Improvement in the information-oriented construction of temporary soil-retaining walls using sparse modeling

This fundamental study investigates how“super-resolution”technology based on sparse modeling,which has attracted attention in various fields,can be applied to the information-oriented construction of temporary soil-retaining walls.The machine learning process adopted here is based on the analytical results of numerical computations that involve many preliminary assumptions related to soilretaining walls,rather than the collection of images utilized in the image reconstruction technology.Consequently,bases for vectors related to the displacement of retaining walls are generated using efficient inverse analysis and“super-resolution”processing from sparse amounts of physical observation data.The purpose is to improve the properties of the inverse problem by artificial interpolation based on numerical analysis.It has been shown that the inverse analysis related to the displacement of retaining walls can be performed efficiently and that highly accurate predictions can be achieved even with limited physical observations.In general,the inverse analysis of retaining walls is an ill-posed problem.However,if the number of apparent observations reconverted by“super-resolution”technology exceeds the number of unknown parameters,then the displacement distribution of a retaining wall can be estimated efficiently.Another original idea is to break down the inverse problem into two separate problems by addressing the earth pressure distribution acting on the retaining wall.This makes it possible to identify the part to which the nonlinear inverse problem can be applied and to facilitate the efficient estimation and interpretation of the results.

Multi-Layer Deep Sparse Representation for Biological Slice Image Inpainting

Biological slices are an effective tool for studying the physiological structure and evolutionmechanism of biological systems.However,due to the complexity of preparation technology and the presence of many uncontrollable factors during the preparation processing,leads to problems such as difficulty in preparing slice images and breakage of slice images.Therefore,we proposed a biological slice image small-scale corruption inpainting algorithm with interpretability based on multi-layer deep sparse representation,achieving the high-fidelity reconstruction of slice images.We further discussed the relationship between deep convolutional neural networks and sparse representation,ensuring the high-fidelity characteristic of the algorithm first.A novel deep wavelet dictionary is proposed that can better obtain image prior and possess learnable feature.And multi-layer deep sparse representation is used to implement dictionary learning,acquiring better signal expression.Compared with methods such as NLABH,Shearlet,Partial Differential Equation(PDE),K-Singular Value Decomposition(K-SVD),Convolutional Sparse Coding,and Deep Image Prior,the proposed algorithm has better subjective reconstruction and objective evaluation with small-scale image data,which realized high-fidelity inpainting,under the condition of small-scale image data.And theOn2-level time complexitymakes the proposed algorithm practical.The proposed algorithm can be effectively extended to other cross-sectional image inpainting problems,such as magnetic resonance images,and computed tomography images.

Sparse Kernel Locality Preserving Projection and Its Application in Nonlinear Process Fault Detection

Locality preserving projection (LPP) is a newly emerging fault detection method which can discover local manifold structure of a data set to be analyzed, but its linear assumption may lead to monitoring performance degradation for complicated nonlinear industrial processes. In this paper, an improved LPP method, referred to as sparse kernel locality preserving projection (SKLPP) is proposed for nonlinear process fault detection. Based on the LPP model, kernel trick is applied to construct nonlinear kernel model. Furthermore, for reducing the computational complexity of kernel model, feature samples selection technique is adopted to make the kernel LPP model sparse. Lastly, two monitoring statistics of SKLPP model are built to detect process faults. Simulations on a continuous stirred tank reactor (CSTR) system show that SKLPP is more effective than LPP in terms of fault detection performance.

Anomaly detection in traffic surveillance with sparse topic model

Most research on anomaly detection has focused on event that is different from its spatial-temporal neighboring events.It is still a significant challenge to detect anomalies that involve multiple normal events interacting in an unusual pattern.In this work,a novel unsupervised method based on sparse topic model was proposed to capture motion patterns and detect anomalies in traffic surveillance.scale-invariant feature transform(SIFT)flow was used to improve the dense trajectory in order to extract interest points and the corresponding descriptors with less interference.For the purpose of strengthening the relationship of interest points on the same trajectory,the fisher kernel method was applied to obtain the representation of trajectory which was quantized into visual word.Then the sparse topic model was proposed to explore the latent motion patterns and achieve a sparse representation for the video scene.Finally,two anomaly detection algorithms were compared based on video clip detection and visual word analysis respectively.Experiments were conducted on QMUL Junction dataset and AVSS dataset.The results demonstrated the superior efficiency of the proposed method.

DOA ESTIMATION USING A SPARSE LINEAR MODEL BASED ON EIGENVECTORS

To reduce high computational cost of existing Direction-Of-Arrival(DOA) estimation techniques within a sparse representation framework,a novel method with low computational com-plexity is proposed.Firstly,a sparse linear model constructed from the eigenvectors of covariance matrix of array received signals is built.Then based on the FOCal Underdetermined System Solver(FOCUSS) algorithm,a sparse solution finding algorithm to solve the model is developed.Compared with other state-of-the-art methods using a sparse representation,our approach also can resolve closely and highly correlated sources without a priori knowledge of the number of sources.However,our method has lower computational complexity and performs better in low Signal-to-Noise Ratio(SNR).Lastly,the performance of the proposed method is illustrated by computer simulations.

Multimodality Prediction of Chaotic Time Series with Sparse Hard-Cut EM Learning of the Gaussian Process Mixture Model

The contribution of this work is twofold： （1） a multimodality prediction method of chaotic time series with the Gaussian process mixture （GPM） model is proposed, which employs a divide and conquer strategy. It automatically divides the chaotic time series into multiple modalities with different extrinsic patterns and intrinsic characteristics, and thus can more precisely fit the chaotic time series. （2） An effective sparse hard-cut expec- tation maximization （SHC-EM） learning algorithm for the GPM model is proposed to improve the prediction performance. SHO-EM replaces a large learning sample set with fewer pseudo inputs, accelerating model learning based on these pseudo inputs. Experiments on Lorenz and Chua time series demonstrate that the proposed method yields not only accurate multimodality prediction, but also the prediction confidence interval SHC-EM outperforms the traditional variational 1earning in terms of both prediction accuracy and speed. In addition, SHC-EM is more robust and insusceptible to noise than variational learning.

Dynamic Global-Principal Component Analysis Sparse Representation for Distributed Compressive Video Sampling

Video reconstruction quality largely depends on the ability of employed sparse domain to adequately represent the underlying video in Distributed Compressed Video Sensing (DCVS). In this paper, we propose a novel dynamic global-Principal Component Analysis (PCA) sparse representation algorithm for video based on the sparse-land model and nonlocal similarity. First, grouping by matching is realized at the decoder from key frames that are previously recovered. Second, we apply PCA to each group (sub-dataset) to compute the principle components from which the sub-dictionary is constructed. Finally, the non-key frames are reconstructed from random measurement data using a Compressed Sensing (CS) reconstruction algorithm with sparse regularization. Experimental results show that our algorithm has a better performance compared with the DCT and K-SVD dictionaries.

Sparse convolutional model with semantic expression for waste electrical appliances recognition

Deep neural networks play an important role in the recognition of waste electrical appliances. However, deep neural network components still lack reliability in decision-making features. To address this problem, a sparse convolutional model with semantic expression(SCMSE) is proposed. First, a low-rank sparse semantic expression component, combining the benefits of residual networks and sparse representation, is adapted to enhance sparse feature extraction and semantic expression. Second, a reliable network architecture is obtained by iterating the optimal sparse solution, enhancing semantic expression. Finally, the results of visualization experiments on the waste electrical appliances dataset demonstrate that the proposed SCMSE can obtain excellent semantic performance.

Human Action Recognition Based on Supervised Class-Specific Dictionary Learning with Deep Convolutional Neural Network Features

Human action recognition under complex environment is a challenging work.Recently,sparse representation has achieved excellent results of dealing with human action recognition problem under different conditions.The main idea of sparse representation classification is to construct a general classification scheme where the training samples of each class can be considered as the dictionary to express the query class,and the minimal reconstruction error indicates its corresponding class.However,how to learn a discriminative dictionary is still a difficult work.In this work,we make two contributions.First,we build a new and robust human action recognition framework by combining one modified sparse classification model and deep convolutional neural network(CNN)features.Secondly,we construct a novel classification model which consists of the representation-constrained term and the coefficients incoherence term.Experimental results on benchmark datasets show that our modified model can obtain competitive results in comparison to other state-of-the-art models.

Sparsity-Assisted Intelligent Condition Monitoring Method for Aero-engine Main Shaft Bearing

Weak feature extraction is of great importance for condition monitoring and intelligent diagnosis of aeroengine.Aimed at achieving intelligent diagnosis of aero-engine main shaft bearing,an enhanced sparsity-assisted intelligent condition monitoring method is proposed in this paper.Through analyzing the weakness of convex sparse model,i.e.the tradeoff between noise reduction and feature reconstruction,this paper proposes an enhanced-sparsity nonconvex regularized convex model based on Moreau envelope to achieve weak feature extraction.Accordingly,a sparsity-assisted deep convolutional variational autoencoders network is proposed,which achieves the intelligent identification of fault state through training denoised normal data.Finally,the effectiveness of the proposed method is verified through aero-engine bearing run-to-failure experiment.The comparison results show that the proposed method is good at abnormal pattern recognition,showing a good potential for weak fault intelligent diagnosis of aero-engine main shaft bearings.

Using the Correlation Criterion to Position and Shape RBF Units for Incremental Modelling

A novel technique is proposed for the incremental construction of sparse radial basis function （RBF） networks. The correlation between an RBF regressor and the training data is used as the criterion to position and shape the RBF node, and it is shown that this is equivalent to incrementally minimise the modelling mean square error. A guided random search optimisation method, called the repeated weighted boosting search, is adopted to append RBF nodes one by one in an incremental regression modelling procedure. The experimental results obtained using the proposed method demonstrate that it provides a viable alternative to the existing state-of-the-art modelling techniques for constructing parsimonious RBF models that generalise well.

Block Principle Component Analysis with Lp-norm for Robust and Sparse Modelling

Block principle and pattern classification component analysis （BPCA） is a recently developed technique in computer vision In this paper, we propose a robust and sparse BPCA with Lp-norm, referred to as BPCALp-S, which inherits the robustness of BPCA-L1 due to the employment of adjustable Lp-norm. In order to perform a sparse modelling, the elastic net is integrated into the objective function. An iterative algorithm which extracts feature vectors one by one greedily is elaborately designed. The monotonicity of the proposed iterative procedure is theoretically guaranteed. Experiments of image classification and reconstruction on several benchmark sets show the effectiveness of the proposed approach.

Robust Error Density Estimation in Ultrahigh Dimensional Sparse Linear Model

This paper focuses on error density estimation in ultrahigh dimensional sparse linear model,where the error term may have a heavy-tailed distribution.First,an improved two-stage refitted crossvalidation method combined with some robust variable screening procedures such as RRCS and variable selection methods such as LAD-SCAD is used to obtain the submodel,and then the residual-based kernel density method is applied to estimate the error density through LAD regression.Under given conditions,the large sample properties of the estimator are also established.Especially,we explicitly give the relationship between the sparsity and the convergence rate of the kernel density estimator.The simulation results show that the proposed error density estimator has a good performance.A real data example is presented to illustrate our methods.

Structured Learning in Biological Domain

Biological domain has been blessed with more and more data from biotechnologies as well as data integration tools.In the renaissance of machine learning and artificial intelligence,there is so much promise of data-driven biological knowledge discovery.However,it is not straight forward due to the complexity of the domain knowledge hidden in the data.At any level,be it atoms,molecules,cells or organisms,there are rich interdependencies among biological components.Machine learning approaches in this domain usually involves analyzing interdependency structures encoded in graphs and related formalisms.In this report,we review our work in developing new Machine Learning methods for these applications with improved performances in comparison with state-of-the-art methods.We show how the networks among biological components can be used to predict properties.

Application of kernel methods in signals modulation classification

A new approach to common signals classification of relevance vector machine （RVM） was presented and two signal classifiers based on kernel methods of support vector machine （SVM） and RVM were compared and analyzed. First several robust features of signals were extracted as the input of classifiers, then the kernel thought was used to map feature vectors impliedly to the high dimensional feature space, and multi-class RVM and SVM classifiers were designed to complete AM, CW, SSB, MFSK and MPSK signals recognition. Simulation result showed that when chose proper parameter, RVM and SVM had comparable accuracy but RVM had less learning time and basis functions. The classification speed of RVM is much faster than SVM.

Methods for Population-Based eQTL Analysis in Human Genetics

Gene expression is a critical process in biological system that is influenced and modulated by many factors including genetic variation. Expression Quantitative Trait Loci（e QTL） analysis provides a powerful way to understand how genetic variants affect gene expression. For genome wide e QTL analysis, the number of genetic variants and that of genes are large and thus the search space is tremendous. Therefore, e QTL analysis brings about computational and statistical challenges. In this paper, we provide a comprehensive review of recent advances in methods for e QTL analysis in population-based studies. We first present traditional pairwise association methods, which are widely used in human genetics. To account for expression heterogeneity, we investigate the methods for correcting confounding factors. Next, we discuss newly developed statistical learning methods including Lasso-based models. In the conclusion, we provide an overview of future method development in analyzing e QTL associations. Although we focus on human genetics in this review, the methods are applicable to many other organisms.