Search

Deep neural network (DNN) architecture based models have high expressive power and learning capacity. However, they are essentially a black box method since it is not easy to mathematically formulate the functions that are learned within its many layers of representation. Realizing this, many researchers have started to design methods to exploit the drawbacks of deep learning based algorithms questioning their robustness and exposing their singularities. In this paper, we attempt to unravel three aspects related to the robustness of DNNs for face recognition: (i) assessing the impact of deep architectures for face recognition in terms of vulnerabilities to attacks, (ii) detecting the singularities by characterizing abnormal filter response behavior in the hidden layers of deep networks; and (iii) making corrections to the processing pipeline to alleviate the problem. Our experimental evaluation using multiple open-source DNN-based face recognition networks, and three publicly available face databases demonstrates that the performance of deep learning based face recognition algorithms can suffer greatly in the presence of such distortions. We also evaluate the proposed approaches on four existing quasi-imperceptible distortions: DeepFool, Universal adversarial perturbations, l 2 , and Elastic-Net (EAD). The proposed method is able to detect both types of attacks with very high accuracy by suitably designing a classifier using the response of the hidden layers in the network. Finally, we present effective countermeasures to mitigate the impact of adversarial attacks and improve the overall robustness of DNN-based face recognition. [ABSTRACT FROM AUTHOR]

In this paper, an existing rule-based intrusion detection system (IDS) is made more intelligent through the application of machine learning. Snort was chosen as it is an open source software and though it was performing well, it showed false positives (FPs). To find the best performing machine learning algorithms (MLAs) to use with Snort so as to improve its detection, we tested some algorithms on three available datasets. Support vector machine (SVM) was chosen along with fuzzy logic and decision tree based on their accuracy. Combined versions of algorithms through ensemble SVM along with other variants were tried on the generated traffic of normal and malicious packets at 10Gbps. Optimized versions of the SVM along with firefly and ant colony optimization (ACO) were also tried, and the accuracy improved remarkably. Thus, the application of combined and optimized MLAs to Snort at 10Gbps worked quite well. [ABSTRACT FROM AUTHOR]

Sentiment analysis techniques are increasingly exploited to categorize the opinion text to one or more predefined sentiment classes for the creation and automated maintenance of review-aggregation websites. In this paper, a Malay sentiment analysis classification model is proposed to improve classification performances based on the semantic orientation and machine learning approaches. First, a total of 2,478 Malay sentiment-lexicon phrases and words are assigned with a synonym and stored with the help of more than one Malay native speaker, and the polarity is manually allotted with a score. In addition, the supervised machine learning approaches and lexicon knowledge method are combined for Malay sentiment classification with evaluating thirteen features. Finally, three individual classifiers and a combined classifier are used to evaluate the classification accuracy. In experimental results, a wide-range of comparative experiments is conducted on a Malay Reviews Corpus (MRC), and it demonstrates that the feature extraction improves the performance of Malay sentiment analysis based on the combined classification. However, the results depend on three factors, the features, the number of features and the classification approach. [ABSTRACT FROM AUTHOR]

In this paper, we analyze the spectrum occupancy in cognitive radio networks (CRNs) using different machine learning techniques. Both supervised techniques [naive Bayesian classifier (NBC), decision trees (DT), support vector machine (SVM), linear regression (LR)] and unsupervised algorithms [hidden Markov model (HMM)] are studied to find the best technique with the highest classification accuracy (CA). A detailed comparison of the supervised and unsupervised algorithms in terms of the computational time and the CA is performed. The classified occupancy status is further utilized to evaluate the blocking probability of secondary user for future time slots, which can be used by system designers to define spectrum-allocation and spectrum-sharing policies. Numerical results show that SVM is the best algorithm among all the supervised and unsupervised classifiers. Based on this, we proposed a new SVM algorithm by combining it with a firefly algorithm (FFA), which is shown to outperform all the other algorithms. [ABSTRACT FROM PUBLISHER]

Graph classification aims to learn models to classify structure data. To date, all existing graph classification methods are designed to target one single learning task and require a large number of labeled samples for learning good classification models. In reality, each real-world task may only have a limited number of labeled samples, yet multiple similar learning tasks can provide useful knowledge to benefit all tasks as a whole. In this paper, we formulate a new multi-task graph classification (MTG) problem, where multiple graph classification tasks are jointly regularized to find discriminative subgraphs shared by all tasks for learning. The niche of MTG stems from the fact that with a limited number of training samples, subgraph features selected for one single graph classification task tend to overfit the training data. By using additional tasks as evaluation sets, MTG can jointly regularize multiple tasks to explore high quality subgraph features for graph classification. To achieve this goal, we formulate an objective function which combines multiple graph classification tasks to evaluate the informativeness score of a subgraph feature. An iterative subgraph feature exploration and multi-task learning process is further proposed to incrementally select subgraph features for graph classification. Experiments on real-world multi-task graph classification datasets demonstrate significant performance gain. [ABSTRACT FROM PUBLISHER]

In recent years, imbalanced learning problem has attracted more and more attentions from both academia and industry, and the problem is concerned with the performance of learning algorithms in the presence of data with severe class distribution skews. In this paper, we apply the well-known statistical model logistic discrimination to this problem and propose a novel method to improve its performance. To fully consider the class imbalance, we design a new cost function which takes into account the accuracies of both positive class and negative class as well as the precision of positive class. Unlike traditional logistic discrimination, the proposed method learns its parameters by maximizing the proposed cost function. Experimental results show that, compared with other state-of-the-art methods, the proposed one shows significantly better performance on measures of recall, g-mean, f-measure, AUC, and accuracy. [ABSTRACT FROM AUTHOR]

Accurate tree-cover estimates are useful in deriving above-ground biomass density estimates from very high resolution (VHR) satellite imagery data. Numerous algorithms have been designed to perform tree-cover delineation in high-to-coarse-resolution satellite imagery, but most of them do not scale to terabytes of data, typical in these VHR data sets. In this paper, we present an automated probabilistic framework for the segmentation and classification of 1-m VHR data as obtained from the National Agriculture Imagery Program (NAIP) for deriving tree-cover estimates for the whole of Continental United States, using a high-performance computing architecture. The results from the classification and segmentation algorithms are then consolidated into a structured prediction framework using a discriminative undirected probabilistic graphical model based on conditional random field, which helps in capturing the higher order contextual dependence relations between neighboring pixels. Once the final probability maps are generated, the framework is updated and retrained by incorporating expert knowledge through the relabeling of misclassified image patches. This leads to a significant improvement in the true positive rates and reduction in false positive rates (FPRs). The tree-cover maps were generated for the state of California, which covers a total of 11 095 NAIP tiles and spans a total geographical area of 163 696 sq. miles. Our framework produced correct detection rates of around 88% for fragmented forests and 74% for urban tree-cover areas, with FPRs lower than 2% for both regions. Comparative studies with the National Land-Cover Data algorithm and the LiDAR high-resolution canopy height model showed the effectiveness of our algorithm for generating accurate high-resolution tree-cover maps. [ABSTRACT FROM PUBLISHER]

Both cost-sensitive classification and online learning have been extensively studied in data mining and machine learning communities, respectively. However, very limited study addresses an important intersecting problem, that is, “Cost-Sensitive Online Classification”. In this paper, we formally study this problem, and propose a new framework for Cost-Sensitive Online Classification by directly optimizing cost-sensitive measures using online gradient descent techniques. Specifically, we propose two novel cost-sensitive online classification algorithms, which are designed to directly optimize two well-known cost-sensitive measures: (i) maximization of weighted sum of sensitivity and specificity, and (ii) minimization of weighted misclassification cost. We analyze the theoretical bounds of the cost-sensitive measures made by the proposed algorithms, and extensively examine their empirical performance on a variety of cost-sensitive online classification tasks. Finally, we demonstrate the application of the proposed technique for solving several online anomaly detection tasks, showing that the proposed technique could be a highly efficient and effective tool to tackle cost-sensitive online classification tasks in various application domains. [ABSTRACT FROM AUTHOR]

In this paper, we propose a multiple kernel learning (MKL) algorithm that is based on the sparse representation-based classification (SRC) method. Taking advantage of the nonlinear kernel SRC in efficiently representing the nonlinearities in the high-dimensional feature space, we propose an MKL method based on the kernel alignment criteria. Our method uses a two step training method to learn the kernel weights and sparse codes. At each iteration, the sparse codes are updated first while fixing the kernel mixing coefficients, and then the kernel mixing coefficients are updated while fixing the sparse codes. These two steps are repeated until a stopping criteria is met. The effectiveness of the proposed method is demonstrated using several publicly available image classification databases and it is shown that this method can perform significantly better than many competitive image classification algorithms. [ABSTRACT FROM AUTHOR]

Subjective pattern recognition is a class of pattern recognition problems, where we not only merely know a few, if any, the strategies our brains employ in making decisions in daily life but also have only limited ideas on the standards our brains use in determining the equality/inequality among the objects. Face recognition is a typical example of such problems. For solving a subjective pattern recognition problem by machinery, application accuracy is the standard performance metric for evaluating algorithms. However, we indeed do not know the connection between algorithm design and application accuracy in subjective pattern recognition. Consequently, the research in this area follows a 'trial and error' process in a general sense: try different parameters of an algorithm, try different algorithms, and try different algorithms with different parameters. This phenomenon can be observed clearly in the nearly 30 years research of the face recognition: although huge advances have been made, no algorithm has ever been shown a potential to be consistently better than most of the algorithms developed earlier; it was even shown that a naïve algorithm can work, in the sense of accuracy, at least no worse than many newly developed ones in a few benchmarks. We argue that, the primary objective of subjective pattern recognition research should be moved to theoretical robustness from application accuracy so that we can evaluate and compare algorithms without or with only few 'trial and error' steps. We in this paper introduce an analytical model for studying the theoretical stabilities of multicandidate Electoral College and Direct Popular Vote schemes (aka regional voting scheme and national voting scheme, respectively), which can be expressed as the a posteriori probability that a winning candidate will continue to be chosen after the system is subjected to noise. This model shows that, in the context of multicandidate elections, generally, Electoral College is more stable than Direct Popular Vote, that the stability of Electoral College increases from that of Direct Popular Vote as the size of the subdivided regions decreases from the original nation size, up to a certain level, and then the stability starts to decrease approaching the stability of Direct Popular Vote as the region size approaches the original unit cell size; and that the stability of Electoral College approaches that of Direct Popular Vote in the two extremities as the region size increases to the original national size or decreases to the unit cell size. It also shows a special situation of white noise dominance with negligibly small concentrated noise, where Direct Popular Vote is surprisingly more stable than Electoral College, although the existence of such a special situation is questionable. We observe that 'high stability' in theory indeed always reveals itself in 'high accuracy' in applications. Extensive experiments on two human face benchmark databases applying an Electoral College framework embedded with standard baseline and newly developed holistic algorithms have been conducted. The impressive improvement by Electoral College over regular holistic algorithms verifies the stability theory on the voting systems. It also shows an evidential support for adopting theoretical stability instead of application accuracy as the primary objective for subjective pattern recognition research. [ABSTRACT FROM AUTHOR]

Pattern recognition has been employed in a myriad of industrial, commercial and academic applications. Many techniques have been devised to tackle such a diversity of applications. Despite the long tradition of pattern recognition research, there is no technique that yields the best classification in all scenarios. Therefore, as many techniques as possible should be considered in high accuracy applications. Typical related works either focus on the performance of a given algorithm or compare various classification methods. In many occasions, however, researchers who are not experts in the field of machine learning have to deal with practical classification tasks without an in-depth knowledge about the underlying parameters. Actually, the adequate choice of classifiers and parameters in such practical circumstances constitutes a long-standing problem and is one of the subjects of the current paper. We carried out a performance study of nine well-known classifiers implemented in the Weka framework and compared the influence of the parameter configurations on the accuracy. The default configuration of parameters in Weka was found to provide near optimal performance for most cases, not including methods such as the support vector machine (SVM). In addition, the k-nearest neighbor method frequently allowed the best accuracy. In certain conditions, it was possible to improve the quality of SVM by more than 20% with respect to their default parameter configuration. [ABSTRACT FROM AUTHOR]

This paper presents a computer vision framework supporting automated infrastructure damage detection, with a specific focus on surface crack detection in concrete. The approach presented is designed to provide a significant increase in robustness relative to existing methods when faced with widely varying field conditions while operating fast enough to be used in large scale applications. In particular, a clustering method for segmentation is developed that exploits inherent characteristics of fracture images to achieve consistent performance, combined with robust feature extraction to improve recognition algorithm classifier outcomes. The approach is shown to perform well in detecting cracks across a broad range of surface and lighting conditions, which can cause existing techniques to exhibit significant reductions in detection accuracy and/or detection speed. [ABSTRACT FROM AUTHOR]

Machine learning techniques for computer vision applications like object recognition, scene classification, etc., require a large number of training samples for satisfactory performance. Especially when classification is to be performed over many categories, providing enough training samples for each category is infeasible. This paper describes new ideas in multiclass active learning to deal with the training bottleneck, making it easier to train large multiclass image classification systems. First, we propose a new interaction modality for training which requires only yes-no type binary feedback instead of a precise category label. The modality is especially powerful in the presence of hundreds of categories. For the proposed modality, we develop a Value-of-Information (VOI) algorithm that chooses informative queries while also considering user annotation cost. Second, we propose an active selection measure that works with many categories and is extremely fast to compute. This measure is employed to perform a fast seed search before computing VOI, resulting in an algorithm that scales linearly with dataset size. Third, we use locality sensitive hashing to provide a very fast approximation to active learning, which gives sublinear time scaling, allowing application to very large datasets. The approximation provides up to two orders of magnitude speedups with little loss in accuracy. Thorough empirical evaluation of classification accuracy, noise sensitivity, imbalanced data, and computational performance on a diverse set of image datasets demonstrates the strengths of the proposed algorithms. [ABSTRACT FROM AUTHOR]

Cooperative coevolution is a successful trend of evolutionary computation which allows us to define partitions of the domain of a given problem, or to integrate several related techniques into one, by the use of evolutionary algorithms. It is possible to apply it to the development of advanced classification methods, which integrate several machine learning techniques into a single proposal. A novel approach integrating instance selection, instance weighting, and feature weighting into the framework of a coevolutionary model is presented in this paper. We compare it with a wide range of evolutionary and nonevolutionary related methods, in order to show the benefits of the employment of coevolution to apply the techniques considered simultaneously. The results obtained, contrasted through nonparametric statistical tests, show that our proposal outperforms other methods in the comparison, thus becoming a suitable tool in the task of enhancing the nearest neighbor classifier. [ABSTRACT FROM PUBLISHER]

Common spatial pattern (CSP) is a popular algorithm for classifying electroencephalogram (EEG) signals in the context of brain–computer interfaces (BCIs). This paper presents a regularization and aggregation technique for CSP in a small-sample setting (SSS). Conventional CSP is based on a sample-based covariance-matrix estimation. Hence, its performance in EEG classification deteriorates if the number of training samples is small. To address this concern, a regularized CSP (R-CSP) algorithm is proposed, where the covariance-matrix estimation is regularized by two parameters to lower the estimation variance while reducing the estimation bias. To tackle the problem of regularization parameter determination, R-CSP with aggregation (R-CSP-A) is further proposed, where a number of R-CSPs are aggregated to give an ensemble-based solution. The proposed algorithm is evaluated on data set IVa of BCI Competition III against four other competing algorithms. Experiments show that R-CSP-A significantly outperforms the other methods in average classification performance in three sets of experiments across various testing scenarios, with particular superiority in SSS. [ABSTRACT FROM AUTHOR]

• A study on learning comprehensible and accurate hybrid trees. • Hybrid trees – a combination of comprehensible and black-box classifieirs. • A novel algorithm evaluated on 23 UCI datasets. • The algorithm provides trade-off between accuracy and comprehensibility. • The algorithm generates hybrid-trees faster that a baseline search algorithm. Finding the best classifiers according to different criteria is often performed by a multi-objective machine learning algorithm. This study considers two criteria that are usually treated as the most important when deciding which classifier to apply in practice: comprehensibility and accuracy. A model that offers a broad range of trade-offs between the two criteria is introduced because they conflict; i.e., increasing one decreases the other. The choice of the model is motivated by the fact that domain experts often formalize decisions based on knowledge that can be represented by comprehensible rules and some tacit knowledge. This approach is mimicked by a hybrid tree that consists of comprehensible parts that originate from a regular classification tree and incomprehensible parts that originate from an accurate black-box classifier. An empirical evaluation on 23 UCI datasets shows that the hybrid trees provide trade-offs between the accuracy and comprehensibility that are not possible using traditional machine learning models. A corresponding hybrid-tree comprehensibility metric is also proposed. Furthermore, the paper presents a novel algorithm for learning MAchine LeArning Classifiers with HybrId TrEes (MALACHITE), and it proves that the algorithm finds a complete set of nondominated hybrid trees with regard to their accuracy and comprehensibility. The algorithm is shown to be faster than the well-known multi-objective evolutionary optimization algorithm NSGA-II for trees with moderate size, which is a prerequisite for comprehensibility. On the other hand, the MALACHITE algorithm can generate considerably larger hybrid-trees than a naïve exhaustive search algorithm in a reasonable amount of time. In addition, an interactive iterative data mining process based on the algorithm is proposed that enables inspection of the Pareto set of hybrid trees. In each iteration, the domain expert analyzes the current set of nondominated hybrid trees, infers domain relations, and sets the parameters for the next machine learning step accordingly. [ABSTRACT FROM AUTHOR]