Search

This article presents an analysis of data extraction for classification using correlation coefficient and fuzzy model. Several traditional methods of data extraction are used for classification that could not provide sufficient information for further step of data analysis on class. It needs refinement of features data to distinguish a class that differs from a traditional class. Thus, it proposes the feature tiny data (subfeature data) to find distinguish class from a traditional class using two methods such as correlation coefficient and fuzzy model to select features as well as subfeature for distinguishing class. In the first approach, the correlation coefficient methods with gradient descent technique are used to select features from the dataset and in the second approach, the fuzzy model with supreme of minimum value is considered to get subfeature data. As per the proposed model, some features (i.e., three features from the acoustic dataset, two features from the QCM dataset, and eight features from the audit dataset, etc.) and subfeatures (as per threshold value like 20 for acoustic; 10 for QCM, and 20 for audit, etc.) are selected based on correlation coefficient as well as fuzzy methods, respectively. Further, the probability approach is used to find the association and availability of subfeature data from the dimensional reduced database. The experimental results show the proposed framework identifies and selects both feature and subfeature data with the effectiveness of the new class. The comparison results of several classifiers on several datasets are explained in the experimental section.

Relation classification (RC) task is one of fundamental tasks of information extraction, aiming to detect the relation information between entity pairs in unstructured natural language text and generate structured data in the form of entity-relation triple. Although distant supervision methods can effectively alleviate the problem of lack of training data in supervised learning, they also introduce noise into the data, and still cannot fundamentally solve the long-tail distribution problem of the training instances. In order to enable the neural network to learn new knowledge through few instances like humans, this work focuses on few-shot relation classification (FSRC), where a classifier should generalize to new classes that have not been seen in the training set, given only a number of samples for each class. To make full use of the existing information and get a better feature representation for each instance, we propose to encode each class prototype in an adaptive way from two aspects. First, based on the prototypical networks, we propose an adaptive mixture mechanism to add label words to the representation of the class prototype, which, to the best of our knowledge, is the first attempt to integrate the label information into features of the support samples of each class so as to get more interactive class prototypes. Second, to more reasonably measure the distances between samples of each category, we introduce a loss function for joint representation learning to encode each support instance in an adaptive manner. Extensive experiments have been conducted on FewRel under different few-shot (FS) settings, and the results show that the proposed adaptive prototypical networks with label words and joint representation learning has not only achieved significant improvements in accuracy, but also increased the generalization ability of few-shot RC models.

In this article, a semisupervised weighting method for feature dimension based on entropy is proposed for classification, dimension reduction, and correlation analysis. For real-world data, different feature dimensions usually show different importance. Generally, data in the same class are supposed to be similar, so their entropy should be small; and those in different classes are supposed to be dissimilar, so their entropy should be large. According to this, we propose a way to construct the weights of feature dimensions with the whole entropy and the innerclass entropies. The weights indicate the contribution of their corresponding feature dimensions in classification. They can be used to improve the performance of classification by giving a weighted distance metric and can be applied to dimension reduction and correlation analysis as well. Some numerical experiments are given to test the proposed method by comparing it with some other representative methods. They demonstrate that the proposed method is feasible and efficient in classification, dimension reduction, and correlation analysis.

In multi-dimensional classification (MDC), a number of class variables are assumed in the output space with each of them specifying the class membership w.r.t. one heterogeneous class space. One major challenge in learning from MDC examples lies in the heterogeneity of class spaces, where the modeling outputs from different class spaces are not directly comparable. To tackle this problem, we propose a new strategy named decomposed label encoding which enables modeling alignment for MDC in an encoded label space derived from one-vs-one (OvO) decomposition. Specifically, the original MDC output space is transformed into a ternary encoded label space by conducting OvO decomposition w.r.t. each class space. Then, the manifold structure in the feature space is exploited to enrich the labeling information in the encoded label space. Finally, the predictive model is induced by fitting the metric-aligned modeling outputs with enriched labeling information. Extensive experiments over twenty benchmark data sets clearly show the superiority of the proposed MDC strategy against state-of-the-art approaches.

In transfer learning model, the source domain samples and target domain samples usually share the same class labels but have different distributions. In general, the existing transfer learning algorithms ignore the interclass differences and intraclass similarities across domains. To address these problems, this article proposes a transfer learning algorithm based on discriminative Fisher embedding and adaptive maximum mean discrepancy (AMMD) constraints, called discriminative Fisher embedding dictionary transfer learning (DFEDTL). First, combining the label information of source domain and part of target domain, we construct the discriminative Fisher embedding model to preserve the interclass differences and intraclass similarities of training samples in transfer learning. Second, an AMMD model is constructed using atoms and profiles, which can adaptively minimize the distribution differences between source domain and target domain. The proposed method has three advantages: 1) using the Fisher criterion, we construct the discriminative Fisher embedding model between source domain samples and target domain samples, which encourages the samples from the same class to have similar coding coefficients; 2) instead of using the training samples to design the maximum mean discrepancy (MMD), we construct the AMMD model based on the relationship between the dictionary atoms and profiles; thus, the source domain samples can be adaptive to the target domain samples; and 3) the dictionary learning is based on the combination of source and target samples which can avoid the classification error caused by the difference among samples and reduce the tedious and expensive data annotation. A large number of experiments on five public image classification datasets show that the proposed method obtains better classification performance than some state-of-the-art dictionary and transfer learning methods. The code has been available at https://github.com/shilinrui/DFEDTL.

This paper is a unique effort to resolve the scaling issues of machine learning using the multi-node environment of Big Data. For this purpose, it incorporates the concept of machine learning with big data. Machine learning is a branch of Artificial Intelligence which trains the computers to learn without being explicitly programmed. Machine learning works on the development of computer programs and software that work according to the input dataset. In machine learning, classification is used to identify the class of instances. In this the category or class or a group provides the new observation to which it belongs on the basis of the training dataset. Training datasets contain those instances whose class is already known. An algorithm that implements classification is known as a classifier. The main aim of this paper is to evaluate how different classifiers work on two different datasets, one is a real time dataset of MP Election nomination (2018), while the other is a standard Breast Cancer dataset, released by University of Wisconsin (1995). There are various classifiers available like Decision Trees, K-Nearest Neighbors, Support Vector Machine, Logistic Regression, Naive Bayes etc. Different classifiers work differently on different datasets, giving different accuracy according to the kind and the size of the dataset. So we have used different classifiers for our work. When we were applying machine learning classifier on a real-time dataset, i.e. MPSEC dataset, Accuracy goes down. Similarly, these classifiers apply to Standard Breast Cancer dataset then results show outstanding accuracy. This research work shows the exciting patterns to study the case of a real-time and standard dataset using different machine learning classifiers with a combination of big data. This research work also incorporates the concept of parallelization using a multi-nodes environment which is able to deal with scalability issues.

Shyness trait is a double-edged personality trait, and could pose a risk in early childhood for later adjustment difficulties. Therefore, it is necessary to pay attention to children with shyness trait and properly supervise them in early education, where the key problem is shyness trait recognition. Although some psychological methods of shyness measurement have been presented, they are high-cost and can only get a one-sided result limited by the targeted subjects. To develop an automated method of shyness trait recognition, we collect a dataset, containing online writing data of 1,754 schoolchildren from an educational website and ground truth labels obtained by a professional scale. The natural implicitness makes shyness trait difficult to be observed from a single point of view and the class imbalanced problem increases the challenge. In this paper, a novel shyness trait recognition framework is proposed, which extracts multi-view features of online writing, including document-view, sentence-view and temporality-view ones. Different strategies with different features are applied to each single-view prediction and the multi-view prediction is made by a weighted voting ensemble. To verify the effectiveness, extensive experiments are conducted, demonstrating that the multi-view prediction significantly outperforms each single-view prediction and some advanced models of multi-view learning.

The first thing addressed in this paper is Plant fruit recognition and its feature extraction, it plays vital role in Agriculture. The point is to construct an exact, quick and solid framework utilizing CNN realities. Automatic fruit detection may reduce human efforts (counting number of fruits and manual identification). The second thing addressed in this paper is plant classification to avoid similarities in many plant fruits. Fruit classification may help fruit sellers to identify and to differentiate various kind of fruits having same similarities. The proposed framework has applied convolutional Neural Net (CNN) to the undertakings of distinguishing natural fruit pictures. In any case, deep learning has been shown as of late to be an extremely incredible picture identification procedure, and CNN is a best in class way to deal with deep learning.

Extraction of information in enterprise applications due to complex distribution of information in large identical sources. In such a scenario, a single approach or method for extracting information text would limit information needs, which would require a high level of work and time. So, it is an open problem in data analysis application to construct a suitable classification model for unsupervised datasets which having multi level of information attributes. It is vital to analyse and discover the valuable pattern information to support the accurate classification. Selection of attributes and construction of relation model always is a challenging due the increasing attribute dimensionality. So, it is important to learn and select the best attributes patterns to enhance the information association through accurate classification. In this paper, we propose an ARP-RW method for construction of Attribute Relation Pattern (ARP) through computing attributes Relation Weightage (RW) to enhance the information classification. The ARP-RW method provides the solution to identifying the relation class of the multi-attributes documents. It initially constructs an attribute relation class matrix (ARCM) utilizing the set of data collection and the RW value of the MLA. Later, it builds an ARP based on the MLA selected to perform accurate classification. The evaluation of the proposal method using the datasets of UCI Machine Learning Repository shows improvisation in comparison the existing approaches.

Machine learning research, particularly in genomics, is often based on wide shaped datasets, i.e. datasets having a large number of features, but a small number of samples. Such configurations raise the possibility of chance influence (the increase of measured accuracy due to chance correlations) on the learning process and the evaluation results. Prior research underlined the problem of generalization of models obtained based on such data. In this paper, we investigate the influence of chance on prediction and show its significant effects on wide shaped datasets. First, we empirically demonstrate how significant the influence of chance in such datasets is by showing that prediction models trained on thousands of randomly generated datasets can achieve high accuracy. This is the case even when using cross-validation. We then provide a formal analysis of chance influence and design formal chance influence estimators based on the dataset parameters, namely its sample size, the number of features, the number of classes and the class distribution. Finally, we provide an in-depth discussion of the formal analysis including applications of the findings and recommendations on chance influence mitigation.

In unsupervised domain adaptation (UDA), a classifier for the target domain is trained with massive true-label data from the source domain and unlabeled data from the target domain. However, collecting fully-true-label data in the source domain is high-cost and sometimes impossible. Compared to the true labels, a complementary label specifies a class that a pattern does not belong to, hence collecting complementary labels would be less laborious than collecting true labels. Thus, in this paper, we propose a novel setting that the source domain is composed of complementary-label data, and a theoretical bound for it is first proved. We consider two cases of this setting, one is that the source domain only contains complementary-label data (completely complementary unsupervised domain adaptation, CC-UDA), and the other is that the source domain has plenty of complementary-label data and a small amount of true-label data (partly complementary unsupervised domain adaptation, PC-UDA). To this end, a complementary label adversarial network} (CLARINET) is proposed to solve CC-UDA and PC-UDA problems. CLARINET maintains two deep networks simultaneously, where one focuses on classifying complementary-label source data and the other takes care of source-to-target distributional adaptation. Experiments show that CLARINET significantly outperforms a series of competent baselines on handwritten-digits-recognition and objects-recognition tasks., arXiv admin note: text overlap with arXiv:2007.14612

Available data in machine learning applications is becoming increasingly complex, due to higher dimensionality and difficult classes. There exists a wide variety of approaches to measuring complexity of labeled data, according to class overlap, separability or boundary shapes, as well as group morphology. Many techniques can transform the data in order to find better features, but few focus on specifically reducing data complexity. Most data transformation methods mainly treat the dimensionality aspect, leaving aside the available information within class labels which can be useful when classes are somehow complex. This paper proposes an autoencoder-based approach to complexity reduction, using class labels in order to inform the loss function about the adequacy of the generated variables. This leads to three different new feature learners, Scorer, Skaler and Slicer. They are based on Fisher's discriminant ratio, the Kullback-Leibler divergence and least-squares support vector machines, respectively. They can be applied as a preprocessing stage for a binary classification problem. A thorough experimentation across a collection of 27 datasets and a range of complexity and classification metrics shows that class-informed autoencoders perform better than 4 other popular unsupervised feature extraction techniques, especially when the final objective is using the data for a classification task., This paper has been accepted for publication by IEEE Transactions on Pattern Analysis and Machine Intelligence

Domain adaptation (DA) refers to generalize a learning technique across the source domain and target domain under different distributions. Therefore, the essential problem in DA is how to reduce the distribution discrepancy between the source and target domains. Typical methods are to embed the adversarial learning technique into deep networks to learn transferable feature representations. However, existing adversarial related DA methods may not sufficiently minimize the distribution discrepancy. In this paper, a domain adaptation method minimum adversarial distribution discrepancy (MADD) is proposed by combining feature distribution with adversarial learning. Specifically, we design a novel divergence metric loss, named maximum mean discrepancy based on conditional entropy (MMD-CE), and embed it in the adversarial domain adaptation network. The proposed MMD-CE loss can address two problems: 1) the misalignment from different class distributions between domains, 2) the equilibrium challenge issue in adversarial domain adaptation. Comparative experiments on Office-31, ImageCLEF-DA, and Office-Home datasets with state-of-the-art methods show that, our method has some advantageous performances.

One-class collaborative filtering (OCCF) problems are prevalent in real-world recommender systems, such as news and music recommendation, but suffer from sparsity issues and particularly lack negatively preferred items. To address the challenge, the state-of-the-art algorithms assigned uninteracted items smaller weights of being negative and performed low-rank approximation over a user-item interaction matrix. However, the prior ratings of uninteracted items are usually set to zero but may not well-defined. To avert the prior ratings of uninteracted items, in this paper, we propose a novel ranking-based implicit regularizer by hypothesizing that users preference scores for uninteracted items are close to each other. The regularizer is then used in a ranking-based OCCF framework to penalize large difference of preference scores between uninteracted items. To optimize model parameters in this framework, we develop a scalable alternating least square algorithm and coordinate descent algorithm, whose time complexity is linearly proportional to data size. Finally, we extensively evaluate the proposed algorithms with six real-world datasets. The results show that the proposed regularizer significantly improves recommendation quality of ranking-based OCCF algorithms, such as BPRMF and RankALS. Moreover, the ranking-based framework with the proposed regularizer outperforms the state-of-the-art recommendation algorithms for implicit feedback.

Over the last decade Deep learning-based models surpasses classical machine learning models in a variety of text classification tasks. The primary challenge with text classification is determining the most appropriate deep learning classifier. Numerous research initiatives incorporated ensemble learning to boost the performance, minimize errors and avoid overfitting. However, the performance of the ensemble-methods is limited by the baseline classifiers and the fusion method. The current study makes the following contributions: First, it proposes a new meta-learning ensemble method that fuses baseline deep learning models using 2-tiers of meta-classifiers. Second, it conducts several experiments on six public benchmark datasets to evaluate the performance of the proposed ensemble. For each benchmark dataset, committees of different deep baseline classifiers are trained, and their best performance is compared with the performance of the proposed ensemble. Furthermore, the paper extends the results by comparing the performance of the proposed ensemble method to other state-of-the-art ensemble methods. The findings indicate that the proposed ensemble method significantly improve the classification accuracy of the baseline deep models. Furthermore, the proposed method outperforms the state-of-art ensemble methods. Finally, using the probability distributions for each class label of the deep baseline models improves the performance of the proposed ensemble method.

The goal of supervised hashing is to construct hash mappings from collections of images and semantic annotations such that semantically relevant images are embedded nearby in the learned binary hash representations. Existing deep supervised hashing approaches that employ classification frameworks with a classification training objective for learning hash codes often encode class labels as one-hot or multi-hot vectors. We argue that such label encodings do not well reflect semantic relations among classes and instead, effective class label representations ought to be learned from data, which could provide more discriminative signals for hashing. In this article, we introduce Adaptive Labeling Deep Hashing (AdaLabelHash) that learns binary hash codes based on learnable class label representations. We treat the class labels as the vertices of a K -dimensional hypercube, which are trainable variables and adapted together with network weights during the backward network training procedure. The label representations, referred to as codewords, are the target outputs of hash mapping learning. In the label space, semantically relevant images are then expressed by the codewords that are nearby regarding Hamming distances, yielding compact and discriminative binary hash representations. Furthermore, we find that the learned label representations well reflect semantic relations. Our approach is easy to realize and can simultaneously construct both the label representations and the compact binary embeddings. Quantitative and qualitative evaluations on several popular benchmarks validate the superiority of AdaLabelHash in learning effective binary codes for image search.

Ground military target recognition plays a crucial role in unmanned equipment and grasping the battlefield dynamics for military applications, but is disturbed by low-resolution and noisy-representation. In this paper, a recognition method, involving a novel visual attention mechanism-based Gabor region proposal sub-network (Gabor RPN) and improved refinement generative adversarial sub-network (GAN), is proposed. Novel central–peripheral rivalry 3D color Gabor filters are proposed to simulate retinal structures and taken as feature extraction convolutional kernels in low-level layer to improve the recognition accuracy and framework training efficiency in Gabor RPN. Improved refinement GAN is used to solve the problem of blurry target classification, involving a generator to directly generate large high-resolution images from small blurry ones and a discriminator to distinguish not only real images vs. fake images but also the class of targets. A special recognition dataset for ground military target, named Ground Military Target Dataset (GMTD), is constructed. Experiments performed on the GMTD dataset effectively demonstrate that our method can achieve better energy-saving and recognition results when low-resolution and noisy-representation targets are involved, thus ensuring this algorithm a good engineering application prospect.

Despite that pre-trained word embedding models have advanced a wide range of natural language processing applications, they ignore the contextual information and meaning within the text. In this paper, we investigate the potential of the pre-trained Arabic BERT (Bidirectional Encoder Representations from Transformers) model to learn universal contextualized sentence representations aiming to showcase its usefulness for Arabic text Multi-class categorization. We propose to exploit the pre-trained AraBERT for contextual text representation learning in two different ways, transfer learning model and feature extractor. On the one hand, we employ the Arabic BERT (AraBERT) model after fine-tuning its parameters on the OSAC datasets to transfer its knowledge for the Arabic text categorization. On the other hand, we inquire into AraBERT performance, as a feature extractor model, by combining it with several classifiers, including CNN, LSTM, Bi-LSTM, MLP, and SVM. Finally, we conduct an exhaustive set of experiments comparing two BERT models, namely AraBERT and multilingual BERT. The findings show that the fine-tuned AraBERT model accomplishes state-of-the-art performance results and attains up to 99% in terms of F1-score and accuracy.

Traditional Deep Neural Network (DNN) security is mostly related to the well-known adversarial input example attack.Recently, another dimension of adversarial attack, namely, attack on DNN weight parameters, has been shown to be very powerful. Asa representative one, the Bit-Flip based adversarial weight Attack (BFA) injects an extremely small amount of faults into weight parameters to hijack the executing DNN function. Prior works of BFA focus on un-targeted attacks that can hack all inputs into a random output class by flipping a very small number of weight bits stored in computer memory. This paper proposes the first work oftargetedBFA based (T-BFA) adversarial weight attack on DNNs, which can intentionally mislead selected inputs to a target output class. The objective is achieved by identifying the weight bits that are highly associated with classification of a targeted output through a class-dependent weight bit searching algorithm. Our proposed T-BFA performance is successfully demonstrated on multiple DNN architectures for image classification tasks. For example, by merely flipping 27 out of 88 million weight bits of ResNet-18, our T-BFA can misclassify all the images from Hen class into Goose class (i.e., 100% attack success rate) in ImageNet dataset, while maintaining 59.35% validation accuracy.

With the memory-resource-limited constraints, class-incremental learning (CIL) usually suffers from the ``catastrophic forgetting'' problem when updating the joint classification model on the arrival of newly added classes. To cope with the forgetting problem, many CIL methods transfer the knowledge of old classes by preserving some exemplar samples into the size-constrained memory buffer. To utilize the memory buffer more efficiently, we propose to keep more auxiliary low-fidelity exemplar samples, rather than the original real-high-fidelity exemplar samples. Such a memory-efficient exemplar preserving scheme makes the old-class knowledge transfer more effective. However, the low-fidelity exemplar samples are often distributed in a different domain away from that of the original exemplar samples, that is, a domain shift. To alleviate this problem, we propose a duplet learning scheme that seeks to construct domain-compatible feature extractors and classifiers, which greatly narrows down the above domain gap. As a result, these low-fidelity auxiliary exemplar samples have the ability to moderately replace the original exemplar samples with a lower memory cost. In addition, we present a robust classifier adaptation scheme, which further refines the biased classifier (learned with the samples containing distillation label knowledge about old classes) with the help of the samples of pure true class labels. Experimental results demonstrate the effectiveness of this work against the state-of-the-art approaches. We will release the code, baselines, and training statistics for all models to facilitate future research.

(Aim) To detect COVID-19 patients more accurately and more precisely, we proposed a novel artificial intelligence model. (Methods) We used previously proposed chest CT dataset containing four categories: COVID-19, community-acquired pneumonia, secondary pulmonary tuberculosis, and healthy subjects. First, we proposed a novel VGG-style base network (VSBN) as backbone network. Second, convolutional block attention module (CBAM) was introduced as attention module into our VSBN. Third, an improved multiple-way data augmentation method was used to resist overfitting of our AI model. In all, our model was dubbed as a 12-layer attention-based VGG-style network for COVID-19 (AVNC) (Results) This proposed AVNC achieved the sensitivity/precision/F1 per class all above 95%. Particularly, AVNC yielded a micro-averaged F1 score of 96.87%, which is higher than 11 state-of-the-art approaches. (Conclusion) This proposed AVNC is effective in recognizing COVID-19 diseases.

Task-related component analysis (TRCA) has been applied successfully in the recently popular SSVEP target recognition methods. However, a spatial filter is trained for each class in TRCA, and the training of each filter uses only the training data of the corresponding class. Therefore the information between classes is ignored in the training process, which leads to classification inefficiency. Aiming at solving this defect in TRCA, we proposed a twodimensional locality preserving projections (2DLPP) method and a two-dimensional linear discriminant analysis (2DLDA) method based on the 2-Norm form of the Pearson’s correlation coefficient. The 2DLPP and 2DLDA methods can simultaneously use the samples of all categories to train the spatial filters so that these two methods can make use of the information between classes to some extent. We also showed that the 2DLPP method and the 2DLDA method performed significantly better than the multiset canonical correlation analysis (MsetCCA), extended CCA (eCCA) and TRCA methods with two public datasets. Therefore, the proposed methods based on 2DLPP or 2DLDA can make more efficient use of sample information and have a great potential for SSVEP target recognition.

Most of the recent image segmentation methods have tried to achieve the utmost segmentation results using large-scale pixel-level annotated data sets. However, obtaining these pixel-level annotated training data is usually tedious and expensive. In this work, we address the task of semisupervised semantic segmentation, which reduces the need for large numbers of pixel-level annotated images. We propose a method for semisupervised semantic segmentation by improving the confidence of the predicted class probability map via two parts. First, we build an adversarial framework that regards the segmentation network as the generator and uses a fully convolutional network as the discriminator. The adversarial learning makes the prediction class probability closer to 1. Second, the information entropy of the predicted class probability map is computed to represent the unpredictability of the segmentation prediction. Then, we infer the label-error map of the segmentation prediction and minimize the uncertainty on misclassified regions for unlabeled images. In contrast to existing semisupervised and weakly supervised semantic segmentation methods, the proposed method results in more confident predictions by focusing on the misclassified regions, especially the boundary regions. Our experimental results on the PASCAL VOC 2012 and PASCAL-CONTEXT data sets show that the proposed method achieves competitive segmentation performance.

Heterogeneous Information Network(HIN) collecitve classification studies the problem of predicting labels for one type of nodes in a HIN which contains multiple types of nodes multiple types of links among them. Previous studies have revealed that exploiting relative importance of links is quite useful to improve node classification performance as connected nodes tend to have similar labels. Most existing approaches exploit the relative importance of links either by directly counting the number of connections among nodes or by learning the weight of each type of link from labeled data only. However, these approaches either neglect the importance of types of links to the class labels or may lead to overfitting problem. We propose a Tensor-based Markov chain (T-Mark) approach, which is able to automatically and simultaneously predict the labels for unlabeled nodes and give the relative importance of types of links that actually improve the classification accuracy. Specifically, we build two tensor equations by using the HIN and features of nodes from both labeled and unlabeled data. A Markov chain-based model is proposed and it is solved by an iterative process to obtain the stationary distributions. Theoretical analyses of the existence and uniqueness of such probability distributions are given. Extensive experimental results demonstrate that T-Mark is able to achieve superior performance in the comparison and obtain reasonable relative importance of links.

Generative adversarial networks (GANs) are subject to catastrophic forgetting when learning stream of data. Inspired by the knowledge of neuroscience, this paper develops a Memory Formation System (MFS) to establish memory for GANs. MFS is composed of three modules, including the identifier, weights upgrade (WU), and weights reactivate (WR). These modules simulate the function of encoding, consolidating, and retrieving in memory formation of human. Identifier creates indexes to label continuous tasks and these indexes are used as a cue when the corresponding tasks are recalled. WU and WR work as synaptic consolidation and system consolidation, respectively. In WU, a novel method, Weight Saliency Measure (WSM) is proposed to measure the saliency of weight. Valuable weights are protected and invaluable weights are released for update when GAN is trained for the new task. In WR, traditional data-replay methods are improved by selecting k representatives in each class and their feature vectors are restored. When pseudo data of this class are regenerated, only the samples with distance less than a threshold can be used in future training. Experimental results based on the testing of continual image generation and continual 3D reconstruction show that MFS can establish a memory system to handle the catastrophic forgetting problem effectively.

In this article, fault diagnosis is of great significance for system health maintenance. For real applications, diagnosis accuracy suffers from unbalanced data patterns, where normal data are usually abundant than anomaly ones, leading to tremendous diagnosis obstacles. Therefore, it is challenging to use only normal data for fault diagnosis under this imbalanced condition. In addition, a single fault diagnosis model can only conduct one fault diagnosis task in most of cases. Accordingly, a one-class generative adversarial detection (OCGAD) framework based on semisupervised learning is proposed to learn one-class latent knowledge for dealing with multiple semisupervised fault diagnosis tasks, i.e., fault detection using only normal knowledge learning, novelty detection from unknown conditional data, and fault classification with unlabeled data. A bi-directional generative adversarial network (Bi-GAN) is first trained with only normal data. A one-class support vector machine is then established using features exacted by Bi-GAN from signals acquired from an attitude sensor for multifunctional fault detection. The presented OCGAD model is validated using an industrial robot with experiments of three fault detection tasks. The results demonstrate that the present model has good performance for dealing with multiple semisupervised diagnosis problems. info:eu-repo/semantics/publishedVersion

Detecting and classifying the occurrence of multiple class events is important in enhancing the wide-area situational awareness ability of power distribution systems. Multiple class event classification based on voltage magnitude measurement has not been largely deployed due to the inadequate elucidation between the occurrence of events and the corresponding eigenvalue perturbations in the measurement data. Herein, a detection and classification method for multiple class events in power distribution systems is proposed based on an eigenvalue fluctuation model. We first elucidated the interrelation between event occurrence and the eigenvalue fluctuations in the system voltage magnitude measurement data. Based on the discovery that different classes of events lead to their respective data eigenvalue behavior, we proposed detection criteria CSRL, CMLP1 and CMLP2 to quantitatively assess these behaviors. Since the different value changes in CSRL, CMLP1 and CMLP2 can indicate the occurrence of different classes of events, we trained and utilized a supervised classifier to achieve multiple event detection and classification. Simulation tests are performed in 7 different test feeders. Eight classes of events can be accurately detected and classified with only voltage magnitude measurement data. An 80% correct detection rate can be obtained with only a 20% measurement device penetration rate.

Applications challenged by the joint problem of concept drift and class imbalance are attracting increasing research interest. This paper proposes a novel Reinforcement Online Active Learning Ensemble for Drifting Imbalanced data stream (ROALE-DI). The ensemble classifier has a long-term stable classifier and a dynamic classifier group which applies a reinforcement mechanism to increases the weight of the dynamic classifiers, which perform better on the minority class, and decreases the weight of the opposite. When the data stream is class imbalanced, the classifiers will lack the training samples of the minority class. To supply training samples, when creating a new classifier, the labeled instances buffer is used to provide instances of the minority class. Then, a hybrid labeling strategy that combines the uncertainty strategy and imbalance strategy is proposed to define whether to obtain the real label of an instance. An experimental evaluation compares the classification performance of the proposed method with semi-supervised and supervised algorithms on both real-world and synthetic data streams. The results show that the ROALE-DI achieves higher Area Under the ROC Curve (AUC) and accuracy values with even fewer real labels, and the labeling cost dynamically adjusts according to the concept drift and class imbalance ratio.

Classical fuzzy rough set often uses fuzzy rough dependency as an evaluation function of feature selection. However, this function only retains the maximum member- ship degree of a sample to one decision class, it can not describe the classification error. Therefore, in this work, a novel criterion function for feature selection is proposed to overcome this weakness. To characterize the classification error rate, we first introduce a class of irreflexive and symmetric fuzzy binary relations to redefine the concepts of fuzzy rough approximations. Then, we propose a novel concept of dependency: inner product dependency to describe the classification error, and construct a criterion function to evaluate the importance of candidate features. The proposed criterion function not only can maintain a maximum dependency function, but also guarantees the minimum classification error. The experimental analysis shows that the proposed criterion function is effective for data sets with a large overlap between different categories.

In this article, a composite anti-disturbance control of disturbance rejection and attenuation for a class of nonlinear positive systems is proposed for the first time to achieve both positivity and stability. The disturbance observer-based control (DOBC) method is utilized to compensate for the effects of the modeled disturbances, while the L₁ control strategy is adopted to suppress the effects of the disturbances which are not compensable to satisfy the L₁ index performance. Besides, two analysis schemes of the closed-loop system are presented. The first scheme is commonly used and is easy for analysis, but it has some conservativeness. The second scheme is proposed novelly with less conservativeness, which releases the positivity constraint of disturbance estimation errors. The corresponding theorems and algorithms are proposed for necessary conditions and solutions. Finally, two examples are presented to illustrate the effectiveness of the proposed control schemes.

Weakly supervised temporal action detection aims at localizing temporal positions of action instances in untrimmed videos with only action class labels. In general, previous methods individually classify each frame based on the appearance information and the short-term motion information, and then integrate consecutive high-response action frames into entities which serve as detected action instances. However, the long-range temporal dependencies between action frames are not fully utilized, and the detection results are more likely to be trapped in the most discriminative action segments. To alleviate this issue, we propose a novel two-branch (i.e., the coarse detection branch and the refining detection branch) detection framework with learning the long-range temporal dependencies for obtaining more accurate detection results, where only action class labels are required. The coarse detection branch is used to localize the most discriminative segments of action instances based on a typical multi-instance learning paradigm under the supervision of action class labels, whereas the refining detection branch is expected to localize the less discriminative segments of action instances via learning the long-range temporal dependencies between frames based on the proposed Transformer-style architecture and learning strategies. This collaboration mechanism takes full advantage of complementary information from the provided action class labels and the natural temporal dependencies between action frames, forming a more comprehensive solution. Consequently, our method obtains more precise detection results. Expectedly, the proposed method outperforms recent weakly supervised temporal action detection methods on dataset THUMOS14 and ActivityNet measured by mAP@tIoU and AR@AN.

MicroRNAs (miRNAs) are a class of non-coding RNAs that play critical role in many biological processes, such as cell growth, development, differentiation and aging. Increasing studies have revealed that miRNAs are closely involved in many human diseases. Therefore, the prediction of miRNA-disease associations is of great significance to the study of the pathogenesis, diagnosis and intervention of human disease. However, biological experimentally methods are usually expensive in time and money, while computational methods can provide an efficient way to infer the underlying disease-related miRNAs. In this study, we propose a novel method to predict potential miRNA-disease associations, called SVAEMDA. Our method mainly consider the miRNA-disease association prediction as semi-supervised learning problem. SVAEMDA integrates disease semantic similarity, miRNA functional similarity and respective Gaussian interaction profile (GIP) similarities. The integrated similarities are used to learn the representations of diseases and miRNAs. SVAEMDA trains a variational autoencoder based predictor by using known miRNA-disease associations, with the form of concatenated dense vectors. Reconstruction probability of the predictor is used to measure the correlation of the miRNA-disease pairs. Experimental results show that SVAEMDA outperforms other stat-of-the-art methods. AUC values of SVAEMDA of global leave-one-out cross validation (LOOCV) and 5-fold cross validation (5-fold CV) are 0.9464 and 0.9428 respectively. In addition, case studies of three common human diseases indicate that SVAEMDA obtains 100 percent of the top 50 predicted candidates in the benchmark databases. Therefore, SVAEMDA can efficiently and accurately predict the potential associations between diseases and miRNAs.

Imbalanced class distribution is an inherent problem in many real-world classification tasks where the minority class is the class of interest. Many conventional statistical and machine learning classification algorithms are subject to frequency bias, and learning discriminating boundaries between the minority and majority classes could be challenging. To address the class distribution imbalance in deep learning, we propose a class rebalancing strategy based on a class-balanced dynamically weighted loss function where weights are assigned based on the class frequency and predicted probability of ground-truth class. The ability of dynamic weighting scheme to self-adapt its weights depending on the prediction scores allows the model to adjust for instances with varying levels of difficulty resulting in gradient updates driven by hard minority class samples. We further show that the proposed loss function is classification calibrated. Experiments conducted on highly imbalanced data across different applications of cyber intrusion detection (CICIDS2017 data set) and medical imaging (ISIC2019 data set) show robust generalization. Theoretical results supported by superior empirical performance provide justification for the validity of the proposed dynamically weighted balanced (DWB) loss function.

Semi-supervised learning aims to learn prediction models from both labeled and unlabeled samples. There has been extensive research in this area. Among existing work, generative mixture models with Expectation-Maximization (EM) is a popular method due to clear statistical properties. However, existing literature on EM-based semi-supervised learning largely focuses on unstructured prediction, assuming that samples are independent and identically distributed. Studies on EM-based semi-supervised approach in structured prediction is limited. This paper aims to fill the gap through a comparative study between unstructured and structured methods in EM-based semi-supervised learning. Specifically, we compare their theoretical properties and find that both methods can be considered as a generalization of self-training with soft class assignment of unlabeled samples, but the structured method additionally considers structural constraint in soft class assignment. We conducted a case study on real-world flood mapping datasets to compare the two methods. Results show that structured EM is more robust to class confusion caused by noise and obstacles in features in the context of the flood mapping application.

Question Answering Systems (QAS) are rising solutions providing exact and precise answers to natural questions. Duplicate Question Detection (DQD), which aims to reuse previous answers, has shown its ability to improve user experience and reduce significantly the response time. However, few Arabic QAS integrate solutions able to detect duplicate questions in their workflow. In this paper, we build a DQD method based on contextual word representation, question classification and forward/backward structured self attention. First, we extract contextual word representation Embeddings from Language Models (ELMo) to map questions into a vector space. Next, we train two models to classify question embedding according to two taxonomies: Hamza et al. and Li & Roth. Then, we introduce a class label matching step to filter out questions that have different class labels. Finally, we propose a Bidirectional Attention Bidirectional LSTM (BiAttention BiLSTM) model that focuses only on keywords to predict whether a question pair is a duplicate or not. We also apply a data augmentation strategy based on symmetry, reflexivity, and transitivity relations to improve the generalization of our model. Various experimentations are performed to evaluate the impact of question classification and pre-processing step on DQD model. The obtained results show that our model achieves good performances as compared to the baseline results.

Poetry has a prominent history in Arabic literature. The classical Arabic poetry has 16 m that vary in rhythm and target purpose. Chanting a poem eloquently requires knowing the poem’s meter and obtaining a diacritized version of its verses (letters inscribed with their short vowels); diacritics are often not inscribed in Arabic texts. This work proposes solutions to classify input Arabic text into the 16 poetry meters and prose. It also investigates the automatic diacritization of Arabic poetry. We adopt machine learning approach using a large dataset of 1657 k verses of poems and prose to develop neural networks to classify and diacritize Arabic poetry. We propose deep and narrow recurrent neural networks with bidirectional long short-term memory cells for solving these problems. The proposed model classifies the input text with an average accuracy of 97.27%, which is significantly higher than previous work. We also propose a solution that achieves an accuracy that approaches 100% when multiple verses of the same poem are available through predicting the class from the aggregate probabilities of the multiple verses. Diacritizing poetry is much harder than diacritizing prose due to the poet’s meticulous selection of phrases and relaxation of some diacritization rules.

Decision tree is a widely used non-parametric technique in machine learning, data mining and pattern recognition. It is simple to understand and interpret, however it faces challenges such as handling higher dimensional and class imbalanced datasets, over-fitting and instability. To overcome some of these issues, vertical partitioning approaches like serial partitioning, theme based partitioning are used in the literature. A vertical partitioning approach divides the feature set into subsets of features (blocks) and makes use of these subsets for subsequent tasks. In this work, we use the ideas of music rhythm tree to propose a novel vertical partitioning technique. It orders the features based on the average correlation strength of the features before partitioning the feature set. The proposed method is proved to be superior by showing an average of 13.8 % , 6 % , 9.8 % , 19.7 % , 9.4 % , and 29.4 % higher classification accuracy over C4.5, Random Forest, Bagging, Adaboost, an ensemble technique and a vertical partitioning technique respectively. Our empirical results on 15 datasets demonstrate that the proposed vertical partitioning method is more stable and better in handling class-imbalanced data. Finally, some popular statistical tests are conducted to validate the statistical significance of the results of the proposed method.

The research of semantic segmentation based on unsupervised domain adaptation greatly alleviates the high-cost bottleneck of manual annotation in deep learning. Inevitably domain gap limits the ability of target domain to learn knowledge from source domain. Previous domain alignment strategies aim to explore maximum domain-invariant space to enlarge knowledge that target domain learns. They based on the assumption that images are symmetrical in categories covering all predefined categories. However, in practical training, there is class asymmetry in image samples inter domains. This will lead to outlier categories and negative transfer, because it lacks the guidance of the corresponding category of the source domain. To tackle this issue, we propose a partial domain adaptive method on semantic segmentation to guide target model to learn category-level knowledge selectively. PAM (Partial Adaptive Map) module is introduced to motivate the target model to obtain more knowledge from non-outliers and less knowledge from outliers to avoid negative transfer. We further analyze the effectiveness of this method from the perspective of JS divergence. Our method has achieved significant improvement without additional discriminators. Experiments on general datasets GTA5 and SYNTHIA can compare with SOTA methods.

Detection of Alzheimer's Disease (AD) from neuroimaging data such as MRI through machine learning has been a subject of intense research in recent years. Recent success of deep learning in computer vision has progressed such research further. However, common limitations with such algorithms are reliance on a large number of training images, and requirement of careful optimization of the architecture of deep networks. In this paper, we attempt solving these issues with transfer learning, where the state-of-the-art VGG architecture is initialized with pre-trained weights from large benchmark datasets consisting of natural images. The network is then fine-tuned with layer-wise tuning, where only a pre-defined group of layers are trained on MRI images. To shrink the training data size, we employ image entropy to select the most informative slices. Through experimentation on the ADNI dataset, we show that with training size of 10 to 20 times smaller than the other contemporary methods, we reach state-of-the-art performance in AD vs. NC, AD vs. MCI, and MCI vs. NC classification problems, with a 4% and a 7% increase in accuracy over the state-of-the-art for AD vs. MCI and MCI vs. NC, respectively. We also provide detailed analysis of the effect of the intelligent training data selection method, changing the training size, and changing the number of layers to be fine-tuned. Finally, we provide Class Activation Maps (CAM) that demonstrate how the proposed model focuses on discriminative image regions that are neuropathologically relevant, and can help the healthcare practitioner in interpreting the model's decision making process.

The methods for combining multiple classifiers based on belief functions require to work with a common and complete (closed) Frame of Discernment (FoD) on which the belief functions are defined before making their combination.This theoretical requirement is however difficult to satisfy in practice because some abnormal (or unknown) objects that do not belong to any pre-defined class of the FoD can appear in real classification applications. The classifiers learnt using different attributes information can provide complementary knowledge which is very useful for making the classification but they are usually based on different FoDs. In order to clearly identify the specific class of the abnormal objects, we propose a new method for combination of classifiers working with incomplete frames of discernment, named CCIF for short. This is a progressive detection method that select and add the detected abnormal objects to the training data set. Because one pattern can be considered as an abnormal object by one classifier and be committed to a specific class by another one, a weighted evidence combination method is proposed to fuse the classification results of multiple classifiers. This new method offers the advantage to make a refined classification of abnormal objects, and to improve the classification accuracy thanks to the complementarity of the classifiers. Some experimental results are given to validate the effectiveness of the proposed method using real data sets.

Aims: Early detection of breast cancer has reduced many deaths. Earlier CAD systems used to be the second opinion for radiologists and clinicians. Machine learning and deep learning have brought tremendous changes in medical diagnosis and imagining. Background: Breast cancer is the most commonly occurring cancer in women and it is the second most common cancer overall. According to the 2018 statistics, there were over 2million cases all over the world. Belgium and Luxembourg have the highest rate of cancer. Objective: A method for breast cancer detection has been proposed using Ensemble learning. 2- class and 8-class classification is performed. Method: To deal with imbalance classification, the authors have proposed an ensemble of pretrained models. Result: 98.5% training accuracy and 89% of test accuracy are achieved on 8-class classification. Moreover, 99.1% and 98% train and test accuracy are achieved on 2 class classification. Conclusion: it is found that there are high misclassifications in class DC when compared to the other classes, this is due to the imbalance in the dataset. In the future, one can increase the size of the datasets or use different methods. In implement this research work, authors have used 2 Nvidia Tesla V100 GPU’s in google cloud platform.

Ninety-two Japanese elementary school sixth-graders (46 boys and 46 girls; 11–12 years old) learned the quota system as part of gender equality education. We used a group performance implicit association test (Mori, Uchida, and Imada, 2008) to evaluate the lesson's effect by assessing the children's image of “woman” before and after the class. The results showed that the image of “woman” among boys improved significantly from neutral to positive through the lesson. We also found that girls’ implicit image of “woman” was statistically higher than that of boys before learning, and was raised considerably after the class.

Object location in indoor environments is challenging when there is no physical contact, a field of view, reflective materials, and an excess of obstacles. Several works propose using Radio Frequency Identification technology (RFID) and machine learning methods to develop location systems in those situations. However, using an object as a target class slows learning and prediction in large-scale environments. To circumvent such problems, we proposed a location system that uses hierarchical classification. We divided the environment into regions to reduce the classifier’s training and the number of predicted classes. To define the regions, we used clustering techniques, indicating which clustering technique achieves the best performance in the proposed scenario. The main contribution of this work is a high-precision location system for large-scale environments. The results showed the proposed system’s implantation in a real environment with 400 target objects with 5 cm of location precision. The accuracy for region detection is 99.36%, while for identifying the object is 99.94%. Additionally, with the proposed hierarchical approach, we showed a reduction of 38.16% and 58.39% in processing time and classifier model size.

Heart disease seriously threatens human life due to high morbidity and mortality. Accurate prediction and diagnosis become more critical for early prevention, detection and treatment. The Internet of Medical Things (IoMT) and Artificial Intelligence (AI) support healthcare services in heart disease monitoring, prediction and diagnosis. However, most prediction models only predict whether people are sick, and rarely further determine the severity of the disease. In this paper, we propose a machine learning based prediction model to achieve binary and multiple classification heart disease prediction simultaneously. We first design a Fuzzy-GBDT algorithm combining fuzzy logic and Gradient Boosting Decision Tree (GBDT) to reduce data complexity and increase the generalization of binary classification prediction. Then, we integrate Fuzzy-GBDT with Bagging to avoid over-fitting. The Bagging-Fuzzy-GBDT for multi-classification prediction further classify the severity of heart disease. Evaluation results demonstrate the Bagging-Fuzzy-GBDT has excellent accuracy and stability in both binary and multiple classification predictions.

Wafer map defect recognition (WMDR) is an important part of the integrated circuit manufacturing system. Although deep neural networks (DNNs) have been applied successfully in WMDR, class imbalance and lack of data with class labels affect their performance significantly. A new generative adversarial network (GAN), multi-granularity generative adversarial network (MGGAN), is proposed for wafer map augmentation and enhancement. To alleviate instability and mode collapse of traditional GANs, the lightweight convolution and a two-way information interaction of three subnetworks are considered. MGGAN consists of an auxiliary feature extractor (AFE), a generator (G) and a discriminator (D) for wafer map generation and WMDR. Firstly, a pretrained deep convolutional neural network (CNN), ResNet101, is employed as AFE to extract multi-granularity features from wafer maps. Secondly, in order to improve effectiveness of the adversarial training, a feature matching term is considered in the objective function of the feature generator to minimize the statistical difference between the real images and the generated images. Finally, MGGAN has been successfully applied to WMDR. The experiment results on an industrial dataset WM-811K demonstrate that MGGAN outperforms other typical GANs that aim to solve class imbalance problems and gains better recognition performance than those state-of-the-art DNNs on WMDR. This indicates effectiveness of MGGAN in image enhancement and generation. MGGAN obtains an accuracy of 88.02% on the original data and the pre-trained ResNet101 obtains 93.43% on the data enhanced by MGGAN.

Time-series shapelets are discriminative subsequences, recently found effective for time series classification (tsc). It is evident that the quality of shapelets is crucial to the accuracy of tsc. However, major research has focused on building accurate models from some shapelet candidates. To determine such candidates, existing studies are surprisingly simple, e.g., enumerating subsequences of some fixed lengths, or randomly selecting some subsequences as shapelet candidates. The major bulk of computation is then on building the model from the candidates. In this paper, we propose a novel efficient shapelet discovery method, called bspcover, to discover a set of high-quality shapelet candidates for model building. Specifically, bspcover generates abundant candidates via Symbolic Aggregate approXimation with sliding window, then prunes identical and highly similar candidates via Bloom filters, and similarity matching, respectively. We next propose a $p$p-Cover algorithm to efficiently determine discriminative shapelet candidates that maximally represent each time-series class. Finally, any existing shapelet learning method can be adopted to build a classification model. We have conducted extensive experiments with well-known time-series datasets and representative state-of-the-art methods. Results show that bspcover speeds up the state-of-the-art methods by more than 70 times, and the accuracy is often comparable to or higher than existing works.

In many practical applications, the data are class imbalanced. Accordingly, it is very meaningful and valuable to investigate the classification of imbalanced data. In the framework of binary imbalanced data classification, the synthetic minority oversampling technique (SMOTE) is the best-known oversampling method. However, for each positive sample, SMOTE generates only k synthetic samples on the lines between the positive sample and its k -nearest neighbors, resulting in three drawbacks: (1) SMOTE cannot effectively extend the training field of positive samples; (2) the generated positive samples lack diversity; (3) SMOTE does not accurately approximate the probability distribution of the positive samples. Therefore, two binary imbalanced data classification methods named BIDC1 and BIDC2 based on diversity oversampling by generative models are proposed. The BIDC1 and BIDC2 conduct diversity oversampling using extreme learning machine autoencoder and generative adversarial network, respectively. Extensive experiments on 26 data sets are conducted to compare the two methods with 14 state-of-the-art methods using five metrics: F-measure, G-means, AUC-area, MMD-score, and Silhouette-score. The experimental results demonstrate that the two methods outperform the other 14 methods.

Although there is increased uptake of active learning approaches in especially developed countries, this is still lagging within the African context. The current study therefore focused on the implementation of group learning at Stellenbosch University, with several modifications versus our earlier, pilot study. Students freely formed small groups at the start of a 5-wk cardiovascular physiology lecture series and were apportioned three separate assignments to complete over this period. This included three in-class group-learning sessions, while students also completed group work outside class times. The active learning element was embedded within a constructive alignment framework. Our data revealed that 75% of the students felt that the in-class sessions was a good use of their time and that they progressively improved their assignment scores, i.e., 67.5 ± 9.3%, 72.4 ± 9.8%, and 76.1 ± 9.5% for the first, second, and final ones, respectively ( P < 0.0001). Moreover, the average class score for their final test (68 ± 15.1%) was higher when compared with the average class score (57.5 ± 19.4%) calculated for the previous 3 yr ( P < 0.0001). This study revealed two major findings: i.e., 1) students displayed a strong positive response regarding the adoption of in-class collaborative group work, and 2) the introduction of such active learning elements correlated with improved student assignment and test scores. Based on these findings, we propose additional modifications (including a shift to more formative assessments) to ensure even greater success with the roll-out of such active learning elements within the African context.

Artificial Neural networks (ANNs) are widely used to solve classification problems for many machine learning applications. When errors occur in the computational units of an ANN implementation due to for example radiation effects, the result of an arithmetic operation can be changed, and therefore, the predicted classification class may be erroneously affected. This is not acceptable when ANNs are used in many safety-critical applications, because the incorrect classification may result in a system failure. Existing error-tolerant techniques usually rely on physically replicating parts of the ANN implementation or incurring in a significant computation overhead. Therefore, efficient protection schemes are needed for ANNs that are run on a processor and used in resource-limited platforms. A technique referred to as Selective Neuron Re-Computation (SNRC), is proposed in this paper. As per the ANN structure and algorithmic properties, SNRC can identify the cases in which the errors have no impact on the outcome; therefore, errors only need to be handled by re-computation when the classification result is detected as unreliable. Compared with existing temporal redundancy-based protection schemes, SNRC saves more than 60 percent of the re-computation (more than 90 percent in many cases) overhead to achieve complete error protection as assessed over a wide range of datasets. Different activation functions are also evaluated. This research was supported by the National Science Foundation Grants CCF-1953961 and 1812467, by the ACHILLES project PID2019-104207RB-I00 and the Go2Edge network RED2018-102585-T funded by the Spanish Ministry of Science and Innovation and by the Madrid Community research project TAPIR-CM P2018/TCS-4496. Publicado