthesized data and real data evaluations.

Scatter Component Analysis: A Unified Framework for Domain Adaptation and Domain Generalization.

This paper addresses classification tasks on a particular target domain in which labeled training data are only available from source domains different from (but related to) the target. Two closely related frameworks, domain adaptation and domain generalization, are concerned with such tasks, where the only difference between those frameworks is the availability of the unlabeled target data: domain adaptation can leverage unlabeled target information, while domain generalization cannot. We propose Scatter Component Analyis (SCA), a fast representation learning algorithm that can be applied to both domain adaptation and domain generalization. SCA is based on a simple geometrical measure, i.e., scatter, which operates on reproducing kernel Hilbert space. SCA finds a representation that trades between maximizing the separability of classes, minimizing the mismatch between domains, and maximizing the separability of data; each of which is quantified through scatter. The optimization problem of SCA can be reduced to a generalized eigenvalue problem, which results in a fast and exact solution. Comprehensive experiments on benchmark cross-domain object recognition datasets verify that SCA performs much faster than several state-of-the-art algorithms and also provides state-of-the-art classification accuracy in both domain adaptation and domain generalization. We also show that scatter can be used to establish a theoretical generalization bound in the case of domain adaptation.

Identification of human sympathetic neurovascular control using multivariate wavelet decomposition analysis.

The dynamic regulation of cerebral blood flow (CBF) is thought to involve myogenic and chemoreflex mechanisms, but the extent to which the sympathetic nervous system also plays a role remains debated. Here we sought to identify the role of human sympathetic neurovascular control by examining cerebral pressure-flow relations using linear transfer function analysis and multivariate wavelet decomposition analysis that explicitly accounts for the confounding effects of dynamic end-tidal Pco2 (PetCO2 ) fluctuations. In 18 healthy participants randomly assigned to the α1-adrenergic blockade group (n = 9; oral Prazosin, 0.05 mg/kg) or the placebo group (n = 9), we recorded blood pressure, middle cerebral blood flow velocity, and breath-to-breath PetCO2 Analyses showed that the placebo administration did not alter wavelet phase synchronization index (PSI) values, whereas sympathetic blockade increased PSI for frequency components ≤0.03 Hz. Additionally, three-way interaction effects were found for PSI change scores, indicating that the treatment response varied as a function of frequency and whether PSI values were PetCO2 corrected. In contrast, sympathetic blockade did not affect any linear transfer function parameters. These data show that very-low-frequency CBF dynamics have a composite origin involving, not only nonlinear and nonstationary interactions between BP and PetCO2 , but also frequency-dependent interplay with the sympathetic nervous system.

Minimum Entropy Rate Simplification of Stochastic Processes.

We propose minimum entropy rate simplification (MERS), an information-theoretic, parameterization-independent framework for simplifying generative models of stochastic processes. Applications include improving model quality for sampling tasks by concentrating the probability mass on the most characteristic and accurately described behaviors while de-emphasizing the tails, and obtaining clean models from corrupted data (nonparametric denoising). This is the opposite of the smoothing step commonly applied to classification models. Drawing on rate-distortion theory, MERS seeks the minimum entropy-rate process under a constraint on the dissimilarity between the original and simplified processes. We particularly investigate the Kullback-Leibler divergence rate as a dissimilarity measure, where, compatible with our assumption that the starting model is disturbed or inaccurate, the simplification rather than the starting model is used for the reference distribution of the divergence. This leads to analytic solutions for stationary and ergodic Gaussian processes and Markov chains. The same formulas are also valid for maximum-entropy smoothing under the same divergence constraint. In experiments, MERS successfully simplifies and denoises models from audio, text, speech, and meteorology.

Detection of Impaired Sympathetic Cerebrovascular Control Using Functional Biomarkers Based on Principal Dynamic Mode Analysis.

This study sought to determine whether models of cerebrovascular function based on Laguerre-Volterra kernels that account for nonlinear cerebral blood flow (CBF) dynamics can detect the effects of functional cerebral sympathetic blockade. We retrospectively analyzed continuous beat-to-beat blood pressure, middle cerebral blood velocity, and partial-pressure of end-tidal CO2 (PETCO2) recordings from eighteen healthy individuals who were treated with either an oral dose of the α1-adrenergic receptor blocker Prazosin or a placebo treatment. The global principal dynamic modes (PDMs) were analyzed using Laguerre-Volterra kernels to examine the nonlinear system dynamics. Our principal findings were: (1) very low frequency (<0.03 Hz) linear components of first-order kernels for BP and PETCO2 are mutually coupled to CBF dynamics with the ability to separate individuals between control and blockade conditions, and (2) the gains of the nonlinear functions associated with low-pass and ≈0.03 Hz global PDMs for the BP are sensitive to sympathetic blockade. Collectively these results suggest that very low frequency global PDMs for BP may have potential utility as functional biomarkers of sympathetic neurovascular dysfunction which can occur in conditions like autonomic failure, stroke and traumatic brain injury.

Non-Linear Characterisation of Cerebral Pressure-Flow Dynamics in Humans.

Cerebral metabolism is critically dependent on the regulation of cerebral blood flow (CBF), so it would be expected that vascular mechanisms that play a critical role in CBF regulation would be tightly conserved across individuals. However, the relationships between blood pressure (BP) and cerebral blood velocity fluctuations exhibit inter-individual variations consistent with heterogeneity in the integrity of CBF regulating systems. Here we sought to determine the nature and consistency of dynamic cerebral autoregulation (dCA) during the application of oscillatory lower body negative pressure (OLBNP). In 18 volunteers we recorded BP and middle cerebral artery blood flow velocity (MCAv) and examined the relationships between BP and MCAv fluctuations during 0.03, 0.05 and 0.07Hz OLBNP. dCA was characterised using project pursuit regression (PPR) and locally weighted scatterplot smoother (LOWESS) plots. Additionally, we proposed a piecewise regression method to statistically determine the presence of a dCA curve, which was defined as the presence of a restricted autoregulatory plateau shouldered by pressure-passive regions. Results show that LOWESS has similar explanatory power to that of PPR. However, we observed heterogeneous patterns of dynamic BP-MCAv relations with few individuals demonstrating clear evidence of a dCA central plateau. Thus, although BP explains a significant proportion of variance, dCA does not manifest as any single characteristic BP-MCAv function.

A Method of Speech Periodicity Enhancement Using Transform-domain Signal Decomposition.

Periodicity is an important property of speech signals. It is the basis of the signal's fundamental frequency and the pitch of voice, which is crucial to speech communication. This paper presents a novel framework of periodicity enhancement for noisy speech. The enhancement is applied to the linear prediction residual of speech. The residual signal goes through a constant-pitch time warping process and two sequential lapped-frequency transforms, by which the periodic component is concentrated in certain transform coefficients. By emphasizing the respective transform coefficients, periodicity enhancement of noisy residual signal is achieved. The enhanced residual signal and estimated linear prediction filter parameters are used to synthesize the output speech. An adaptive algorithm is proposed for adjusting the weights for the periodic and aperiodic components. Effectiveness of the proposed approach is demonstrated via experimental evaluation. It is observed that harmonic structure of the original speech could be properly restored to improve the perceptual quality of enhanced speech.

Graph-preserving sparse nonnegative matrix factorization with application to facial expression recognition.

In this paper, a novel graph-preserving sparse nonnegative matrix factorization (GSNMF) algorithm is proposed for facial expression recognition. The GSNMF algorithm is derived from the original NMF algorithm by exploiting both sparse and graph-preserving properties. The latter may contain the class information of the samples. Therefore, GSNMF can be conducted as an unsupervised or a supervised dimension reduction method. A sparse representation of the facial images is obtained by minimizing the l(1)-norm of the basis images. Furthermore, according to the graph embedding theory, the neighborhood of the samples is preserved by retaining the graph structure in the mapped space. The GSNMF decomposition transforms the high-dimensional facial expression images into a locality-preserving subspace with sparse representation. To guarantee convergence, we use the projected gradient method to calculate the nonnegative solution of GSNMF. Experiments are conducted on the JAFFE database and the Cohn-Kanade database with unoccluded and partially occluded facial images. The results show that the GSNMF algorithm provides better facial representations and achieves higher recognition rates than nonnegative matrix factorization. Moreover, GSNMF is also more robust to partial occlusions than other tested methods.

Auditory-model based robust feature selection for speech recognition.

It is shown that robust dimension-reduction of a feature set for speech recognition can be based on a model of the human auditory system. Whereas conventional methods optimize classification performance, the proposed method exploits knowledge implicit in the auditory periphery, inheriting its robustness. Features are selected to maximize the similarity of the Euclidean geometry of the feature domain and the perceptual domain. Recognition experiments using mel-frequency cepstral coefficients (MFCCs) confirm the effectiveness of the approach, which does not require labeled training data. For noisy data the method outperforms commonly used discriminant-analysis based dimension-reduction methods that rely on labeling. The results indicate that selecting MFCCs in their natural order results in subsets with good performance.

Squared error as a measure of perceived phase distortion.

Based on two well-known auditory models, it is investigated whether the squared error between an original signal and a phase-distorted signal is a perceptually relevant measure for distortions in the Fourier phase spectrum of periodic signals obtained from speech. Both the performance of phase vector quantizers and the direct relationship between the squared error and two perceptual distortion measures are studied. The results indicate that for small values the squared error correlates well to the perceptual measures. However, for large errors, an increase in squared error does not, on average, lead to an increase in the perceptual measures. Empirical rate-perceptual distortion curves and listening tests confirm that, for low to medium codebook sizes, the average perceived distortion does not decrease with increasing codebook size when the squared error is used as encoding criterion.