We introduce a new Blind Source Separation Approach called binNTF which operates on tensor-valued binary datasets. Assuming that several simultaneously acting sources or elementary causes are generating the observed data, the objective of our approach is to uncover the underlying sources as well as their individual contribution to each observation with a minimum number of assumptions in an unsupervised fashion. We motivate, develop and demonstrate our method in the context of binary wafer test data which evolve during microchip fabrication. In this application, we also have to deal with incomplete datasets which can occur due to the commonly used stop-on-first-fail testing procedure or result from the aggregation of several distinct tests into BIN categories.

A novel structure learning algorithm for Bayesian Networks based on the Phyasrum Solver is introduced. First, the algorithm calculates pair wise correlation coefficients in the dataset. Within an initially fully connected Physarum-Maze, the length of the connections is given by the inverse correlation coefficient between the connected nodes. Then, the shortest indirect paths between each two nodes is determined using the Physarum Solver. In each iteration, a score of the surviving edges is increased. Based on that score, the highest ranked connections are combined to form a Bayesian Network. The novel Physarum Learner method is evaluated with different configurations and compared to the LAGD Hill Climber showing comparable performance regarding the quality of training results and increased time efficiency for large datasets.

In this paper we propose a novel method for brain SPECT image feature extraction based on the Empirical Mode Decomposition (EMD). The proposed method applied to assist the diagnosis of Alzheimer Disease (AD) selects the most discriminant voxels for classification from the transformed EMD feature space. In particular, high-frequency components of the EMD transformation are found to retain regional differences in functional activity which is characteristic of AD. The EMD represents a fully data-driven, unsupervised and additive signal decomposition and does not need any a priori defined basis system. Several experiments were carried out on a balanced SPECT database collected from the “Virgen de las Nieves” Hospital in Granada (Spain), containing 96 recordings and yielding up to 100% accuracy in separating AD and normal controls, and then outperforming recently proposed CAD systems.

Biomedical signals are in general non-linear and non-stationary which renders them difficult to analyze with classical time series analysis techniques. Empirical Mode Decomposition (EMD) in conjunction with a Hilbert spectral transform, together called Hilbert-Huang Transform, is ideally suited to extract informative components which are characteristic of underlying biological or physiological processes. The method is fully adaptive and generates a complete set of orthogonal basis functions, called Intrinsic Mode Functions (IMFs), in a purely data-driven manner. Amplitude and frequency of IMFs may vary over time which renders them different from conventional basis systems and ideally suited to study non-linear and non-stationary time series. However, biomedical time series are often recorded over long time periods. This generates the need for efficient EMD algorithms which can analyze the data in real time. No such algorithms yet exist which are robust, efficient and easy to implement. The contribution shortly reviews the technique of EMD and related algorithms and develops an on-line variant, called Sliding Empirical Mode Decomposition (SEMD), which is shown to perform well on large scale time series.

A novel network based on Linsker-type Hebbian learning is analyzed in its dynamical behavior. The network combines a coupled dynamics of fast and slow states and is prone to internal parametrical fluctuations as well as external noises. Robustness represents a crucial property of the network to attenuate the effects of internal fluctuation and external noise. In this study, we formulate this novel neural network as a coupled nonlinear differential systems operating at different time-scales under vanishing perturbations. We determine conditions for the existence of a global uniform attractor of the perturbed biological system. By using a Lyapunov function for the coupled system, we derive a maximal upper bound for the fast time scale associated with the fast state. Finally, two examples are given to confirm the applicability of the developed theoretical framework.

Due to external stimuli, biomedical signals are in general non-linear and non-stationary. Intelligent signal processing is crucial to unravel the information content buried in biomedical time series. Empirical Mode Decomposition is ideally suited to extract all pure oscillatory modes which are contained in the signal. These modes, called Intrinsic Mode Functions (IMFs), represent a complete set of locally orthogonal basis functions with time-varying amplitude and frequency. The contribution discusses the application of an online variant, called SEMD, to non-stationary biomedical time series recorded during neuromonitoring.

Due to external stimuli, biomedical signals are in general non-linear and non-stationary. Empirical Mode Decomposition in conjunction with a Hilbert spectral transform, together called Hilbert-Huang Transform, is ideally suited to extract essential components which are characteristic of the underlying biological or physiological processes. The method is fully adaptive and generates the basis to represent the data solely from these data and based on them. The basis functions, called Intrinsic Mode Functions (IMFs) represent a complete set of locally orthogonal basis functions whose amplitude and frequency may vary over time. The contribution reviews the technique of EMD and related algorithms and discusses illustrative applications.

In this letter, we propose a novel least-mean-square (LMS) algorithm for filtering speech sounds in the adaptive noise cancellation (ANC) problem. It is based on the minimization of the squared Euclidean norm of the difference weight vector under a stability constraint defined over theaposterioriestimation error. To this purpose, the Lagrangian methodology has been used in order to propose a nonlinear adaptation rule defined in terms of the product of differential inputs and errors which means a generalization of the normalized (N)LMS algorithm. The proposed method yields better tracking ability in this context as shown in the experiments which are carried out on the AURORA 2 and 3 speech databases. They provide an extensive performance evaluation along with an exhaustive comparison to standard LMS algorithms with almost the same computational load, including the NLMS and other recently reported LMS algorithms such as the modified (M)-NLMS, the error nonlinearity (EN)-LMS, or the normalized data nonlinearity (NDN)-LMS adaptation.

MOTIVATION: Modern machine learning methods based on matrix decomposition techniques, like independent component analysis (ICA) or non-negative matrix factorization (NMF), provide new and efficient analysis tools which are currently explored to analyze gene expression profiles. These exploratory feature extraction techniques yield expression modes (ICA) or metagenes (NMF). These extracted features are considered indicative of underlying regulatory processes. They can as well be applied to the classification of gene expression datasets by grouping samples into different categories for diagnostic purposes or group genes into functional categories for further investigation of related metabolic pathways and regulatory networks. RESULTS: In this study we focus on unsupervised matrix factorization techniques and apply ICA and sparse NMF to microarray datasets. The latter monitor the gene expression levels of human peripheral blood cells during differentiation from monocytes to macrophages. We show that these tools are able to identify relevant signatures in the deduced component matrices and extract informative sets of marker genes from these gene expression profiles. The methods rely on the joint discriminative power of a set of marker genes rather than on single marker genes. With these sets of marker genes, corroborated by leave-one-out or random forest cross-validation, the datasets could easily be classified into related diagnostic categories. The latter correspond to either monocytes versus macrophages or healthy vs Niemann Pick C disease patients.

BACKGROUND: The analysis of high-throughput gene expression data sets derived from microarray experiments still is a field of extensive investigation. Although new approaches and algorithms are published continuously, mostly conventional methods like hierarchical clustering algorithms or variance analysis tools are used. Here we take a closer look at independent component analysis (ICA) which is already discussed widely as a new analysis approach. However, deep exploration of its applicability and relevance to concrete biological problems is still missing. In this study, we investigate the relevance of ICA in gaining new insights into well characterized regulatory mechanisms of M-CSF dependent macrophage differentiation. RESULTS: Statistically independent gene expression modes (GEM) were extracted from observed gene expression signatures (GES) through ICA of different microarray experiments. From each GEM we deduced a group of genes, henceforth called sub-mode. These sub-modes were further analyzed with different database query and literature mining tools and then combined to form so called meta-modes. With them we performed a knowledge-based pathway analysis and reconstructed a well known signal cascade. CONCLUSION: We show that ICA is an appropriate tool to uncover underlying biological mechanisms from microarray data. Most of the well known pathways of M-CSF dependent monocyte to macrophage differentiation can be identified by this unsupervised microarray data analysis. Moreover, recent research results like the involvement of proliferation associated cellular mechanisms during macrophage differentiation can be corroborated.

This paper presents a novel voice activity detector (VAD) for improving speech detection robustness in noisy environments and the performance of speech recognition systems in real-time applications. The algorithm is based on a generalized complex Gaussian (GCG) observation model and defines an optimal likelihood ratio test (LRT) involving multiple and correlated observations (MCO) based on jointly Gaussian probability distribution functions (jGpdf). An extensive analysis of the proposed methodology for a low dimensional observation model demonstrates 1) the improved robustness of the proposed approach by means of a clear reduction of the classification error as the number of observations is increased, and 2) the tradeoff between the number of observations and the detection performance. The proposed strategy is also compared to different VAD methods including the G.729, AMR, and AFE standards, as well as other recently reported algorithms showing a sustained advantage in speech/nonspeech detection accuracy and speech recognition performance.

Nonnegative Matrix Factorization (NMF) has proven to be a useful tool for the analysis of nonnegative multivariate data. Gene expression profiles naturally conform to assumptions about data formats raised by NMF. However, it is known not to lead to unique results concerning the component signals extracted. In this paper we consider an extension of the NMF algorithm which provides unique solutions whenever the underlying component signals are sufficiently sparse. A new sparseness measure is proposed most appropriate to suitably transformed gene expression profiles. The resulting fitness function is discontinuous and exhibits many local minima, hence we use a genetic algorithm for its optimization. The algorithm is applied to toy data to investigate its properties as well as to a microarray data set related to Pseudo-Xanthoma Elasticum (PXE).

In underdetermined blind source separation, more sources are to be extracted from less observed mixtures without knowing both sources and mixing matrix. k-means-style clustering algorithms are commonly used to do this algorithmically given sufficiently sparse sources, but in any case other than deterministic sources, this lacks theoretical justification. After establishing that mean-based algorithms converge to wrong solutions in practice, we propose a median-based clustering scheme. Theoretical justification as well as algorithmic realizations (both online and batch) are given and illustrated by some examples.

In this work an automatic assignment tool for estimated independent components within an independent component analysis is presented. The tool is applied to the problem of removing the water resonance and related artifacts from multi-dimensional proton NMR spectra. The algorithm uses local PCA to approximate the water artifact and defines a suitable cost function which is optimized using simulated annealing. The blind extraction of artifact-related source signals is effected by a recently developed algorithm called dAMUSE.

We present a method based on singular spectrum analysis to remove ocular artifacts (EOG) from an electroencephalogram (EEC). After embedding the EEG signals in a feature space of time-delayed coordinates, feature vectors are clustered and the principal components (PCs) are computed locally within each cluster. Then we assume that the EOG artifact is associated with the PCs belonging to the largest eigenvalues. We incorporate a minimum description length (IMDL) criterion to estimate the number of eigenvectors needed to represent the EOG artifact faithfully. The EOG signal thus extracted is then subtracted from the original EEG signal to obtain the corrected EEG signal we are interested in.

We present a denoising algorithm for enhancing noisy signals based on local independent component analysis (ICA). This is done by applying ICA to the signal in localized delayed coordinates. The components resembling the signals can be detected by various criteria depending on the nature of the signal. Estimators of kurtosis or the variance of the autocorrelation have been considered. The algorithm proposed can favorably be applied to the problem of denoising multidimensional data like images or fMRI data sets. In comparison to denoising algorithms using wavelets, Wiener filters and kernel PCA the local PCA and ICA algorithms perform considerably better. We provide applications of the algorithm to images and the analysis of protein NMR spectra.

We discuss a comparative 3D spatial analysis of fMRI data taken during a combined word perception and motor task. We show that a classical GLM analysis using SPM does not yield reasonable results. Only with BSS techniques using the fastICA algorithm can we get meaningful and interesting results. The event-based experiment was part of a study to investigate the network of neurons involved in the perception of speech and the decoding of auditory speech stimuli. Corresponding to 4 different stimuli different independent components (IC) could be identified in the auditory cortex and, most interesting, an IC representing a network of 3 simultaneously active areas in the inferior frontal gyrus could be detected.

Multidimensional /sup 1/H NMR spectra of biomolecules dissolved in light water are contaminated by an intense water artifact. Generalized eigenvalue decomposition methods using congruent matrix pencils are used to separate the water artefact from the protein spectra. Due to the statistical separation process, however, noise is introduced into the reconstructed spectra. Hence Kernel-based denoising techniques are discussed to obtain noise- and artifact-free 2D NOESY NMR spectra of proteins.

In the first part of the paper, the one-dimensional functional equation g(y(t))=cg(z(t)) with known functions y and z and constant c is studied. Its indeterminacies are calculated, and an algorithm for approximating g is proposed. Then, this linearization identification algorithm is applied to the postnonlinear blind source separation (BSS) problem. In the case of bounded sources, a self-organizing map is used to approximate the boundary, and the postnonlinearity estimation is reduced to the one-dimensional equation from above. For super Gaussian sources, the density maxima are interpolated by performing linear BSS within concentric rings. Postnonlinearity estimation using ring approximation separates the mixtures.

In this work a modified version of AMUSE, called dAMUSE, is proposed. The main modification consists in increasing the dimension of the data vectors by joining delayed versions of the observed mixed signals. With the new data a matrix pencil is computed and its generalized eigendecomposition is performed as in AMUSE. We will show that in this case the output (or independent) signals are filtered versions of the source signals. Some numerical simulations using artificially mixed signals as well as biological data (RR and QT intervals of Electrocardiogram) are presented.

A neural network model for the identification and classification of malign and benign skin lesions from ALA-induced fluorescence images is presented. A self-organizing feature map or generative topographic mapping is used to cluster images patches according to their inherent local features, which then can be extracted with ICA. These components are used to distinguish skin cancer from benign lesions achieving an average classification rate of 70% so far.

Guided by the principles of geometric independent component analysis (ICA), we present a new approach (SOM-ICA) to linear geometric ICA using self-organizing map (SOM). We observe a considerable improvement in separation quality of different distributions, albeit at high computational costs. The SOMICA algorithm is therefore primarily interesting from a theoretical point of view bringing together ICA and SOMs; this intersection could lead to new proofs in geometric ICA based on similar theorems in the SOM theory.

Geometric algorithms for linear independent component analysis (ICA) have recently received some attention due to their pictorial description and their relative ease of implementation. The geometric approach to ICA was proposed first by Puntonet and Prieto (1995). We will reconsider geometric ICA in a theoretic framework showing that fixed points of geometric ICA fulfill a geometric convergence condition (GCC), which the mixed images of the unit vectors satisfy too. This leads to a conjecture claiming that in the nongaussian unimodal symmetric case, there is only one stable fixed point, implying the uniqueness of geometric ICA after convergence. Guided by the principles of ordinary geometric ICA, we then present a new approach to linear geometric ICA based on histograms observing a considerable improvement in separation quality of different distributions and a sizable reduction in computational cost, by a factor of 100, compared to the ordinary geometric approach. Furthermore, we explore the accuracy of the algorithm depending on the number of samples and the choice of the mixing matrix, and compare geometric algorithms with classical ICA algorithms, namely, Extended Infomax and FastICA. Finally, we discuss the problem of high-dimensional data sets within the realm of geometrical ICA algorithms.