Search

We discuss the sparse Canonical Correlation Analysis (CCA) problem in the context of high-dimensional multi-view problems, where we aim to discover interpretable association structures among multiple random vectors via their respective views with an emphasis on setting where the number of observations is too few compared to the number of covariates. Throughout this text, we use the term view define as observations of a random vector on an ordered set of subjects, which is the same for observations of all other random vectors involved in the analysis. We denote each view by Xi ∈ R n×pi , i = 1, . . . , m, where m is the number of random vectors, or equivalently number of views. In the first two chapters we consider linear association structures shared among multiple views, where the objective is to learn sparse linear combinations of multiple sets of covariates such that they are maximally correlated. In the first chapter we introduce a new approach to the sparse CCA, where we learn the sparsity pattern of the canonical directions in the first stage by casting this problem as two successively shrinking concave minimization programs which are solved via a first-order algorithm, and in the second stage we solve a small CCA problem by considering the sparsity patterns estimated in the first stage. We demonstrate via simulations that, in comparison to other available methods, our approach demonstrates superior convergence properties and capability to recover the underlying sparsity patterns and the magnitudes of the non-zero elements of the canonical directions, as well as, significantly lower computational cost. We then apply our method to a multi-omic environmental genetics study on fruit flies, where we hypothesise about the mechanism of adaptation of this model organism to environmental pesticides.In the second chapter we tackle a shared short-coming of sparse PCA and sparse CCA methods, which is that, in case of estimating multiple components or canonical directions for each view, these directions are not orthogonal to each other, which diminishes interpretability. While all other approaches estimate canonical directions one-by-one via the contraction scheme, we offer a block scheme where we estimate the first d canonical directions simultaneously. In this setting, we can more easily impose orthogonality, and also encourage disjoint sets of non-zero elements within multiple directions, resulting in more interpretable models. We also extended our model to what we call sparse Directed CCA, where we use an accessory variable, defined in the text, to try to capture variations related to a certain hypothesis, rather than the dominant variations which might be proven irrelevant to the main hypothesis. As a validating example, we apply our method to the lung cancer multi-omics available on The Cancer Genome Atlas, using survival data as our accessory variable. While regular sparse CCA exclusively identified correlation structures dominated by and communities separated by gender, our directed sparse CCA correctly identified two underlying communities which were significantly separated by survival.In the final chapter, we generalize our framework to discover non-linear association structures by proposing a two-stage sparse kernel CCA algorithm. We learn maximally aligned kernels in the first stage via sparse Multiple Kernel Learning (MKL), and then solve a KCCA problem in the second stage using learned kernels. We perform sparse MKL by forming an alignment matrix where its elements are the sample Hilbert Schmidt Independence Criterion of base kernels of pairs of views. These base kernels are functions of small sets of covariates of each view; therefore our sparse MKL approach provides interpretable solutions, as sparse convex linear combinations of base kernels. We finally provide an Apache Spark implementation of our methods introduced throughout the dissertation which makes users capable of running our methods on very high-dimensional datasets, e.g. observations on millions of Single Nucleotide Polymorphism loci, using distributed computing. We call this package SparKLe.R versions of our algorithms are also available. MuLe, BLOCCS, and SparKLe-R implements our methods presented in Chapters 1,2, and 3, respectively.

In clinical medicine, magnetic resonance imaging (MRI) is one of the most important tools for diagnosis, triage, prognosis, and treatment planning. However, MRI suffers from an inherent slow data acquisition process because data is collected sequentially in k-space. In recent years, most MRI reconstruction methods proposed in the literature focus on holistic image reconstruction rather than enhancing the edge information. This work steps aside this general trend by elaborating on the enhancement of edge information. Specifically, we introduce a novel parallel imaging coupled dual discriminator generative adversarial network (PIDD-GAN) for fast multi-channel MRI reconstruction by incorporating multi-view information. The dual discriminator design aims to improve the edge information in MRI reconstruction. One discriminator is used for holistic image reconstruction, whereas the other one is responsible for enhancing edge information. An improved U-Net with local and global residual learning is proposed for the generator. Frequency channel attention blocks (FCA Blocks) are embedded in the generator for incorporating attention mechanisms. Content loss is introduced to train the generator for better reconstruction quality. We performed comprehensive experiments on Calgary-Campinas public brain MR dataset and compared our method with state-of-the-art MRI reconstruction methods. Ablation studies of residual learning were conducted on the MICCAI13 dataset to validate the proposed modules. Results show that our PIDD-GAN provides high-quality reconstructed MR images, with well-preserved edge information. The time of single-image reconstruction is below 5ms, which meets the demand of faster processing., Comment: 33 pages, 13 figures, Applied Intelligence

Clustering data with both numeric and categorical attributes is of great importance as such data are ubiquitous in real-world problems. Multi-view learning approaches have proven to be more effective and having better generalisation ability compared to single-view learning in many problems. However, most of the existing clustering algorithms developed for mixed numeric and categorical data are single-view. In this research, we propose a novel multi-view clustering algorithm based on the k-prototypes (which we term Multi-view K-Prototypes) for clustering mixed data. To the best of our knowledge, our proposed Multi-view K-Prototypes is the first multi-view version of the well-known k-prototypes algorithm. To cluster the mixed data over multiple views, we present a novel representation prototype of cluster centres in the scenario of multiple views, and we also devise formulas for updating the cluster centres over each view. Then we propose the concept of consensus cluster centres to output the final clustering result. Finally, we carried out a series of experiments on four benchmark datasets to assess the performance of the proposed Multi-view K-Prototypes clustering. Experimental results show that the Multi-view K-Prototypes algorithm outperforms the seven state-of-the-art algorithms in most cases.

This paper proposes a weighted hybrid multi-view fusion method for the semi-supervised classification problem. Instead of getting access to the features from different views directly, this method utilizes the square losses of the multi-view classifiers to exploit the between-view relationship, which preserves the privacy of data. Considering the different prediction capability of classifiers on multiple views, an objective function with the constraint of rank consistency is constructed to weight view-specific learners adaptively, where the constraint makes each view-specific learner improve its performance by exploring the predicted results of other learners. Furthermore, an iterative algorithm based on the Variant Alternating Splitting Augmented Lagrangian Method (VASALM) and the quadratic programming method is developed to optimize the objective function. Experimental results on different real-world datasets demonstrate the effectiveness of the proposed method for multi-view learning. The experiments also analyze parameter sensitivity and convergency of the optimization algorithm. Refereed/Peer-reviewed

Social networks are extensively exploited by third-party consumers such as researchers and advertisers to understand user characteristics and behaviors. In general, before network data is published, sensitive relationships should be anonymized to prevent the compromise of individual privacy. To quantify the guarantee level of privacy-preserving mechanisms and mitigate users' privacy concerns, numerous studies concerning network data de-anonymization have been carried out. However, most existing studies focus on single-view data, and privacy protection for multi-view data that is ubiquitous in the era of big data has not been yet extensively explored. In this study, we are interested in answering the following question: Are the traditional privacy protection methods still valid for the anonymization of multi-view data? In this study, we propose a Multi-View Low-Rank Coding (MVLRC) based network data de-anonymization framework to assess the vulnerability of privacy protection techniques by accurately reconstructing a large portion of the original data. Specifically, the framework assumes that in principle, the target and auxiliary networks have common structural patterns, and they can be modeled together to infer the hidden structure of the target network. The essential components of our work include the following: (1) a robust network representation model for structural pattern learning; (2) the network representation based multi-view modeling of target network and auxiliary network; (3) the inference of the anonymized links via target network reconstruction. Experimental results on synthetic networks and three real-world networks demonstrate that auxiliary networks can be utilized by malicious adversaries for privacy inference attacks. Thus, the privacy protection of multi-view network data needs more sophisticated anonymization techniques.

By the development of social media, sentiment analysis has changed to one of the most remarkable research topics in the field of natural language processing which tries to dig information from textual data containing users’ opinions or attitudes toward a particular topic. In this regard, deep neural networks have emerged as promising techniques that have been extensively used for this aim in recent years and obtained significant results. Considering the fact that deep neural networks can automatically extract features from data, it can be claimed that intermediate representations extracted from these networks can be also used as appropriate features. While different deep neural networks are able to extract various types of features due to their distinct structures, we decided to combine features extracted from heterogeneous neural networks using multi-view classifiers to enhance the overall performance of document-level sentiment analysis by considering the correlation between them. The proposed multi-view deep network makes use of intermediate features extracted from convolutional and recursive neural networks to perform classification. Based on the results of the experiments, the proposed multi-view deep network not only outperforms single-view deep neural networks but also has superior efficiency and generalization performance.

This paper represents a semi-supervised learning framework, which integrates multi-view learning, extreme learning machine (ELM) and graph-based semi-supervised learning. The aim is to expand the scope of adaptation of non-negative sparse graph (NNSG) framework, under a multi-view condition and a non-linear relationship. The proposed multi-view learning method will be adaptive since when data is single-view the framework will degenerate into an embedded framework for NNSG framework. The proposed ELM method also will be adaptive since the number of hidden layer neuron will change with different number of input and output layer neuron. The combination of both proposed methods outperforms traditional graph-based semi-supervised learning, such as flexible manifold embedding (FME) and NNSG framework, which can not establish an affinity matrix for multi-view and can not establish a non-liner model for unknown data. Unlike traditional graph-based semi-supervised learning methods, which only can label propagation and build linear regression models for single or multi-view data, our proposed method has an obvious advantage that is applicable to any single or multi-view data, and builds linear or non-linear models. We provides extensive experiments on four public database in order to evaluate the performance of the proposed method. These experiments demonstrate significant improvement over the state-of-the-art algorithms in label propagation and processing of new data.

Takagi-Sugeno-Kang (TSK) fuzzy systems are well known for their good balance between approximation accuracy and interpretability. In this paper, we propose a deep view-reduction TSK fuzzy system termed as DVR-TSK-FS in which two powerful mechanisms associating with a deep structure are developed: 1) during the multi-view learning in each component, a sample-distribution-dependent parameter is defined to control the learning of the weight of each view. The parameter is not fixed by users, it is set according to the feature space in advance such that the learnt weight of each view indeed reflects the amount of pattern information involved in each view; 2) during the iteration of DRV-TSK-FS in each component, weak views are automatically reduced by comparing the learnt weight with a fixed threshold which is also automatically set according to the number of objects and the dimension of the feature space. 3) All components are linked in a stacked way based on the stacked generalization principle such that the outputs of all previous components are augmented into the current one which can help open the manifold structure of the original feature space. DRV-TSK-FS is testified on a multi-view EEG dataset for epileptic EEG signals recognition.

In multi-view multi-label learning, each object is represented by multiple data views, and belongs to multiple class labels simultaneously. Generally, all the data views have a contribution to the multi-label learning task, but their contributions are different. Besides, for each data view, each class label is only associated with a subset data features, and different features have different contributions to each class label. In this paper, we propose a novel framework VLSF for multi-view multi-label learning, i.e., multi-view multi-label learning with View-Label-Specific Features. Specifically, we first learn a low dimensional label-specific data representation for each data view and construct a multi-label classification model based on it by exploiting label correlations and view consensus, and learn the contribution weight of each data view to multi-label learning task for all the class labels jointly. Then, the final prediction can be made by combing the prediction results of all the classifiers and the learned contribution weights. The extensive comparison experiments with the state-of-the-art approaches manifest the effectiveness of the proposed method VLSF.

Multi-view learning mechanism, which enhances learning performance by training multi-model data sets, is a popular filed in recent years. Multi-view generalized eigenvalue proximal support vector machine (MvGSVM), as a most recently proposed classifier, has been shown to be successful in multi-model classification, which incorporates multi-view learning into classical GEPSVM. However, this method is still based on squared L2-norm distance measure, thus its robustness is not guaranteed in the presence of outliers. To address this problem, we propose a robust multi-view GEPSVM based on Lp-norm minimization and Ls-norm maximization. But, the introduction of the Lp-norm and Ls-norm makes the problem different from the generalized eigenvalue problem. So an efficient iterative algorithm is designed to solve this problem, and we also give the proof of convergence of the algorithm. The performances in extensive experiments demonstrate the effectiveness and robustness of the algorithm.

Graph-based multi-view feature learning methods learn a low-dimensional embedding of the data by modeling the affinity correlations with a graph to reduce the dimension. However, the learned low-dimensional representation relies on a fixed graph that is potentially inaccurate and unreliable. Besides, the graph construction and the projection matrix leaning are separated into two independent processes. To tackle the problems, we propose a robust unsupervised multi-view feature learning method with a dynamic graph. The dynamic graph structure is constructed adaptively and the robust projection matrix is learned simultaneously. Specifically, we adaptively learn a dynamic graph which captures the intrinsic multiple view-specific relations of samples. Robust projection matrix learning suppresses the adverse noises and preserves the intrinsic graph structure. Moreover, the assigned weights are learned automatically for each view without any extra parameter. We finally develop an efficient alternative optimization algorithm to solve the objective formulation. The extensive experiments conducted on several multi-view datasets demonstrate the effectiveness of our proposed method.

Grading of glioma is crucial for treatment decision making as well as prognostic assessments. In clinical routines, radiologists grade gliomas with multiple complementary magnetic resonance imaging (MRI) sequences, which is yet challenging for glioma prediction models. In this paper, we take full advantages of four commonly used MRI sequences to propose non-invasive grading of glioma based on a variant of maximum entropy discrimination (MED) and decision tree. First, radiomics features calculation is, respectively, performed on T1-weighted imaging, T2-weighted imaging, fluid attenuation inversion recovery imaging, and contrast-enhanced T1-weighted imaging. Then, radiomics features are integrated to build a glioma prediction model named four-sequence MED (FSMED) according to the assumption that the classification margin of different sequences is consistent. Finally, we propose a multi-MED decision tree (MMEDT) model to obtain the grading of gliomas based on the output of FSMED and the results of MED on each sequence. Validation experiments are conducted on a data set collected from Henan Provincial People's Hospital (GliomaHPPH2018) and Multimodal Brain Tumor Image Segmentation Benchmark 2017 (BraTS2017). The results of these two data sets demonstrate the high-prediction performance of our method. The average areas under the curve (AUC) of MMEDT are 0.9119, 0.8184, and 0.9084 for GliomaHPPH2018, BraTS2017, and their merged set, respectively, with the corresponding average sensitivities of 92.55%, 87.85%, and 87.91%, and average specificities of 92.57%, 81.36%, and 87.39%.

© 2017 Elsevier B.V. In multi-view clustering, datasets are comprised of different representations of the data, or views. Although each view could individually be used, exploiting information from all views together could improve the cluster quality. In this paper a new model Multi-View Kernel Spectral Clustering (MVKSC) is proposed that performs clustering when two or more views are available. This model is formulated as a weighted kernel canonical correlation analysis in a primal-dual optimization setting typical of Least Squares Support Vector Machines (LS-SVM). The primal model includes, in particular, a coupling term, which enforces the clustering scores corresponding to the different views to align. Because of the out-of-sample extension, this model is easily applied to large-scale datasets. The performance of the proposed model is shown on synthetic and real-world datasets, as well as on some large-scale datasets. Experimental comparisons with a number of other methods show that using multiple views improves the clustering results and that the proposed method is competitive with other state-of-the-art algorithms in terms of clustering accuracy and runtime. Especially on the large-scale datasets the advantage of the proposed method is clearly shown, as it is able to handle larger datasets than the other state-of-the-art algorithms. ispartof: An International Journal on Information Fusion vol:44 pages:46-56 status: published

Background: The brain magnetic resonance imaging (MRI) image segmentation method mainly refers to the division of brain tissue, which can be divided into tissue parts such as white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). The segmentation results can provide a basis for medical image registration, 3D reconstruction, and visualization. Generally, MRI images have defects such as partial volume effects, uneven grayscale, and noise. Therefore, in practical applications, the segmentation of brain MRI images has difficulty obtaining high accuracy.Materials and Methods: The fuzzy clustering algorithm establishes the expression of the uncertainty of the sample category and can describe the ambiguity brought by the partial volume effect to the brain MRI image, so it is very suitable for brain MRI image segmentation (B-MRI-IS). The classic fuzzy c-means (FCM) algorithm is extremely sensitive to noise and offset fields. If the algorithm is used directly to segment the brain MRI image, the ideal segmentation result cannot be obtained. Accordingly, considering the defects of MRI medical images, this study uses an improved multiview FCM clustering algorithm (IMV-FCM) to improve the algorithm’s segmentation accuracy of brain images. IMV-FCM uses a view weight adaptive learning mechanism so that each view obtains the optimal weight according to its cluster contribution. The final division result is obtained through the view ensemble method. Under the view weight adaptive learning mechanism, the coordination between various views is more flexible, and each view can be adaptively learned to achieve better clustering effects.Results: The segmentation results of a large number of brain MRI images show that IMV-FCM has better segmentation performance and can accurately segment brain tissue. Compared with several related clustering algorithms, the IMV-FCM algorithm has better adaptability and better clustering performance.

The growth of Internet and the pervasiveness of ICT have led to a radical change in social relationships. One of the drawbacks of this change is the exposure of individuals to threats during online activities. In this context, thetechno-regulationparadigm is inspiring new ways to safeguard legally interests by means of tools allowing to hamper breaches of law. In this paper, we focus on the exposure of individuals to specific online threats when interacting with smartphones. We propose a novel techno-regulatory approach exploiting machine learning techniques to provide safeguards against threats online. Specifically, we study a set of touch-based gestures to distinguish between underages or adults who is accessing a smartphone, and so to guarantee protection. To evaluate the proposed approach’s effectiveness, we developed an Android app to build a dataset consisting of more than 9000 touch-gestures from 147 participants. We experimented bothsingle-viewandmulti-viewlearning techniques to find the best combination of touch-gestures able of distinguishing between adults and underages. Results show that the multi-view learning combining scrolls, swipes, and pinch-to-zoom gestures, achieves the best ROC AUC (0.92) and accuracy (88%) scores.

Background: Traditionally, machine learning algorithms have been simply applied for software defect prediction by considering single-view data, meaning the input data contains a single feature vector. Nevertheless, different software engineering data sources may include multiple and partially independent information, which makes the standard single-view approaches ineffective. Objective: In order to overcome the single-view limitation in the current studies, this article proposes the usage of a multi-view learning method for software defect classification problems. Method: The Multi-View k-Nearest Neighbors (MVKNN) method was used in the software engineering field. In this method, first, base classifiers are constructed to learn from each view, and then classifiers are combined to create a robust multi-view model. Results: In the experimental studies, our algorithm (MVKNN) is compared with the standard k-nearest neighbors (KNN) algorithm on 50 datasets obtained from different software bug repositories. The experimental results demonstrate that the MVKNN method outperformed KNN on most of the datasets in terms of accuracy. The average accuracy values of MVKNN are 86.59%, 88.09%, and 83.10% for the NASA MDP, Softlab, and OSSP datasets, respectively. Conclusion: The results show that using multiple views (MVKNN) can usually improve classification accuracy compared to a single-view strategy (KNN) for software defect prediction.

Multiple kernel clustering (MKC) optimally utilizes a group of pre-specified base kernels to improve clustering performance. Among existing MKC algorithms, the recently proposed late fusion MKC methods demonstrate promising clustering performance in various applications and enjoy considerable computational acceleration. However, we observe that the kernel partition learning and late fusion processes are separated from each other in the existing mechanism, which may lead to suboptimal solutions and adversely affect the clustering performance. In this article, we propose a novel late fusion multiple kernel clustering with proxy graph refinement (LFMKC-PGR) framework to address these issues. First, we theoretically revisit the connection between late fusion kernel base partition and traditional spectral embedding. Based on this observation, we construct a proxy self-expressive graph from kernel base partitions. The proxy graph in return refines the individual kernel partitions and also captures partition relations in graph structure rather than simple linear transformation. We also provide theoretical connections and considerations between the proposed framework and the multiple kernel subspace clustering. An alternate algorithm with proved convergence is then developed to solve the resultant optimization problem. After that, extensive experiments are conducted on 12 multi-kernel benchmark datasets, and the results demonstrate the effectiveness of our proposed algorithm. The code of the proposed algorithm is publicly available at https://github.com/wangsiwei2010/graphlatefusion_MKC.

In this study, we propose a novel data analysis approach that can be used for multi-view analysis and integration of heterogeneous temporal data originating from multiple sources. The proposed approach consists of several distinctive layers: (i) select a suitable set (view) of parameters in order to identify characteristic behaviour within each individual source (ii) exploit an alternative set (view) of raw parameters (or high-level features) to derive some complementary representations (e.g. related to source performance) of the results obtained in the first layer with the aim to facilitate comparison and mediation across the different sources (iii) integrate those representations in an appropriate way, allowing to trace back similar cross-source performance to certain characteristic behaviour of the individual sources. The validity and the potential of the proposed approach has been demonstrated on a real-world dataset of a fleet of wind turbines. © Springer Nature Switzerland AG 2020.

BackgroundIn recent years, to investigate challenging bioinformatics problems, the utilization of multiple genomic and proteomic sources has become immensely popular among researchers. One such issue is feature or gene selection and identifying relevant and non-redundant marker genes from high dimensional gene expression data sets. In that context, designing an efficient feature selection algorithm exploiting knowledge from multiple potential biological resources may be an effective way to understand the spectrum of cancer or other diseases with applications in specific epidemiology for a particular population.ResultsIn the current article, we design the feature selection and marker gene detection as a multi-view multi-objective clustering problem. Regarding that, we propose an Unsupervised Multi-View Multi-Objective clustering-based gene selection approach called UMVMO-select. Three important resources of biological data (gene ontology, protein interaction data, protein sequence) along with gene expression values are collectively utilized to design two different views. UMVMO-selectaims to reduce gene space without/minimally compromising the sample classification efficiency and determines relevant and non-redundant gene markers from three cancer gene expression benchmark data sets.ConclusionA thorough comparative analysis has been performed with five clustering and nine existing feature selection methods with respect to several internal and external validity metrics. Obtained results reveal the supremacy of the proposed method. Reported results are also validated through a proper biological significance test and heatmap plotting.

International audience; Nowadays, we can observe a multiplication of multi-view data in domains such as marketing, bank administration, survey analysis, or social networks: We are dealing with large data bases that share a fair amount of data representing the same individual with different features depending on the data base. In this context, one can use Machine Learning methods to analyze this fragmented data across several heterogeneous sources (called views). Such analysis is subject to several difficulties: First, not all individual will be present and represented in all data sites and views. And second, this type of cross site analysis raises several ethical questions on privacy issues as no local site should have direct access to data from the other sources. To solve these problems, we present a method called the Cooperative Reconstruction System which aims at reconstructing information missing in some views in a multi-view context using information available in the other views. Furthermore, our method considers privacy issues and therefore achieves said reconstruction without direct data transfer from one view to another.

Metric learning is a class of efficient algorithms for EEG signal classification problem. Usually, metric learning method deals with EEG signals in the single view space. To exploit the diversity and complementariness of different feature representations, a new auto-weighted multi-view discriminative metric learning method with Fisher discriminative and global structure constraints for epilepsy EEG signal classification called AMDML is proposed to promote the performance of EEG signal classification. On the one hand, AMDML exploits the multiple features of different views in the scheme of the multi-view feature representation. On the other hand, considering both the Fisher discriminative constraint and global structure constraint, AMDML learns the discriminative metric space, in which the intraclass EEG signals are compact and the interclass EEG signals are separable as much as possible. For better adjusting the weights of constraints and views, instead of manually adjusting, a closed form solution is proposed, which obtain the best values when achieving the optimal model. Experimental results on Bonn EEG dataset show AMDML achieves the satisfactory results.

Recognition of motor imagery intention is one of the hot current research focuses of brain-computer interface (BCI) studies. It can help patients with physical dyskinesia to convey their movement intentions. In recent years, breakthroughs have been made in the research on recognition of motor imagery task using deep learning, but if the important features related to motor imagery are ignored, it may lead to a decline in the recognition performance of the algorithm. This paper proposes a new deep multi-view feature learning method for the classification task of motor imagery electroencephalogram (EEG) signals. In order to obtain more representative motor imagery features in EEG signals, we introduced a multi-view feature representation based on the characteristics of EEG signals and the differences between different features. Different feature extraction methods were used to respectively extract the time domain, frequency domain, time-frequency domain and spatial features of EEG signals, so as to made them cooperate and complement. Then, the deep restricted Boltzmann machine (RBM) network improved by t-distributed stochastic neighbor embedding(t-SNE) was adopted to learn the multi-view features of EEG signals, so that the algorithm removed the feature redundancy while took into account the global characteristics in the multi-view feature sequence, reduced the dimension of the multi-visual features and enhanced the recognizability of the features. Finally, support vector machine (SVM) was chosen to classify deep multi-view features. Applying our proposed method to the BCI competition IV 2a dataset we obtained excellent classification results. The results show that the deep multi-view feature learning method further improved the classification accuracy of motor imagery tasks.

Multi-view learning analyzes the information from several perspectives and has largely been applied on semi-supervised contexts. It has not been extensively analyzed for inducing interpretable rule-based classifiers. We present a multi-view and grammar-based genetic programming model for inducing rules for semi-supervised contexts. It evolves several populations and views, and promotes both accuracy and agreement among the views. This work details how and why common practices may not produce the expected results when inducing rule-based classifiers under this methodology.

Purpose: Automated segmentation of abnormal tissues in medical images is considered as an essential part of those computer-aided detection and diagnosis systems which analyze medical images. However, automated segmentation of abnormalities is a challenging task due to the limitations of imaging technologies and complex structure of abnormalities, including low contrast between normal and abnormal tissues, shape diversity, appearance inhomogeneity, and the vague boundaries of abnormalities. Therefore, more intelligent segmentation techniques are required to tackle these challenges. Materials and Methods: In this study, a method, which is called MMTDNN, is proposed to segment and detect medical image abnormalities. MMTDNN, as a multi-view learning machine, utilizes convolutional neural networks in a massive training strategy. Moreover, the proposed method has four phases of preprocessing, view generation, pixel-level segmentation, and post-processing. The International Symposium on Biomedical Imaging (ISBI)-2016 dataset is used for the evaluation of the proposed method. Results: The performance of the proposed method has been evaluated on the task of skin lesion segmentation as one of the challenging applications of abnormal tissue segmentation. Both qualitative and quantitative results demonstrate outstanding performance. Meanwhile, the accuracy of 0.973, the Jaccard index of 0.876, and the Dice similarity coefficient of 0.931 have been achieved. Conclusion: In conclusion, the experimental result demonstrates that the proposed method outperforms stateof-the-art methods of skin lesion segmentation.

Maritime surveillance videos provide crucial on-spot kinematic traffic information (traffic volume, ship speeds, headings, etc.) for varied traffic participants (maritime regulation departments, ship crew, ship owners, etc.) which greatly benefits automated maritime situational awareness and maritime safety improvement. Conventional models heavily rely on visual ship features for the purpose of tracking ships from maritime image sequences which may contain arbitrary tracking oscillations. To address this issue, we propose an ensemble ship tracking framework with a multi-view learning algorithm and wavelet filter model. First, the proposed model samples ship candidates with a particle filter following the sequential importance sampling rule. Second, we propose a multi-view learning algorithm to obtain raw ship tracking results in two steps: extracting a group of distinct ship contour relevant features (i.e., Laplacian of Gaussian, local binary pattern, Gabor filter, histogram of oriented gradient, and canny descriptors) and learning high-level intrinsic ship features by jointly exploiting underlying relationships shared by each type of ship contour features. Third, with the help of the wavelet filter, we performed a data quality control procedure to identify abnormal oscillations in the ship positions which were further corrected to generate the final ship tracking results. We demonstrate the proposed ship tracker&rsquo, s performance on typical maritime traffic scenarios through four maritime surveillance videos.

Demand forecasting is one of the main challenges for retailers and wholesalers in any industry. Proper demand forecasting gives business valuable information about potential profits and helps managers in taking targeted decisions on business growth strategies. Nowadays almost all organizations use different data sources or databases for nearly every aspect of their operations so that the knowledge on products on sale belongs to several independent views. The methodology described in this paper addresses the issue of product demand forecasting in fashion industry exploiting a multi-view learning approach. In particular, we show how the integration and connection among multiple views improves results accuracy. In real-life applications not all the views are usually available before a product is put on the market but the utility of a proper demand forecasting increases if the prediction is available before the product launch. We show that missing views can be reconstructed by means of common latent factors; in particular, this paper presents a learning procedure that describes the connection between different views. This connection allows data integration from multiple sources and can be extended to the special case of partial data representation. The nearest neighbors in the latent space play a special role for this process and for a general improvement of the forecast quality. We experimented the proposed methodology on real fashion retail sales showing that multi-view latent learning provides a system that is able to reconstruct satisfactorily non yet available views and can be used to predict the volumes of sales well before the goods are put on the market.

Since their beginnings, social networks have affected the way people communicate and interact with each other. The continuous growing and pervasive use of social media offers interesting research opportunities for analysing the behaviour and interactions of users. Nowadays, interactions are not only limited to social relations, but also to reading and writing activities. Thus, multiple and complementary information sources are available for characterising users and their activities. One task that could benefit from the integration of those multiple sources is community detection. However, most techniques disregard the effect of information aggregation and continue to focus only on one aspect: the topological structure of networks. This paper focuses on how to integrate social and content-based information originated in social networks for improving the quality of the detected communities. A technique for integrating both the multiple information sources and the semantics conveyed by asymmetric relations is proposed and extensively evaluated on two real-world datasets. Experimental evaluation confirmed the differentiated impact that each information source has on the quality of the detected communities, and shed some light on how to improve such quality by combining both social and content-based information. Fil: Tommasel, Antonela. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tandil. Instituto Superior de Ingeniería del Software. Universidad Nacional del Centro de la Provincia de Buenos Aires. Instituto Superior de Ingeniería del Software; Argentina Fil: Godoy, Daniela Lis. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Tandil. Instituto Superior de Ingeniería del Software. Universidad Nacional del Centro de la Provincia de Buenos Aires. Instituto Superior de Ingeniería del Software; Argentina

A great number of studies show that considering information from multiple views results in better performance than their single-view counterparts. However, many previous works aiming to improve classification performance fail to effectively tackle the inter-view correlation or the intra-class variability. Analysis dictionary learning (ADL) has theoretic significance and practical potential in classification tasks. Based on the ADL, this paper proposes a new method, namely, multi-view analysis dictionary learning (MvADL) for image classification. Specifically, multi-view analysis dictionaries are designed to reduce the intra-class variability in the transformed space in the multi-view scenario. Then, a marginalized classification term is incorporated to integrate the semantic information into the basic dictionary learning model. In the marginalized classification term, a marginalized target learning strategy is applied to improve the flexibility and discriminability of the whole model. Besides, an iteratively optimizing algorithm is designed to solve the proposed MvADL. Experiments on benchmark data sets demonstrate the superiority of our proposed method.

In this paper, we propose a robust change detection method for intelligent visual surveillance. This method, named $M^{4}CD$ , includes three major steps. First, a sample-based background model that integrates color and texture cues is built and updated over time. Second, multiple heterogeneous features (including brightness variation, chromaticity variation, and texture variation) are extracted by comparing the input frame with the background model, and a multi-view learning strategy is designed to online estimate the probability distributions for both foreground and background. The three features are approximately conditionally independent, making multi-view learning feasible. Pixel-wise foreground posteriors are then estimated with Bayes rule. Finally, the Markov random field (MRF) optimization and heuristic post-processing techniques are used sequentially to improve accuracy. In particular, a two-layer MRF model is constructed to represent pixel-based and superpixel-based contextual constraints compactly. Experimental results on the CDnet dataset indicate that $M^{4}CD$ is robust under complex environments and ranks among the top methods.

The problem of multi-view transformation is associated with transforming available source views of a given object into unknown target views. To solve this problem, a Mutual-Encoding InfoGenerative Adversarial Networks (MEIGANs)-based algorithm is proposed in this paper. A mutual-encoding representation learning network is proposed to obtain multi-view representations, i.e., it guarantees through encoders different views of the same object are mapped to the common representation, which carries enough information with respect to the object itself. An InfoGenerative Adversarial Networks-based transformation network is proposed to transform multi-views of the given object, which carries the representation information in the generative models and discriminative models, guaranteeing the synthetic transformed view matches the source view. The advantages of the MEIGAN are that it bypasses direct mappings among different views, and can solve the problem of missing views in training data and the problem of mapping between transformed views and source views. Finally, experiments on incomplete data to complete data restoration tasks on MNIST, CelebA, and multi-view angle transformation tasks on 3-D rendered chairs and multi-view clothing show the proposed algorithm yields satisfactory transformation results.

With the increase popularity of multi-view data, multi-view learning has attracted vital attentions in pattern recognition as well as machine learning. Most of existing methods apply in traditional single view learning. However, these methods neglect the complementary information among the views. The aim of multi-view is to discover complementary information and enhance the single view learning result. Multi-view is capable of capture incomplete and different types of information from multiple sources. However, multi-views may contain redundant information. Many multi-view methods assume that multi-views are generated from various view-specific generation matrices. This paper proposes the multi-view collaborative representation classification (MVCRC) algorithm which contains the information of different views and the connection of view-to-view. Experimental results conducted on five practical databases are used to confirm the effectiveness of the proposed approach.

Many sensor nodes have been widely deployed in the physical world to gather various environmental information, such as water quality, earthquake, and huge dam safety. Due to the limitation in the batter power, memory, and computational capacity, missing data can occur at arbitrary sensor nodes and time slots. In extreme situations, some sensors may lose readings at consecutive time slots. The successive missing data takes the side effects on the accuracy of real-time monitoring as well as the performance on the data analysis in the wireless sensor networks. Unfortunately, existing solutions to the missing data filling cannot well uncover the complex non-linear spatial and temporal relations. To address these problems, a DNN (Deep Neural Network) multi-view learning method (DNN-MVL) is proposed to fill the successive missing readings. DNN-MVL mainly considers five views: global spatial view, global temporal view, local spatial view, local temporal view, and semantic view. These five views are modeled with inverse distance of weight interpolation, bidirectional simple exponential smoothing, user-based collaborative filtering, mass diffusion-based collaborative filtering with the bipartite graph, and structural embedding, respectively. The results of the five views are aggregated to a final value in a multi-view learning algorithm with DNN model to obtain the final filling readings. Experiments on large-scale real dam deformation data demonstrate that DNN-MVL has a mean absolute error about 6.5%, and mean relative error 21.4%, and mean square error 8.17% for dam deformation data, outperforming all of the baseline methods.

International audience; Radiomics is a medical imaging technique that aims at extracting a large amount of features from one or several modalities of medical images, in order to help diagnose and treat diseases like cancers. Many recent studies have shown that Radiomics features can offer a lot of useful information that physicians cannot extract from these images, and can be efficiently associated with other information like gene or protein data. However, most of the classification studies in Radiomics report the use of feature selection methods without identifying the underlying machine learning challenges. In this paper, we first show that the Radiomics classification problem should be viewed as a high dimensional, low sample size, multi-view learning problem. Then, we propose a dissimilarity-based method for merging the information from the different views, based on Random Forest classifiers. The proposed approach is compared to different state-of-the-art Radiomics and multi-view solutions, on different public multi-view datasets as well as on Radiomics datasets. In particular, our experiments show that the proposed approach works better than the state-of-the-art methods from the Radiomics, as well as from the multi-view learning literature.

International audience; In this paper, we address the problem of learning from im-balanced multi-class datasets in a supervised setting when multiple descriptions of the data-also called views-are available. Each view incorporates various information on the examples, and in particular, depending on the task at hand, each view might be better at recognizing only a subset of the classes. Establishing a sort-of cooperation between the views is needed for all the classes to be equally recognized-a crucial problem particularly for imbalanced datasets. The novelty of our work consists in capitalizing on the complementariness of the views so that each class can be processed by the most appropriate view(s); thus improving the per-class performances of the final classifier. The main contribution of this paper are two ensemble learning methods based on recent theoretical works on the use of the confusion matrix's norm as an error measure, while empirical results show the benefits of the proposed approaches.

Patients in hospitals, particularly in critical care, are susceptible to many complications affecting morbidity and mortality. Digitized clinical data in electronic medical records can be effectively used to develop machine learning models to identify patients at risk of complications early and provide prioritized care to prevent complications. However, clinical data from heterogeneous sources within hospitals pose signi ficant modeling challenges. In particular, unstructured clinical notes are a valuable source of information containing regular assessments of the patient's condition but contain inconsistent abbreviations and lack the structure of formal documents. Our contributions in this paper are twofold. First, we present a new preprocessing technique for extracting features from informal clinical notes that can be used in a classifi cation model to identify patients at risk of developing complications. Second, we explore the use of collective matrix factorization, a multi-view learning technique, to model heterogeneous clinical data text-based features in combination with other measurements, such as clinical investigations, comorbidities, and demographic data. We present a detailed case study on postoperative respiratory failure using more than 700 patient records from the MIMIC II database. Our experiments demonstrate the ef ficacy of our preprocessing technique in extracting discriminatory features from clinical notes as well as the bene fits of multi-view learning to combine clinical measurements with text data for predicting complications.

This paper introduces a novel multi-view multi-learner (MVML) active learning method, in which the different views are generated by a genetic algorithm (GA). The GA-based view generation method attempts to construct diverse, sufficient, and independent views by considering both inter- and intra-view confidences. Hyperspectral data inherently owns high dimensionality, which makes it suitable for multi-view learning algorithms. Furthermore, by employing multiple learners at each view, a more accurate estimation of the underlying data distribution can be obtained. We also implemented a spectral-spatial graph-based semi-supervised learning (SSL) method as the classifier, which improved the performance of the classification task in comparison with supervised learning. The evaluation of the proposed method was based on three different benchmark hyperspectral data sets. The results were also compared with other state-of-the-art AL-SSL methods. The experimental results demonstrated the efficiency and statistically significant superiority of the proposed method. The GA-MVML AL method improved the classification performances by 16.68%, 18.37%, and 15.1% for different data sets after 40 iterations.

Hyperspectral image (HSI) Clustering is an unsupervised task, which segments pixels into different groups without using labeled samples. In this paper, spatial-spectral based multi-view low-rank sparse subspace clustering (SSMLC) algorithm is proposed. Due to significant number of spectra bands HSI contains much more information than a regular image. These spectral information can be considered as multiview. In this paper, the spectral partitioning is applied to generate spectral views which contain correlated bands. Morphological features of the original HSI are taken as another view which contains spatial features. Principal components construct another view, which eliminates the noise in the original dataset. After the multi-view dataset is formed, multi-view low-rank sparse subspace clustering is applied to segment HSI. Our experiments show that the performance of the proposed SSMLC is better than other that of clustering algorithms such as sparse subspace clustering and low-rank sparse subspace clustering.

Cancer diagnosis and treatment often require a personalized analysis for each patient nowadays, due to the heterogeneity among the different types of tumor and among patients. Radiomics is a recent medical imaging field that has shown during the past few years to be promising for achieving this personalization. However, a recent study shows that most of the state-of-the-art works in Radiomics fail to identify this problem as a multi-view learning task and that multi-view learning techniques are generally more efficient. In this work, we propose to further investigate the potential of one family of multi-view learning methods based on Multiple Classifiers Systems where one classifier is learnt on each view and all classifiers are combined afterwards. In particular, we propose a random forest based dynamic weighted voting scheme, which personalizes the combination of views for each new patient for classification tasks. The proposed method is validated on several real-world Radiomics problems., 10 pages

In biomedical research, many different types of patient data can be collected, such as various types of omics data and medical imaging modalities. Applying multi-view learning to these different sources of information can increase the accuracy of medical classification models compared with single-view procedures. However, collecting biomedical data can be expensive and/or burdening for patients, so that it is important to reduce the amount of required data collection. It is therefore necessary to develop multi-view learning methods which can accurately identify those views that are most important for prediction. In recent years, several biomedical studies have used an approach known as multi-view stacking (MVS), where a model is trained on each view separately and the resulting predictions are combined through stacking. In these studies, MVS has been shown to increase classification accuracy. However, the MVS framework can also be used for selecting a subset of important views. To study the view selection potential of MVS, we develop a special case called stacked penalized logistic regression (StaPLR). Compared with existing view-selection methods, StaPLR can make use of faster optimization algorithms and is easily parallelized. We show that nonnegativity constraints on the parameters of the function which combines the views play an important role in preventing unimportant views from entering the model. We investigate the performance of StaPLR through simulations, and consider two real data examples. We compare the performance of StaPLR with an existing view selection method called the group lasso and observe that, in terms of view selection, StaPLR is often more conservative and has a consistently lower false positive rate., Comment: 26 pages, 9 figures. Accepted manuscript

The primary objective of sentence segmentation process is to determine the sentence boundaries of a stream of words output by the automatic speech recognizers. Statistical methods developed for sentence segmentation requires a significant amount of labeled data which is time-consuming, labor intensive and expensive. In this work, we propose new multi-view semi-supervised learning strategies for sentence boundary classification problem using lexical, prosodic, and morphological information. The aim is to find effective semi-supervised machine learning strategies when only small sets of sentence boundary labeled data are available. We primarily investigate two semi-supervised learning approaches, called self-training and co-training. Different example selection strategies were also used for co-training, namely, agreement, disagreement and self-combined. Furthermore, we propose three-view and committee-based algorithms incorporating with agreement, disagreement and self-combined strategies using three disjoint feature sets. We present comparative results of different learning strategies on the sentence segmentation task. The experimental results show that the sentence segmentation performance can be highly improved using multi-view learning strategies that we proposed since data sets can be represented by three redundantly sufficient and disjoint feature sets. We show that the proposed strategies substantially improve the average baseline F-measure of 67.66% to 75.15% and 64.84% to 66.32% when only a small set of manually labeled data is available for Turkish and English spoken languages, respectively. This material is based upon work supported by the Scientific and Technological Research Council of Turkey (TUBITAK) (Project Number: 107E182 and Project Number: 111E228), Isik University Scientific Research Projects Fund (Project Number: 09A301 and Project Number: 14A201), TUBITAK BIDEB and J. William Fulbright Post-Doctoral Research Fellowship, USA fundings at SRI-International, Speech Technology and Research (STAR) Lab., Menlo Park, CA, USA and International Computer Science Institute (ICSI) Speech Group, University of California at Berkeley, CA, USA. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the funding agencies. The authors thank Gokhan Tur, Dilek Hakkani- Tur, Benoit Favre, Sebastien Cuendet, Murat Saraclar, Siddika Parlak, Erinc Dikici, Izel D. Revidi, Cenk Demiroglu and Fatih Ozaydin and Bogazici University Signal and Image Processing (BUSIM) Group for many helpful discussions. The authors also thank the anonymous reviewers for their useful comments on an earlier version of this paper. Publisher's Version

In this paper, three novel classification algorithms aiming at (semi-)supervised action classification are proposed. Inspired by the effectiveness of discriminant subspace learning techniques and the fast and efficient Extreme Learning Machine (ELM) algorithm for Single-hidden Layer Feedforward Neural networks training, the ELM algorithm is extended by incorporating discrimination criteria in its optimization process, in order to enhance its classification performance. The proposed Discriminant ELM algorithm is extended, by incorporating proper regularization in its optimization process, in order to exploit information appearing in both labeled and unlabeled action instances. An iterative optimization scheme is proposed in order to address multi-view action classification. The proposed classification algorithms are evaluated on three publicly available action recognition databases providing state-of-the-art performance in all the cases.

In this work, a novel unsupervised hashing algorithm, referred to as t-USMVH, and its extension to unsupervised deep hashing, referred to as t-UDH, are proposed to support large-scale video-to-video retrieval. To improve robustness of the unsupervised learning, t-USMVH combines multiple types offeature representations and effectively fuses them by examining a continuous relevance score based on a Gaussian estimation over pairwise distances, and also a discrete neighbor score based on the cardinality of reciprocal neighbors. To reduce sensitivity to scale changes for mapping objects that are far apart from each other, Student t-distribution is used to estimate the similaritybetween the relaxed hash code vectors for keyframes. This results in more accurate preservation of the desired unsupervised similarity structure in the hash code space. By adapting the corresponding optimization objective and constructing the hash mapping function via a deep neural network, we develop a robust unsupervised training strategy for a deep hashing network. Theefficiency and effectiveness of the proposed methods are evaluated on two public video collections via comparisons against multiple classical and state-of-the-art methods.

Several computational approaches have been proposed for inferring the affective state of the user, motivated for example by the goal of building improved interfaces that can adapt to the user@?s needs and internal state. While fairly good results have been obtained for inferring the user state under highly controlled conditions, a considerable amount of work remains to be done for learning high-quality estimates of subjective evaluations of the state in more natural conditions. In this work, we discuss how two recent machine learning concepts, multi-view learning and multi-task learning, can be adapted for user state recognition, and demonstrate them on two data collections of varying quality. Multi-view learning enables combining multiple measurement sensors in a justified way while automatically learning the importance of each sensor. Multi-task learning, in turn, tells how multiple learning tasks can be learned together to improve the accuracy. We demonstrate the use of two types of multi-task learning: learning both multiple state indicators and models for multiple users together. We also illustrate how the benefits of multi-task learning and multi-view learning can be effectively combined in a unified model by introducing a novel algorithm.

Multi-view learning techniques are necessary when data is described by multiple distinct feature sets because single-view learning algorithms tend to overfit on these high-dimensional data. Prior successful approaches followed either consensus or complementary principles. Recent work has focused on learning both the shared and private latent spaces of views in order to take advantage of both principles. However, these methods can not ensure that the latent spaces are strictly independent through encouraging the orthogonality in their objective functions. Also little work has explored representation learning techniques for multi-view learning. In this paper, we use the denoising autoencoder to learn shared and private latent spaces, with orthogonal constraints -- disconnecting every private latent space from the remaining views. Instead of computationally expensive optimization, we adapt the backpropagation algorithm to train our model.

A main goal of data visualization is to find, from among all the available alternatives, mappings to the 2D/3D display which are relevant to the user. Assuming user interaction data, or other auxiliary data about the items or their relationships, the goal is to identify which aspects in the primary data support the user\'s input and, equally importantly, which aspects of the user\'s potentially noisy input have support in the primary data. For solving the problem, we introduce a multi-view embedding in which a latent factorization identifies which aspects in the two data views (primary data and user data) are related and which are specific to only one of them. The factorization is a generative model in which the display is parameterized as a part of the factorization and the other factors explain away the aspects not expressible in a two-dimensional display. Functioning of the model is demonstrated on several data sets., Comment: IEEE International Conference on Acoustic, Speech and Signal Processing 2016

International audience; In many applications, observations are available with different views. This is, for example, the case with image-text classification, multilingual document classification or document classification on the web. In addition, unlabeled multiview examples can be easily acquired, but assigning labels to these examples is usually a time consuming task. We describe a multiview self-learning strategy which trains different voting classifiers on different views. The margin distributions over the unlabeled training data, obtained with each view-specific classifier are then used to estimate an upper-bound on their transductive Bayes error. Minimizing this upper-bound provides an automatic margin-threshold which is used to assign pseudo-labels to unlabeled examples. Final class labels are then assigned to these examples, by taking a vote on the pool of the previous pseudo-labels. New view-specific classifiers are then trained using the labeled and pseudo-labeled training data. We consider applications to image-text classification and to multilingual document classification. We present experimental results on the NUS-WIDE collection and on Reuters RCV1-RCV2 which show that despite its simplicity, our approach is competitive with other state-of-the-art techniques.

In this paper, we employ Single-hidden Layer Feedforward Neural networks in order to perform human action recognition based on multiple action representations. In order to determine both optimized network and action representation combination weights, we propose an optimization process that jointly minimizes the overall network training error and the within-class variance of the training data in the corresponding hidden layer spaces. The proposed approach has been evaluated by using the state-of-the-art Bag of Features-based action video representation on three publicly available action recognition databases, where it outperforms two commonly used video representation combination approaches, as well as the best single-descriptor classification outcome.

Multi-view algorithms, such as co-training and co-EM, utilize unlabeled data when the available attributes can be split into independent and compatible subsets. Experiments have shown that multi-view learning is sometimes beneficial for problems for which the independence assumption is not satisfied. In practice, unfortunately, it is not possible to measure the dependency between two attribute sets; hence, there is no criterion which allows to decide whether multi-view learning is applicable. We conduct experiments with various text classification problems and investigate on the effectiveness of the co-trained SVM and the co-EM SVM under various conditions, including violations of the independence assumption. We identify the error correlation coefficient of the initial classifiers as an elaborate indicator of the expected benefit of multi-view learning. Multi-view algorithms, such as co-training and co-EM, utilize unlabeled data when the available attributes can be split into independent and compatible subsets. Experiments have shown that multi-view learning is sometimes beneficial for problems for which the independence assumption is not satisfied.In practice, unfortunately, it is not possible to measure the dependency between two attribute sets; hence, there is no criterion which allows to decide whether multi-view learning is applicable. We conduct experiments with various text classification problems and investigate on the effectiveness of the co-trained SVM and the co-EM SVM under various conditions, including violations of the independence assumption. We identify the error correlation coefficient of the initial classifiers as an elaborate indicator of the expected benefit of multi-view learning.

Multi-instance (MI) learning is different than standard propositional classification, as it uses a set of bags containing many instances as input. While the instances in each bag are not labeled, the bags themselves are, as positive or negative. In this paper, we present a novel multi-view, two-level classification framework to address the generalized multi-instance problems. We first apply supervised and unsupervised learning methods to transform a MI dataset into a multi-view, single meta-instance dataset. Then we develop a multi-view learning approach that can integrate the information acquired by individual view learners on the meta-instance dataset from the previous step, and construct a final model. Our empirical studies show that the proposed method performs well compared to other popular MI learning methods., 2nd IEEE International Conference on Big Data, IEEE Big Data 2014, October 27-30, 2014, Washington, DC, USA