Search

Unsupervised domain adaptation is intended to construct a reliable model for the unlabeled target samples using the well-labeled but differently distributed source samples. To tackle the domain shift issue, learning domain-invariant feature representations across domains is important, and most of the existing methods have concentrated on this goal. However, these methods rarely take into consideration the group discriminability of the feature representation, which is detrimental to the final recognition. Therefore, this article proposes a novel unsupervised domain adaptation method, named marginal subspace learning with group low-rank (MSL-GLR), to extract both domain-invariant and discriminative feature representations. Specifically, MSL-GLR uses the retargeting strategy to relax the regression matrix, such that the regression values would be forced to satisfy a margin maximization criterion for the requirement of correct classification. Moreover, MSL-GLR imposes a class-induced low-rank constraint, which enables the samples of each class to be located in their respective subspace. In this way, the distance between samples from the same class can be decreased and the discriminant ability of the projection is greatly improved. Furthermore, with the help of alternating direction method of multipliers (ADMM), an efficient algorithm is presented to solve the resulting optimization problem. Finally, the effectiveness of the proposed MSL-GLR is demonstrated by comprehensive evaluations on multiple domain adaptation benchmark datasets.

With the increasing scale of obtained multi-view data, how to deal with large-scale multi-view data quickly and efficiently is a significant problem. In this paper, a novel supervised multi-view feature selection method based on locally sparse regularization and block computing is proposed to solve the problem. Specifically, the multi-view dataset is firstly divided into sub-blocks according to classes and views. Then with the aid of the Alternating Direction Method of Multipliers (ADMM), a sharing sub-model is proposed to perform feature selection on each class by integrating each view’s locally sparse regularizers and shared loss that makes all views share a common penalty and regresses samples to their labels. Finally, all the sharing sub-models are fused to form the final general additive feature selection model, in which each sub-block adjusts its corresponding variables to perform block-based feature selection. In the optimization process, the proposed model can be decomposed into multiple separate subproblems, and an efficient optimization algorithm is proposed to solve them quickly. The comparison experiments with several state-of-the-art feature selection methods show that the proposed method is superior in classification accuracy and training speed.

In recent years, combining the multiple views of data to perform feature selection has been popular. As the different views are the descriptions from different angles of the same data, the abundant information coming from multiple views instead of the single view can be used to improve the performance of identification. In this paper, through the view weighted strategy, we propose a novel robust supervised multiview feature selection method, in which the robust feature selection is performed under the effect of l2,1-norm. The proposed model has the following advantages. Firstly, different from the commonly used view concatenation that is liable to ignore the physical meaning of features and cause over-fitting, the proposed method divides the original space into several subspaces and performs feature selection in the subspaces, which can reduce the computational complexity. Secondly, the proposed method assigns different weights to views adaptively according to their importance, which shows the complementarity and the specificity of views. Then, the iterative algorithm is given to solve the proposed model, and in each iteration, the original large-scale problem is split into the small-scale subproblems due to the divided original space. The performance of the proposed method is compared with several related state-of-the-art methods on the widely used multiview datasets, and the experimental results demonstrate the effectiveness of the proposed method.

Laplacian support vector machine (LapSVM), which is based on the semi-supervised manifold regularization learning framework, performs better than the standard SVM, especially for the case where the supervised information is insufficient. However, the use of hinge loss leads to the sensitivity of LapSVM to noise around the decision boundary. To enhance the performance of LapSVM, we present a novel semi-supervised SVM with the asymmetric squared loss (asy-LapSVM) which deals with the expectile distance and is less sensitive to noise-corrupted data. We further present a simple and efficient functional iterative method to solve the proposed asy-LapSVM, in addition, we prove the convergence of the functional iterative method from two aspects of theory and experiment. Numerical experiments performed on a number of commonly used datasets with noise of different variances demonstrate the validity of the proposed asy-LapSVM and the feasibility of the presented functional iterative method.

The conventional classification models implicitly assume that the distributions of data employed for training and test are identical. However, the assumption is rarely valid in many practical applications. In order to alleviate the difference between the distributions of the training and test sets, in this paper, we propose a regularized subspace learning framework based on the low-rank representation technique for unsupervised domain adaptation. Specifically, we introduce a regularization term of the subspace projection matrix to deal with the ill-conditioned problem and obtain a unique numerical solution. Meanwhile, we impose a structured sparsity-inducing regularizer on the error term so that the proposed method can filter out the outlier information, and therefore improve the performance. The extensive comparison experiments on benchmark data sets demonstrate the effectiveness of the proposed method.

The prevailing steganalysis detector trained by a source is used to recognize images from another different source, the detection accuracy typically drops owing to the mismatch between the two sources. In contrast to previous mismatched steganalysis methods, in this paper, we develop an unsupervised subspace learning-based method which has some differences from the ones common used in mismatched steganalysis. The proposed method employs low-rank and sparse constraints on the reconstruction coefficient matrix to maintain the global and local structures of the data. In this way, we can obtain new feature representations so that the feature distributions of the training and test data are close. We further promote the performance of the proposed method by employing the l2,1-norm on the error matrix. Comprehensive experiments on the JPEG mismatched steganalysis are conducted, and the experimental results show that the proposed method can improve the detection accuracy.

Supervised multi-view feature selection has recently been proven to be a valid way for reducing the dimensionality of data. This paper proposes a robust supervised multi-view feature selection method based on weighted shared loss and maximum margin criterion. Specifically, we fuse all weighted views to establish a shared loss term to maintain the complementary information of views, where the weight of each view can be automatically adjusted according to its contribution to the final task. Moreover, we adopt the capped l 2 -norm rather than the widely used l 2 -norm to measure the loss and remove the potential outliers. Further, we introduce the maximum margin criterion (MMC) into the feature selection and design a weighted view-based MMC such that it is a proper regularization associated with the proposed model. In this way, the inter-class and intra-class structure information of data is well exploited, which makes the margins of the samples belonging to the different classes increase and the margins of the samples belonging the same class decrease. Then the discriminative features of all views can be selected. The corresponding problem can be solved by the alternating direction method of multipliers (ADMM), which can realize the view-block calculation and reduce the computational complexity. Comprehensive experiments on various benchmark datasets and the application on the hydraulic system confirm the validity of our method.

Image steganalysis is a very important research topic in the field of information security. The existing feature based image steganalysis methods have achieved the appealing performance. The performance of them greatly depends on the quality of the hand-crafted steganalysis feature vectors, such as Cartesian Calibration PEV (CC-PEV), DCT Residuals (DCTR), and so on. However, these feature vectors may contain some redundant elements that will reduce the discrimination power and increase the computation cost. In this paper, a novel feature selection model is proposed for JPEG image steganalysis. Specifically, the proposed model imposes an \(l_{2,1}\)-structural constraint on the projection matrix for feature selection. Further, to make the model insensitive to noises and outliers, a capped \(l_2\)-norm based loss function is adopted. Moreover, a graph-based manifold regularization term which exploits the intrinsic local geometric structure of the data is added into the objective function to select the effective feature elements. Finally, an alternately iterative optimization algorithm with proven convergence is given to solve the proposed model. The extensive experiments on three state-of-the-art JPEG steganographic algorithms with 0.1 and 0.2 embedding rates and two JPEG quality factors show that the proposed model can effectively remove some irrelevant and redundant elements meanwhile retaining high detection accuracy.

In many practical applications, the assumption that the distributions of the data employed for training and test are identical is rarely valid, which would result in a rapid decline in performance. To address this problem, the domain adaptation strategy has been developed in recent years. In this paper, we propose a novel unsupervised domain adaptation method, referred to as Robust Adaptation Regularization based on Within-Class Scatter (WCS-RAR), to simultaneously optimize the regularized loss, the within-class scatter, the joint distribution between domains, and the manifold consistency. On the one hand, to make the model robust against outliers, we adopt an l2,1-norm based loss function in virtue of its row sparsity, instead of the widely-used l2-norm based squared loss or hinge loss function to determine the residual. On the other hand, to well preserve the structure knowledge of the source data within the same class and strengthen the discriminant ability of the classifier, we incorporate the minimum within-class scatter into the process of domain adaptation. Lastly, to efficiently solve the resulting optimization problem, we extend the form of the Representer Theorem through the kernel trick, and thus derive an elegant solution for the proposed model. The extensive comparison experiments with the state-of-the-art methods on multiple benchmark data sets demonstrate the superiority of the proposed method.

Most of the existing linguistic steganalysis methods mainly focus on detecting steganographic texts which are generated by embedding secret information into a type of text medium using one steganographic algorithm. But in practical applications, a large number of the steganographic texts may be hybrid ones which are generated by embedding secret information into different types of text media using different steganographic algorithms. In this paper, inspired by transfer learning, a novel linguistic steganalysis method is proposed to detect hybrid steganographic texts. The proposed method first uses the pre-trained BERT language model to obtain initial context-dependent word representations. Then the extracted features are fed into attentional Long Short-Term Memory (LSTM) to get the final contextual representations of sentences. The experimental results show that the proposed method can better satisfy the practical application demands than the existing linguistic steganalysis methods.

Domain adaptation learning is proposed as an effective technology for leveraging rich supervision knowledge from the related domain(s) to learn a reliable classifier for a new domain. One popular kind of domain adaptation methods is based on feature representation. However, such methods fail to consider the within-class and between-class relations after obtaining the new representation. In addition, they do not consider the negative effects of features that might be redundant or irrelevant to the final classification. To this end, a novel domain-invariant feature learning method based on the maximum margin criterion and sparsity technique for unsupervised domain adaptation is proposed in this paper, referred to as discriminative and informative joint distribution adaptation (DIJDA). Specifically, DIJDA adopts the maximum margin criterion in the adaptation process such that the transformed samples are near to those in the same class but segregated from those in different classes. As a result, the discriminative knowledge referred from source labels can be transferred to target domain effectively. Moreover, DIJDA imposes a row-sparsity constraint on the transformation matrix, which enforces rows of the matrix corresponding to inessential feature attributes to be all zero. Therefore, the most informative feature attributes can be extracted. Compared with several state-of-the-art methods, DIJDA substantially improves the classification results on five widely used benchmark datasets, which demonstrates the effectiveness of the proposed method.