Abstract: The matrix completion problem has been studied broadly under many underlying conditions. In many real-life scenarios, we could expect elements from distinct columns or distinct positions to have a different cost. In this paper, we explore this generalization under adaptive conditions. We approach the problem under two different cost models. The first one is that entries from different columns have different observation costs, but, within the same column, each entry has a uniform cost. The second one is any two entry has different observation cost, despite being the same or different columns. We provide complexity analysis of our algorithms and provide tightness guarantees.
Abstract: High-performance computing (HPC) based emulators can be used to model the scattering from multiple stationary and moving targets for RADAR applications. These emulators rely on the RADAR Cross Section (RCS) of the targets being available in complex scenarios. Representing the RCS using tables generated from EM simulations is oftentimes cumbersome leading to large storage requirements. In this paper, we proposed a spherical harmonic based anisotropic scatterer model to represent the RCS of complex targets. The problem of finding the locations and reflection profiles of all scatterers can be formulated as a linear least square problem with a special sparsity constraint. We solve this problem using a modified Orthogonal Matching Pursuit algorithm. The results show that the spherical harmonic based scatterer model can effectively represent the RCS data of complex targets.
Abstract: In this work, we use machine learning and data analysis techniques to predict the one-year mortality of cirrhotic patients. Data from 2,322 patients with liver cirrhosis are collected at a single medical center. Different machine learning models are applied to predict one-year mortality. A comprehensive feature space including demographic information, comorbidity, clinical procedure and laboratory tests is being analyzed. A temporal pattern mining technic called Frequent Subgraph Mining (FSM) is being used. Model for End-stage liver disease (MELD) prediction of mortality is used as a comparator. All of our models statistically significantly outperform the MELD-score model and show an average 10% improvement of the area under the curve (AUC). The FSM technic itself does not improve the model significantly, but FSM, together with a machine learning technique called an ensemble, further improves the model performance. With the abundance of data available in healthcare through electronic health records (EHR), existing predictive models can be refined to identify and treat patients at risk for higher mortality. However, due to the sparsity of the temporal information needed by FSM, the FSM model does not yield significant improvements. Our work applies modern machine learning algorithms and data analysis methods on predicting one-year mortality of cirrhotic patients and builds a model that predicts one-year mortality significantly more accurate than the MELD score. We have also tested the potential of FSM and provided a new perspective of the importance of clinical features.
Abstract: Mixing in the hyperspectral imaging occurs due to the low spatial resolutions of the used cameras. The existing pure materials “endmembers” in the scene share the spectra pixels with different amounts called “abundances”. Unmixing of the data cube is an important task to know the present endmembers in the cube for the analysis of these images. Unsupervised unmixing is done with no information about the given data cube. Sparsity is one of the recent approaches used in the source recovery or unmixing techniques. The l1-norm optimization problem “basis pursuit” could be used as a sparsity-based approach to solve this unmixing problem where the endmembers is assumed to be sparse in an appropriate domain known as dictionary. This optimization problem is solved using proximal method “iterative thresholding”. The l1-norm basis pursuit optimization problem as a sparsity-based unmixing technique was used to unmix real and synthetic hyperspectral data cubes.
Abstract: In this work, we exploit two assumed properties of the abundances of the observed signatures (endmembers) in order to reconstruct the abundances from hyperspectral data. Joint-sparsity is the first property of the abundances, which assumes the adjacent pixels can be expressed as different linear combinations of same materials. The second property is rank-deficiency where the number of endmembers participating in hyperspectral data is very small compared with the dimensionality of spectral library, which means that the abundances matrix of the endmembers is a low-rank matrix. These assumptions lead to an optimization problem for the sparse unmixing model that requires minimizing a combined l2,p-norm and nuclear norm. We propose a variable splitting and augmented Lagrangian algorithm to solve the optimization problem. Experimental evaluation carried out on synthetic and real hyperspectral data shows that the proposed method outperforms the state-of-the-art algorithms with a better spectral unmixing accuracy.
Abstract: The convergence rate of the least-mean-square (LMS)
algorithm deteriorates if the input signal to the filter is correlated.
In a system identification problem, this convergence rate can be
improved if the signal is white and/or if the system is sparse. We
recently proposed a sparse transform domain LMS-type algorithm
that uses a variable step-size for a sparse system identification.
The proposed algorithm provided high performance even if the
input signal is highly correlated. In this work, we investigate the
performance of the proposed TD-LMS algorithm for a large number
of filter tap which is also a critical issue for standard LMS algorithm.
Additionally, the optimum value of the most important parameter is
calculated for all experiments. Moreover, the convergence analysis
of the proposed algorithm is provided. The performance of the
proposed algorithm has been compared to different algorithms in a
sparse system identification setting of different sparsity levels and
different number of filter taps. Simulations have shown that the
proposed algorithm has prominent performance compared to the other
algorithms.
Abstract: This paper presents a subband adaptive filter (SAF)
for a system identification where an impulse response is sparse
and disturbed with an impulsive noise. Benefiting from the uses
of l1-norm optimization and l0-norm penalty of the weight vector
in the cost function, the proposed l0-norm sign SAF (l0-SSAF)
achieves both robustness against impulsive noise and much improved
convergence behavior than the classical adaptive filters. Simulation
results in the system identification scenario confirm that the proposed
l0-norm SSAF is not only more robust but also faster and more
accurate than its counterparts in the sparse system identification in
the presence of impulsive noise.
Abstract: This paper presents a normalized subband adaptive
filtering (NSAF) algorithm to cope with the sparsity condition of
an underlying system in the context of compressive sensing. By
regularizing a weighted l1-norm of the filter taps estimate onto the
cost function of the NSAF and utilizing a subgradient analysis,
the update recursion of the l1-norm constraint NSAF is derived.
Considering two distinct weighted l1-norm regularization cases, two
versions of the l1-norm constraint NSAF are presented. Simulation
results clearly indicate the superior performance of the proposed
l1-norm constraint NSAFs comparing with the classical NSAF.
Abstract: This work presents a new type of the affine projection
(AP) algorithms which incorporate the sparsity condition of a
system. To exploit the sparsity of the system, a weighted l1-norm
regularization is imposed on the cost function of the AP algorithm.
Minimizing the cost function with a subgradient calculus and
choosing two distinct weighting for l1-norm, two stochastic gradient
based sparsity regularized AP (SR-AP) algorithms are developed.
Experimental results exhibit that the SR-AP algorithms outperform
the typical AP counterparts for identifying sparse systems.
Abstract: Image retrieval is a topic where scientific interest is currently high. The important steps associated with image retrieval system are the extraction of discriminative features and a feasible similarity metric for retrieving the database images that are similar in content with the search image. Gabor filtering is a widely adopted technique for feature extraction from the texture images. The recently proposed sparsity promoting l1-norm minimization technique finds the sparsest solution of an under-determined system of linear equations. In the present paper, the l1-norm minimization technique as a similarity metric is used in image retrieval. It is demonstrated through simulation results that the l1-norm minimization technique provides a promising alternative to existing similarity metrics. In particular, the cases where the l1-norm minimization technique works better than the Euclidean distance metric are singled out.
Abstract: In this paper a one-dimension Self Organizing Map
algorithm (SOM) to perform feature selection is presented. The
algorithm is based on a first classification of the input dataset on a
similarity space. From this classification for each class a set of
positive and negative features is computed. This set of features is
selected as result of the procedure. The procedure is evaluated on an
in-house dataset from a Knowledge Discovery from Text (KDT)
application and on a set of publicly available datasets used in
international feature selection competitions. These datasets come
from KDT applications, drug discovery as well as other applications.
The knowledge of the correct classification available for the training
and validation datasets is used to optimize the parameters for positive
and negative feature extractions. The process becomes feasible for
large and sparse datasets, as the ones obtained in KDT applications,
by using both compression techniques to store the similarity matrix
and speed up techniques of the Kohonen algorithm that take
advantage of the sparsity of the input matrix. These improvements
make it feasible, by using the grid, the application of the
methodology to massive datasets.
Abstract: User-based Collaborative filtering (CF), one of the
most prevailing and efficient recommendation techniques, provides
personalized recommendations to users based on the opinions of other
users. Although the CF technique has been successfully applied in
various applications, it suffers from serious sparsity problems. The
cloud-model approach addresses the sparsity problems by
constructing the user-s global preference represented by a cloud
eigenvector. The user-based CF approach works well with dense
datasets while the cloud-model CF approach has a greater
performance when the dataset is sparse. In this paper, we present a
hybrid approach that integrates the predictions from both the
user-based CF and the cloud-model CF approaches. The experimental
results show that the proposed hybrid approach can ameliorate the
sparsity problem and provide an improved prediction quality.
Abstract: In this paper an efficient incomplete factorization preconditioner is proposed for the Least Mean Squares (LMS) adaptive filter. The proposed preconditioner is approximated from a priori knowledge of the factors of input correlation matrix with an incomplete strategy, motivated by the sparsity patter of the upper triangular factor in the QRD-RLS algorithm. The convergence properties of IPLMS algorithm are comparable with those of transform domain LMS(TDLMS) algorithm. Simulation results show efficiency and robustness of the proposed algorithm with reduced computational complexity.
Abstract: Sparse representation has long been studied and several
dictionary learning methods have been proposed. The dictionary
learning methods are widely used because they are adaptive. In this
paper, a new dictionary learning method for audio is proposed. Signals
are at first decomposed into different degrees of Intrinsic Mode
Functions (IMF) using Empirical Mode Decomposition (EMD)
technique. Then these IMFs form a learned dictionary. To reduce the
size of the dictionary, the K-means method is applied to the dictionary
to generate a K-EMD dictionary. Compared to K-SVD algorithm, the
K-EMD dictionary decomposes audio signals into structured
components, thus the sparsity of the representation is increased by
34.4% and the SNR of the recovered audio signals is increased by
20.9%.