Search

Accurate diagnosis of Lithium-ion batteries (Li-ion batteries) degradation plays a critical role in improving the maintenance of energy storage technology. This paper presents a method based on a novel deep network model combined with a data transformation technique to diagnose Li-ion battery degradation modes. Different from conventional studies based on specific experimental and numerical methods to estimate and predict the degradation, the proposed method is based on data-driven approach, by leveraging datasets consisting of voltage/capacity curves, these were converted into incremental capacity (IC) curves and then transformed into images using the gramian angular summation field (GASF) technique. The study adopted two models: Inception-v3 and the proposed model, both underwent fine-tuning and a subsequent transfer learning process. Degradation modes, namely loss of lithium inventory (LLI) and the loss of active materials in both the positive (LAMPE) and negative electrodes (LAMNE), were diagnosed in relation to IC curves. Finally, the model was tested using two different datasets, and the results showed that the proposed method achieved high performance, especially across three Li-ion batteries, three degradation modes, three cells, and various cycles (totaling 378 cases) the proposed method outperformed in 233 cases, thereby outperforming other methods in comparison. Our method provides a flexible data-driven approach that accurately predicts various degradation modes across different cell chemistries throughout their lifespan. [ABSTRACT FROM AUTHOR]

Distributed generation sources (e.g. photovoltaic, wind power) are raised to fit the energy demands recently in proficient way. Further, adoption of renewable energy sources to the smart grid causes severe issues in quality of power signals. These problems can be overcome with correct identification and classification of power quality disturbances (PQDs) efficiently by the both producer and consumer. In this work, a new method for classification of PQDs based on signal processing technique and deep convolutional neural network (DCNN) is proposed. The proposed method compress the raw data of PQDs by using discrete wavelet transform (DWT) and multi-resolution analysis (MRA). Moreover, CNN has the ability to automatic feature selection and classification in a single stage instead of conventional methods i.e. artificial neural networks (ANNs). In this paper, DWT 1D-CNN is designed to extract and optimize multi-scale characteristics of disturbances. Further, 1-D convolutional, pooling and batch-normalization layers are added to facilitate normalize data, better generalization ability and, effective classification process against noisy data. A typical 11Kv distribution network is simulated in MATLAB to prove the validity of the proposed method in an essential part of smart grid. [ABSTRACT FROM AUTHOR]

In power transformers, accurate diagnosis prediction plays a crucial role in ensuring the reliability and efficiency of electrical systems. This paper presents an enhanced diagnostic prediction approach for power transformers by combining the Synthetic Minority Over-sampling Technique (SMOTE) and Deep Neural Networks (DNN). Our approach addresses the challenge of imbalanced data in power transformer diagnosis prediction by leveraging SMOTE to generate synthetic samples and balance the dataset. By effectively handling rare events through the augmentation of minority class instances, our predictive model becomes more robust. Additionally, using deep neural networks enables the exploration of complex patterns and relationships within the transformer dataset. The DNN architecture integrates multiple layers of learnable features, enabling the prediction model to extract high-level representations and achieve improved accuracy and performance. To evaluate the effectiveness of our proposed approach, we conducted extensive experiments using real-world power transformer datasets. Comparative analysis of the results demonstrates that the combination of SMOTE and DNN outperforms traditional classification techniques applied to imbalanced datasets. The proposed model achieves a higher training accuracy of 98.22 % and a testing accuracy of 94.6 % and mitigates the risks associated with misdiagnosis, contributing to power systems' overall reliability and maintenance efficiency. [ABSTRACT FROM AUTHOR]

This paper presents an effective and accurate method to analyze the network traffic risk of enterprise public cloud financial system by using deep learning algorithm. To build the analytical model, the collected data is preprocessed to extract relevant features that contribute to risk analysis. The pre-processed data is used to train the deep learning model. The model is designed to learn patterns and relationships in network traffic data. The network traffic risk analysis module is constructed by combining the deep learning model with the system architecture of the enterprise public cloud financial system. Through a trained deep learning model, the module is able to accurately identify abnormal behavior in network traffic and provide corresponding risk assessments. The results show that the proposed system module can accurately identify abnormal behavior in network traffic and provide risk assessment. [ABSTRACT FROM AUTHOR]

The rapid growth of online music education has led to increased security risks from cyber intrusions. This paper proposes public cloud networks oriented deep neural networks for effective intrusion detection in online music education environments. Specifically, a novel intrusion detection framework is developed, comprising fuzzy logic based feature selection, chronological salp swarm algorithm optimized deep belief networks, and gated recurrent unit integrated convolutional neural networks. Detailed methodologies are presented for the fuzzy logic system, chronological salp optimization, deep belief network architecture, and convolutional neural networks. Comprehensive experiments are conducted on the NSL-KDD and CICIDS2017 datasets. Various deep neural networks are evaluated and compared, including multi-layer perceptrons, convolutional neural networks, deep belief networks, recurrent neural networks, and the proposed models. Experimental results demonstrate that the proposed fuzzy feature selection and chronological salp swarm algorithm optimized deep belief network achieves a test accuracy of 97.33 %, outperforming other peer models. The gated recurrent unit integrated convolutional neural network obtains a test accuracy of 98.46 %, superior to state-of-the-art methods. While the experiments on the newly created dataset for intrusion detection in cloud-based online music education demonstrate that the proposed models outperform the benchmarks. The experiments verify the effectiveness of the proposed deep learning frameworks for intrusion detection in online music education cloud networks. [ABSTRACT FROM AUTHOR]

This paper proposes a new method for calculating joint-state posteriors of mixed-audio features using deep neural networks to be used in factorial speech processing models. The joint-state posterior information is required in factorial models to perform joint-decoding. The novelty of this work is its architecture which enables the network to infer joint-state posteriors from the pairs of state posteriors of stereo features. This paper defines an objective function to solve an underdetermined system of equations, which is used by the network for extracting joint-state posteriors. It develops the required expressions for fine-tuning the network in a unified way. The experiments compare the proposed network decoding results to those of the vector Taylor series method and show 2.3% absolute performance improvement in the monaural speech separation and recognition challenge. This achievement is substantial when we consider the simplicity of joint-state posterior extraction provided by deep neural networks. [ABSTRACT FROM AUTHOR]

Representing three-dimensional (3D) objects by multiple views has become a common solution to the problem of 3D object retrieval. It has gained excellent achievements because of its remarkable adaptability and flexibility. In this paper, we develop a multi-view-based Siamese convolutional neural network for 3D object retrieval. It consists of two sub-networks which have the same architecture and also share the same set of weights. First, we generate a set of RGB images and binary images for each 3D object to capture local and global features. Second, the two sub-networks take corresponding images as input and avoid camera constraint by using average fusion layers. The final compact descriptors are learned by integrating features of sub-networks and then used for retrieval. Experimental results on two benchmarks, the PSB dataset and ETH dataset, testify that our proposed method receives superior retrieval performance compared to state-of-the-art methods. [ABSTRACT FROM AUTHOR]

There is a growing demand for power from day to day. At present, the development achieved in power production from wind is highly significant. In this work, an optimized nonlinear neural network architectural model integrated with a modified firefly algorithm and particle swarm optimization is proposed to perform multistep wind-speed forecasting for specific target sites. Considering these aspects, this paper intends to predict wind speed, as its influence is high in generating wind power. The weights and bias values of the nonlinear neural network model are optimized employing the proposed optimization algorithm in order to achieve the minimum-error criterion. The computed results establish the effectiveness of the forecasting accuracy and the minimal error rate in comparison with existing methods available in the literature. [ABSTRACT FROM AUTHOR]

This paper proposes an Artificial Neural Network (ANN) model for AC flash-over voltage prediction of Silicone Rubber (SiR) insulators based on the experimental tests. In flash-over tests, the flash-over voltages of four different types of SiR insulators are measured under the uniform, longitudinal non-uniform, fan-shaped non-uniform and combined longitudinal, and fan-shaped non-uniform pollution conditions. The proposed ANN model is designed with six inputs data (insulator geometry and pollution parameters) and one output data (flash-over voltage). In order to validate the model, three different types of SiR insulators are tested under different pollution conditions. Then, their flash-over voltages are predicted using the proposed ANN model. The validation of the model shows that the absolute values of relative errors between flash-over voltages of test and prediction are less than 6%. This indicates a high efficiency of ANN technique in the flash-over voltage prediction of SiR insulators. [ABSTRACT FROM AUTHOR]

Multi-layer perceptron neural networks (MLP-NN) are widely utilized in forecasting applications. However, optimal training of these networks is still a challenge. A comprehensive assessment of MLP training approaches comprising of three stages is performed in the present paper. First, the prediction performance is evaluated using twelve training algorithms. Next, optimization algorithms are utilized to enhance the best obtained network parameters obtained from the first step and the performance of eight optimization algorithms is evaluated. Finally, a novel modification is used to improve the performance of the optimization algorithms. The proposed methodologies are applied to two case-studies and statistical metrics are employed for their efficiency evaluation. Wavelet transformation is used to extract the features which will be fed to the MLP-NN as input data. [ABSTRACT FROM AUTHOR]

• An enhanced Hybrid Intelligent System is proposed to predict the energy generated by a solar thermal system. • Supervised (neural networks) and unsupervised learning (clustering) are combined an applied to a real-world case study. • Best regression results are obtained when applying a Multilayer Perceptron trained with the Bayesian Regularization and Levenberg-Marquardt algorithms to k-means clustered datasets. There is wide acknowledgement that solar energy is a promising and renewable source of electricity. However, complementary sources are sometimes required, due to its limited capacity, in order to satisfy user demand. A Hybrid Intelligent System (HIS) is proposed in this paper to optimize the range of possible solar energy and power grid combinations. It is designed to predict the energy generated by any given solar thermal system. To do so, the novel HIS is based on local models that implement both supervised learning (artificial neural networks) and unsupervised learning (clustering). These techniques are combined and applied to a real-world installation located in Spain. Alternative models are compared and validated in this case study with data from a whole year. With an optimum parameter fit, the proposed system managed to calculate the solar energy produced by the panel with an error that was lower than 10−4 in 86% of cases. [ABSTRACT FROM AUTHOR]

• Interpretable DL Model: Our unique deep learning model interprets modulation mechanisms, efficiently extracting time and phase features with minimal resources. • Efficient Performance: Our approach achieves top performance on RML2016 and RML2018 datasets with fewer model parameters. • Fastest Embedded Speed: Outperforming 8 DL-based methods, our approach demonstrates the fastest speed on Jetson Nano and Raspberry Pi systems. Automatic Modulation Classification (AMC) constitutes a fundamental technology for enabling automatic demodulation in Cognitive Communication Systems (CCS). Due to the size, weight, and power (SWaP) constraints of embedded computers employed in CCS, there are limited computational and memory resources. While deep neural networks possess strong feature representation and high accuracy recognition capabilities, they usually come with a high number of network parameters and high computational complexity, thereby reducing the real-time processing ability of CCS. Therefore, neural network structures intended for CCS must be lightweight and computationally efficient. In this paper, we propose a high-performance and resource-friendly network model based on an analysis of the modulation mechanism of communication signals. The network extracts phase features and short-time features sequentially using directional convolutional filters. Long short-term memory (LSTM) units are then used to extract long-term features, and only one fully connected layer is used for classification. Experiments with a standard dataset consisting of 11 communication modulation types demonstrate that our proposed model achieves an accuracy greater than 84.5%, even when the signal-to-noise ratio (SNR) is 0 dB, and the model has only 29187 parameters. On a Jetson Nano embedded platform, the model achieves a processing speed of up to 375366 in-phase and quadrature samples/s. Overall, the results suggest that our proposed approach is both lightweight and highly efficient, making it more suitable for CCS applications. [Display omitted] [ABSTRACT FROM AUTHOR]

This paper presents the Weight-aware Approximate Multiplication Architecture (W-AMA) for Deep Neural Networks (DNNs). Considering the Gaussian-like weight distribution, it deploys an accuracy-configurable computing component to improve the computational efficiency. Two techniques for effectively integrating the W-AMA into DNN accelerator are presented: (1) A Cartesian Genetic Programming (CGP) based approximate multiplier is designed and selectable to compute the Least Significant Bit (LSB) for a higher accuracy mode. The Reward–Penalty-Coefficient (RPC) is proposed to achieve the internal-compensation. (2) The Hessian-Aware-Approximation (HAA) method is utilized for hybrid approximate modes cross-layer mapping. Based on the W-AMA, an energy-efficient DNN accelerator is proposed and evaluated on 28 nm technology. It can achieve the energy efficiency of 9.6 TOPS/W, and the computational energy efficiency can be improved by 1.5 × compared with the standard units, with an 0.52% accuracy loss on CIFAR-10 using ResNet-18. [Display omitted] • Two suggestions support the effective design of approximate circuits in neural networks. • The weight-aware approximate architecture enables multiplications with dynamically configurable accuracy. • A DNN accelerator using Hessian-aware-approximation maps each layer at its appropriate computational accuracy. [ABSTRACT FROM AUTHOR]

• Division-based Morris-Lecar (DBML) neuron model based on a new divider module and exponential terms. • Low-cost high-performance of original ML implementation using proposed DBML model by replacing complex non-linear function of original ML model with the multiplier-less exponential functions and simple division process of DBML. • Following the original ML model in terms of different spiking patterns and dynamical behaviors with employing DBML model. • Large-scale network implementation. Simulation, modeling, and hardware implementation of different parts of central nervous system (CNS) are very attractive research areas in terms of neuromorphic engineering aspects. Considering the neuronal parts of the human brain, the hardware implementation of neural models of spiking neural networks should be evaluated from preparing a simple efficient model and optimal hardware implementation. In this paper, Morris-Lecar (ML) model as a biological neuronal model with nonlinear differential equations and hyperbolic functions is discussed. This original model with non-linear complex functions needs high-computational high-cost hardware implementation on Field-Programmable Gate Arrays (FPGA). This paper presents a new divider module and exponential term (2 X ) to provide low-cost and high-speed implementation of original ML called Division-Based ML (DBML). DBML model suggests simpler equations using division-based exponential terms, logical shifts and simple arithmetic operations. Thus, high-performance neuronal design is presented following original ML model in terms of time domain and different dynamic behaviors. Global architecture based on digital hardware of the proposed DBML model including a novel divider module and Exponential-term evaluation [Display omitted] [ABSTRACT FROM AUTHOR]

• We proposed a 9-layer deep convolutional neural network for leaf disease identification. • We used six different types of image augmentation methods, and compared their validation performances using data augmentation and not using data augmentation. • Fine-tuning the batch size, dropout and epoch improves the performance of our model. • We validated that the proposed model performs better than transfer learning techniques. • Our model produced an average testing accuracy of 96.46%, which was better than the state-of-the-art machine learning approaches. In this paper, we proposed a novel plant leaf disease identification model based on a deep convolutional neural network (Deep CNN). The Deep CNN model is trained using an open dataset with 39 different classes of plant leaves and background images. Six types of data augmentation methods were used: image flipping, gamma correction, noise injection, principal component analysis (PCA) colour augmentation, rotation, and scaling. We observed that using data augmentation can increase the performance of the model. The proposed model was trained using different training epochs, batch sizes and dropouts. Compared with popular transfer learning approaches, the proposed model achieves better performance when using the validation data. After an extensive simulation, the proposed model achieves 96.46% classification accuracy. This accuracy of the proposed work is greater than the accuracy of traditional machine learning approaches. The proposed model is also tested with respect to its consistency and reliability. [ABSTRACT FROM AUTHOR]

Diabetic retinopathy (DR) is an illness occurring in the eye due to increase in blood glucose level. Among people in the age group of 70, 50% of deaths are attributed to diabetes. Early identification and appropriate treatment can reduce the loss of sight in many DR patients. Once the symptoms of DR are recognized, the severity of the disease should be evaluated for administering the right medication. This paper focuses on the classification of DR fundus images according to the severity of the disease using convolutional neural network with the application of suitable Pooling, Softmax and Rectified Linear Activation Unit (ReLU) layers to obtain a high level of accuracy. The performance of the proposed algorithm has been validated using Messidor database. In the case of healthy images, images of stage1, stage 2 and stage 3 of diabetic retinopathy, classification accuracies of 96.6% and 96.2%, 95.6% and 96.6% have been achieved. [ABSTRACT FROM AUTHOR]

Image quality enhancement is an important image processing task, wherein denoising an image is essential for accurate image diagnoses, because the presence of noise in an image produces incorrect information. This paper proposes double density wavelet transform based intelligent techniques for image denoising. This hybrid technique combines wavelet and neural networking methods, performed and validated using different standard images and extant denoising methods. These images were degraded using a variety of noise levels to simulate actual noise degradation. The results reveal the high-peak signal-to-noise ratio of the proposed system, which is thus considered effective compared to state-of-the-art techniques. [ABSTRACT FROM AUTHOR]

Robotic vehicles and their applications have gained widespread significance in recent times in order to automate every sphere of human activities. Besides, it invites a great deal of attention in environments where human presence or intervention becomes very difficult and in most cases not being possible. The paper therefore proposes a voice based movement and evolves a direction control scheme for a robotic vehicle in a Light Fidelity (Li-Fi) medium with a view to exploit its available enormous bandwidth. Choice of Li-Fi as a medium owes to its quick response time and ability to respond to voice based commands. It engages the role of an artificial neural network (ANN) for recognizing the syllables through Levenberg–Marquardt training algorithm. The scheme relies on two sets of instructions issued to the vehicle controlled by the Arduino controller, which allows acting in accordance to the instruction recognized by the voice recognition module. [ABSTRACT FROM AUTHOR]

Abstract With increasing demand for a renewable source of energy and the rapid development of smart cities and cluster computing, wind energy is gaining widespread significance on a global basis. There has been the unprecedented increase in annual installation of wind turbines reaching 60 GW for the first time in history from 51.7 GW of wind energy installation in 2014. Most of the high-power installations utilized double-fed induction generators (DFIG) due to low power losses, which reduced the cost of power converters required in the circuit. Nevertheless, natural factors, such as varying wind speed, influence the output of generators which results in the unsteady generation of power thus affecting the stability of the system. A neural network-based DFIG integrated with super capacitor energy storage system (SCESS) is proposed in this paper. Modeling and simulation have been carried out in MATLAB/Simulink and the observations indicate the significant improvement in the stability of the system. [ABSTRACT FROM AUTHOR]

Highlights • "Seddulbahir," the first Turkish national cyber-firewall system is proposed. • Artificial neural network radial basis function (RBF-NN) approach is used to generate security rules for 21 different attacks from 66 different network service flows. • National Cyber-technologies of developed countries are investigated. • Current state of the Turkey Cyber-infrastructure is highlighted. • Economic Impact of Volumetric DNS attacks on Turkey between 14th-24th December 2015 are exposed. • Required structure of the cyber-technologies to enhance country-wide cyber-security is elaborated. Abstract This research introduces "Seddulbahir," the first Turkish national cyber-firewall system. Proposed as the first of its kind, Seddulbahir resists possible cyber-attack threats against Turkey's internet infrastructure. Seddulbahir uses the artificial neural network radial basis function (RBF-NN) approach to generate rules that can detect 21 different attacks classified as probing, DoS (Denial of Service), U2R (User-to-Root), and R2L (Remote-to-Local) through the analysis of 66 different network service flows and based on three communication protocols. Results indicate that our proposed approach detects abnormal traffic with high efficiency, low cost, and has a wide detection range, providing accurate, flexible, and effective results in contrast to traditional methods. This paper also outlines the current state of cyberspace in Turkey, provides an analysis and detailed model of the latest cyber-attacks generated against Turkey, and investigates the working mechanisms of national cyber firewall systems used by developed countries such as Russia, China, and the United States. [ABSTRACT FROM AUTHOR]

Abstract Diabetic retinopathy (DR) is a common complication of diabetes and one of the major causes of blindness in the active population. Many of the complications of DR can be prevented by blood glucose control and timely treatment. Since the varieties and the complexities of DR, it is really difficult for DR detection in the time-consuming manual diagnosis. This paper is to attempt towards finding an automatic way to classify a given set of fundus images. We bring convolutional neural networks (CNNs) power to DR detection, which includes 3 major difficult challenges: classification, segmentation and detection. Coupled with transfer learning and hyper-parameter tuning, we adopt AlexNet, VggNet, GoogleNet, ResNet, and analyze how well these models do with the DR image classification. We employ publicly available Kaggle platform for training these models. The best classification accuracy is 95.68% and the results have demonstrated the better accuracy of CNNs and transfer learning on DR image classification. [ABSTRACT FROM AUTHOR]

Abstract Artificial neural networks are an integral part of emerging technologies, and ongoing research has shown that they can be applied to a variety of applications. This paper proposes a new cryptographic algorithm using chaotic neural networks, whose function is enhanced by construction with polynomials that exhibit chaos, namely, nonlinear Hermite and Chebyshev polynomials. These polynomials incorporate a memristor conductance, which is used as an activation function in the chaotic neural networks. Further, a function of the weights obtained from the chaotic neural networks, is used to generate the initial values that are used in the cryptographic process. The encryption algorithm employed here is inspired by the Lai–Massey block cipher with cubic and two-dimensional logistic maps, and the evaluation of these chaotic equations is performed using correlation values. The correlation values between the cipher and plain text are also examined to determine the undecipherability of the message to be sent on a public channel. [ABSTRACT FROM AUTHOR]

Abstract Nasopharyngeal carcinoma (NPC) is a multifaceted cancer tumor that makes its diagnosis challenging. NPC has a consistently diffusive enlargement that makes its resection exceptionally challenging. The pathological identification of NPC and comparing typical and anomalous tissues require a set of advanced strategies for the extraction of features. The use of medical images to diagnoses NPC tumor depends on tumor shape, region, and intensity. This paper proposes a novel approach for diagnosing NPC from endoscopic images. The approach includes a trainable segmentation for identifying NPC tissues, genetic algorithm for selecting the best features, and support vector machine for classifying NPC. The proposed approach is validated by comparing the number of classified NPC cases against the manual approach of ENT specialists. The approach shows a high precision of 95.15%, sensitivity of 94.80%, and specificity of 95.20%. Additionally, the optimized feature selection provides straightforward and efficient classification results. [ABSTRACT FROM AUTHOR]

Abstract: In this paper we have addressed the problem of finding a path through a maze of a given size. The traditional ways of finding a path through a maze employ recursive algorithms in which unwanted or non-paths are eliminated in a recursive manner. Neural networks with their parallel and distributed nature of processing seem to provide a natural solution to this problem. We present a biologically inspired solution using a two level hierarchical neural network for the mapping of the maze as also the generation of the path if it exists. For a maze of size S the amount of time it takes would be a function of S (O(S)) and a shortest path (if more than one path exists) could be found in around S cycles where each cycle involves all the neurons doing their processing in a parallel manner. The solution presented in this paper finds all valid paths and a simple technique for finding the shortest path amongst them is also given. The results are very encouraging and more applications of the network setup used in this report are currently being investigated. These include synthetic modeling of biological neural mechanisms, traversal of decision trees, modeling of associative neural networks (as in relating visual and auditory stimuli of a given phenomenon) and surgical micro-robot trajectory planning and execution. [Copyright &y& Elsevier]

Abstract Convolutional neural networks (CNNs) have been a focus area of machine learning in recent years, and they are widely used in vision and speech processing because of their superior performance. However, CNNs are usually resource-heavy to ensure higher accuracy, i.e., an accurate network with millions of parameters requires high performance computing devices. This prevents the use of CNNs in resource-limited hardware. In this paper, we propose a novel CNN simplification method to prune feature maps with relatively low discriminability magnitudes, which can produce a simplified CNN with reduced computational cost. Specifically, we define the critical points among the discriminability values of feature maps in each convolutional layer, and use these critical points to easily find the best pruning number of feature maps. Our experimental results show that in each convolutional layer of the VGG model, 15.6% to 59.7% of feature maps can be pruned without any loss of accuracy in classification tasks. [ABSTRACT FROM AUTHOR]

Highlights • A new approach for solving vehicle tracking problem in the traffic videos is proposed. • Feature extraction in frames of the video is performed using neuro optimization in varying time intervals. • Matching of features is done using neural networks in the consecutive frames that result into stable features. • The work is of great value to track features in the traffic surveillance systems. Abstract This paper presents an efficient feature-based tracking system to detect vehicles in a number of challenging conditions like lighting, occlusion, and darkness. A novel approach for vehicle tracking is proposed using an unsupervised feature matching technique. The system is fully functional under varying conditions because most of the salient vehicle features are tracked from matching of features in different objects. A feature-based vehicle tracking is proposed for a real-time traffic surveillance system. By analyzing the features of vehicles and their corresponding matched features, salient discriminative features of vehicles are calculated. The tracking of target vehicles is performed from the calculation of winner pixels in the consecutive frames using an unsupervised feature matching. To increase the accuracy of vehicle feature classification, orientation of feature descriptor of target vehicles tracked in the video frames is taken into consideration. Experimental results show that features classification rates of 96.4% and 92.7% for different vehicle sets can be achieved using the feature of aspect ratio. The proposed method is compared with recent feature-based method and Kalman filter-based method that results into better detection performance. The method can track the target vehicle under different situations like rotation, scaling, illumination and many others requiring less computation and providing better accuracy. Graphical abstract Image, graphical abstract [ABSTRACT FROM AUTHOR]

The paper proposes a novel adaptive control of the human musculoskeletal arm model in order to simulate reaching movements of the human upper extremity. The controller is based on a normalized radial basis function neural network, which takes the Actor-Critic structure. The normalized radial basis function neural network simultaneously approximates the policy function of the actor network and the value function of the critic network. An adaptive adjustment mechanism is dynamically established to realize the state space constructions. The approach could effectively overcome the curse of dimensionality caused by state space division and always keep the structure in an optimal state. The human arm model adopts a Hill-type model with two joints, six muscles. As a validation, numerical simulations are performed to achieve the reaching movements of the human upper extremity. The results show that the adaptive control can well instruct the human arm model to mimic the reaching movements. [ABSTRACT FROM AUTHOR]

Live virtual machine (VM) migration improves the performance of cloud data center in terms of energy efficiency, fault tolerance, and availability. The workload handled by cloud data center is dynamic in nature. This increases the resource requirement of either the migrated virtual machine or collocated virtual machine at any time leading to further migration. Inappropriately handled live VM migration imposes severe application performance degradation. In this paper, a combined forecasting technique to predict the resource requirement of any virtual machine is proposed. Based on the current and predicted resource utilization, live migration is performed by Combined Forecast Load-Aware technique. Experiments were carried out to evaluate the performance of the proposed technique on live VM migration. The outcomes indicate that the proposed approach has minimized the number of migrations, energy usage, and the message overhead when compared with the existing state-of-art technique. [ABSTRACT FROM AUTHOR]

Since clustering ensemble was proposed, it has rapidly attracted much attention. This paper makes an overview of recent research on clustering ensemble about generative mechanism, selective clustering ensemble, consensus function and application. Twelve clustering ensemble algorithms are described and compared to choose a basic one. The experiment shows that using k-means with different initializations as generative mechanism and average-linkage agglomerative clustering as consensus function is the best one. As ensemble size increases, the performance of clustering ensemble improves. The basic clustering ensemble algorithm with suitable ensemble size is compared with clustering algorithms and the experiment shows that clustering ensemble is better than clustering. The influence of diversity on clustering ensemble is instructive to selecting members. The experiment shows that selecting members in high quality and big diversity for low-dimensional data sets, and selecting members in high quality and median diversity for high-dimensional data sets are better than traditional clustering ensemble. [ABSTRACT FROM AUTHOR]

Recently, convolutional neural network (CNN) has achieved impressive results in object classification tasks. Through various modifications and careful design of the inner structure of CNN, its performance has already become human-competitive, according to recent reports. To some extent, the testing accuracy depends largely on the decision boundaries produced by classifiers, which classify different objects into specific feature spaces. Hence, the relationships among samples in the feature space are critical. However, the softmax loss function that is used in most CNN models, does not directly contain the relationship information. In this paper, we propose a novel loss function, named inter-class constraint loss function, that maximizes the distance between different classes. Together with softmax loss, we can obtain larger inter-class distances and smaller intra-class distances in CNN, thus significantly improving the accuracy in classification. We achieve substantial improvements for the SVHN, CIFAR-10 and CIFAR-100 datasets using our proposed loss function. [ABSTRACT FROM AUTHOR]

This paper proposes a human intuition-based concept of Case-Based Reasoning (CBR) for Static Security Assessment (SSA) of power systems. Prior to security assessment, a composite Security Index (SI) defined in terms of line overload and voltage violation, is computed for each training pattern to label them into either secure or insecure class. Out of all phasor measurements obtained using Phasor Measurement Units (PMUs), only the real and reactive power flows, which are identified as relevant features are given as inputs to the proposed CBR system. The proposed approach has been implemented on IEEE 14 bus, IEEE 30 bus and Indian 246 bus networks and the results have been compared with those obtained using Support Vector Machine (SVM) and Artificial Neural Network (ANN) based classifiers. The results indicate that the proposed approach is more efficient, reliable and outperforms the above said classifiers in terms of classification accuracy at a comparable computation cost. [ABSTRACT FROM AUTHOR]

Human emotion analysis is one of the challenging tasks in today's scenario. The success rate of human emotion recognition has high implication in practical applications such as Human Machine Interaction, anomaly detection, surveillance, etc. Artificial Neural Networks (ANN) is one of the highly favored computational intelligence techniques for human emotion recognition. However, the performance of traditional ANN is not satisfactory in case of applications such as human emotion analysis. This leads to the necessity of modified ANN with better performance than the conventional systems. In this paper, we propose Circular Back Propagation Neural Network (CBPN) and Deep Kohonen Neural Network (DKNN) to overcome drawbacks of the traditional neural networks regarding computational complexity and accuracy. Performance of the proposals is explored in classifying different emotions of humans using Electroencephalography (EEG) signals. It has been validated that the proposals have better performance than the related methods. [ABSTRACT FROM AUTHOR]

This paper examines an observer-based backstepping terminal sliding mode controller (BTSMC) for 3 degrees of freedom overhead transmission line de-icing robot manipulator (OTDIRM). The control law for tracking of the OTDIRM is formulated by the combination of BTSMC and neural network (NN) based approximation. For the precise trajectory tracking performance and enhanced disturbance rejection, NN-based adaptive observer backstepping terminal sliding mode control (NNAOBTSMC) is developed. To obviate local minima problem, the weights of both NN observer and NN approximator are adjusted off-line using particle swarm optimization. The radial basis function neural network-based observer is used to estimate tracking position and velocity vectors of the OTDIRM. The stability of the proposed control methods is verified with the Lyapunov stability theorem. Finally, the robustness of the proposed NNAOBTSMC is checked against input disturbances and uncertainties. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT) Technology has the potential to capture real-time health related parameters everywhere. Henceforth, in this paper, the Cloud Centric IoT (CCIoT) Technology is utilized to assess the health related attributes of a trainee during exercise sessions in a gymnasium. The proposed system have the capabilities to predict the probabilistic vulnerability to health parameters of a trainee during workouts. For this purpose, back-propagation based Artificial Neural Network (ANN) technique is used as a prediction model, layered into three stages, i.e. Monitoring, Learning, and Prediction. Also, the probabilistic vulnerability is represented in real-time using color-coded technique, depicting the health state of the trainee. The proposed system has been validated using an experiment in which five people were monitored for six days at different gymnasiums. Results are compared with different state-of-the-art techniques for determining the overall effectiveness of the proposed system. [ABSTRACT FROM AUTHOR]

Neuromorphic computing based on three-dimensional inetgraed circuits (3D-NCs) offers a novel hardware implementation of neuromorphic computing, and provides high device density, massively parallel signal processing capability, low power consumption, and direct analog signal processing capability. In this paper, by replacing conventional CPUs based on Von Neumann architecture with 3D-NCs, a novel neuromorphic computing based cloud robotics (NC-robotics) system is proposed, which is constructed by 1) cloud server center using 3D-NCs as computing units, 2) neuromorphic robotics based on neural network control technology. Besides the benefits of normal Cloud Robotics platform, this NC- Robotics system has more advantages on massive parallel-computing, analog signals processing, and lower power consumption. In order to implement this NC––Robotics system, a novel 3D-NCs architecture combining vertical RRAM structure is investigated and its concise equivalent circuit model is created, evaluated, and analyzed through SPICE simulations. [ABSTRACT FROM AUTHOR]

Relocation is one of the most common problems in Simultaneous Localization and Mapping (SLAM). This paper presents a novel relocation method, using unsupervised deep learning algorithm to extract the feature of Light Detection and Ranging (LiDAR) data, and narrows the scope of relocation by classifying these features to reduce the randomness of the relocation. Compared with the other methods which is based on matching the manual feature points, this method avoids some limitations of manual features. We modify the Particle Filter SLAM (PF-SLAM), and use our relocation method to replace the original method for experimentation. The experimental results demonstrate that this method can be relocation whit high success rate only use a small amount of computational resource in a short time. Training neural network will consume a lot of computing resources, we also propose a cloud computing framework to the implementation of this method for the mobile robot which computational resources are limited. [ABSTRACT FROM AUTHOR]

Many medical solutions exist in the vast medical logs, and how to use them to make precise recommendations for medical pathways has become a current research hotspot. This paper proposed an algorithmic framework that combines trace clustering, process discovery, and neural network techniques to discover models of medical pathways considering complications from medical logs. First, trace clustering divided the treatment pathways set in medical logs into multiple subsets with similar behaviours. Then multiple streamlined process models were mined by process discovery techniques. The reasonable medical pathways were extracted from each process model. A cluster of trained neural network models was used to determine the case characteristic labels and obtain the attention of events in the medical pathways. The output was finally integrated to generate a model of medical pathways considering complications. The results of the experiments using the Sepsis Cases dataset showed that the average simplicity of the generated process models was 0.695, the average accuracy of the neural network models was 93.3%, and the medical pathways model score was about 0.879. [ABSTRACT FROM AUTHOR]

Landslides are among the many devastating natural calamities that cause damage to life and property. Predicting landslides is an important task to enable preventive measures to be made on time. This paper presents an analysis of univariate time-series forecasting data using an auto regressive integrated moving average (ARIMA) model, a generalized autoregressive conditional heteroskedasticity (GARCH) model, and a dynamic neural network (DNN) model. These techniques rely on the objective of the forecasting, the type of forecasted data, and whether an automatic or manual approach is to be used for forecasting. Different techniques were analyzed on 15-meter landslide sensor data. The objective of this paper is to suggest a best method among well-known models for landslide forecasting. The demonstrated result shows that a dynamic neural network model is best in class for time-series landslide forecasting. Furthermore, upon objectively evaluating the three well-known techniques, the DNN model exhibited a minimum error rate of approximately 0.01 in comparison to other implemented techniques. [ABSTRACT FROM AUTHOR]

In this paper a novel hand-crafted local quadruple pattern (LQPAT) is proposed for facial image recognition and retrieval. Most of the existing hand-crafted descriptors encode only a limited number of pixels in the local neighborhood. Under unconstrained environment the performance of these descriptors tends to degrade drastically. Major problem in increasing the local neighborhood is that, it also increases the feature length of the descriptor. The proposed descriptor tries to overcome these problems by defining an efficient encoding structure with optimal feature length. The proposed descriptor encodes relations amongst the neighbors in quadruple space. Two micro patterns are computed from the local relationships to form the descriptor. The retrieval and recognition accuracies of the proposed descriptor has been compared with state of the art hand crafted descriptors on bench mark databases namely; Caltech-face, LFW, Color-FERET, and CASIA-face-v5. Result analysis shows that the proposed descriptor performs well under uncontrolled variations. [ABSTRACT FROM AUTHOR]

Target classification and recognition (TCR) of high resolution remote-sensing image is the important ability for earth observation system and unmanned autonomous system. It is difficult to improve the precision of TCR because of different imaging mechanism. In this paper, we propose a brain-inspired computing model for TCR using cognitive computing and deep learning. Accordingly, we have built an ensemble learning algorithm based on deep spiking convolutional neural network and hierarchical latent Dirichlet allocation. The hierarchical features were extracted from remote-sensing image. Then a TCR algorithm for small sample sizes and complex target was designed, which uses the incremental and reinforcement learning based on object-oriented and multi-scale data argumentation. Experimental results demonstrate that our algorithm has state-of-the-art performance on public data sets of optical remote-sensing image and synthetic aperture image. The model proposed can provide reference to explore an essential significance in brain-inspired intelligence, and has significant value in military and civil affairs. [ABSTRACT FROM AUTHOR]

In this paper, we present our work on analysis and classification of smiled vowels, chuckling (or shaking) vowels and laughter syllables. This work is part of a larger framework that aims at assessing the level of amusement in speech using the audio modality only. Indeed all of these three categories occur in amused speech and are considered to contribute in the expression of different levels of amusement. We first analyze these three amused speech components on the acoustic level. Then, we improve a classification system we previously developed. With a limited amount of data and features, we are able to obtain good classification results with different systems. Among the compared systems, the best one achieved 82.8% of accuracy, therefore outperforming chance. [ABSTRACT FROM AUTHOR]

This paper presents two random neural network (RNN) based context-aware decision making frameworks to improve adaptive modulation and coding (AMC) in long-term evolution (LTE) downlink systems. In the first framework, AMC is modelled as a traditional classification problem with the aim to maximize the probability of correct classification. The second framework seeks to optimize the throughput as opposed to simply maximizing the probability of the correct classification. To model the second framework, we developed a hybrid cognitive engine (CE) architecture by integrating an RNN based learning algorithm with genetic algorithm (GA) based reasoning. RNN inherent properties help CE to comply with the essential CE design requirement (i.e. concurrent long-term-learning, low computational complexity, and fast decision making). The performance of RNN is compared with artificial neural networks (ANN) and state-of-the-art effective exponential SINR mapping (EESM) algorithm. A comprehensive analysis of the proposed RNN based AMC scheme is presented by jointly incorporating the effect of different schedulers, feedback delays, and multi-antenna diversity on the throughput of an orthogonal frequency-division multiple access (OFDMA) system. The critical analysis of the first framework revealed that RNN based CE can achieve comparable results with faster adaptation, even in severe environment changes without the need of retraining compared to ANN. The analysis of the second approach demonstrated RNNs faster adaptation as compared to ANN and showed upto 253% gain in user throughput. RNN based CE efficiently exploited the channel quality information feedback delay to improve system throughput and helped cell-edge and cell-centre users to experience much better services in terms of achieved throughput as compared to EESM. [ABSTRACT FROM AUTHOR]

In this paper the design of a novel high-speed, low-cost autonomous line following robot that combines human expert knowledge and experiential data extracted through neural network training has been proposed and implemented. The proposed autonomous robot overcomes the disadvantages of earlier designs like oscillations in motion, exception handling, high-cost and high-energy consumption. The robot uses a novel square-topology infrared sensor matrix that enables it to anticipate a turn by sensing the track ahead. Three different strategies to control high-speed line following robots are presented: a neural network based control strategy, a learning vector quantization based strategy and a reinforcement learning based strategy. The final design based on a hybrid neural network based control strategy smoothly followed the benchmark track with just 4 neurons and used a low-cost 8-bit microcontroller to implement the control system. [ABSTRACT FROM AUTHOR]

The Ultraviolet (UV) Non-line-of-sight (NOLS) communications are becoming a research hotspot in recent years. For wireless UV networks, there are some well-known existing issues such as network connectivity, power control and channel assignment. In this paper, a Fast Power Control and Channel Assignment scheme (FPCCA) is proposed to reduce interference and save energy. Firstly, the FPCCA is compared with the traditional channel assignment which without power control. It is proved that the approach of FPCCA can save energy more than 10%. Then, the FPCCA is both applied in the UV and Wireless Mesh Networks (WMN). The performance of FPCCA with different angles and network traffic loads is simulated and analyzed. When the traffic load is fixed, the performance metrics of FPCCA in UV are better than that in WMN. Meanwhile, the FPCCA not only ensures the quickness and accuracy of channel assignment, but also saves energy and improves the throughput. [ABSTRACT FROM AUTHOR]

• One dimensional emotion recognition method is proposed to realize the dimensional emotion recognition by combing the attention mechanism and global second-order deep representations. • The residual attention network is used to extract emotionally relevant features to solve the negative influence of emotionally irrelevant features on the recognition performance. • The global second-order pooling network is integrated to learn the interdependence relationship between the long-distance features, increasing the nonlinear emotion representation learning ability of the trained model. • The experiments on the Affectnet dataset demonstrate the validity of the proposed model which is superior to some existing methods. For the emotion recognition tasks, the emotion-related features and the emotionally irrelevant ones are fully extracted by the deep neural network is particularly desirable. In this paper, a dimensional emotion recognition algorithm is proposed by combing the attention mechanism and global second-order feature representations. More specially, the residual attention network (RAN) is utilized to select the features related to the task, and then the global second-order pooling network is employed to model the correlation between these features to complete the modeling of the feature interdependence relationship, improving the deficiency of the RAN in this point so as to enhance the capacity of extracting the emotion-related features and the emotionally irrelevant ones. The experiments on the AffectNet dataset demonstrate the validity of the proposed model, which is comparable to and even superior to those from the support vector regression method, the convolutional neural network (CNN)-based method, and the recently-developed CNN-based variants. [Display omitted] [ABSTRACT FROM AUTHOR]

Capsule Network (CapsNet) complements the invariance properties of the convolutional neural network with equivariance through pose estimation. While CapsNet achieves a very decent performance with a shallow architecture, it suffers from heavily parameters learning in the primary capsule layer, and has a poor performance on large or complex datasets. To tackle these problems, this paper benefits CapsNet from two techniques of model compression: bottleneck residual block and automated gradual pruning. Specifically, this paper designs an end-to-end framework, denoted as Deft Capsule (DCaps). In order to reduce the number of need-to-learn parameters, this paper applies the bottleneck residual blocks to the primary capsule layer. Furthermore, this paper obtains automated gradual pruning techniques in dynamic routing to improve the performance of DCaps. Experiment results on four classic datasets demonstrate that DCaps learns fast and significantly outperforms CapsNet in image classification tasks. [ABSTRACT FROM AUTHOR]

All witnessed the terrible effects of the COVID-19 pandemic on the health and work lives of the population across the world. It is hard to diagnose all infected people in real time since the conventional medical diagnosis of COVID-19 patients takes a couple of days for accurate diagnosis results. In this paper, a novel learning framework is proposed for the early diagnosis of COVID-19 patients using hybrid deep fusion learning models. The proposed framework performs early classification of patients based on collected samples of chest X-ray images and Coswara cough (sound) samples of possibly infected people. The captured cough samples are pre-processed using speech signal processing techniques and Mel frequency cepstral coefficient features are extracted using deep convolutional neural networks. Finally, the proposed system fuses extracted features to provide 98.70% and 82.7% based on Chest-X ray images and cough (audio) samples for early diagnosis using the weighted sum-rule fusion method. [ABSTRACT FROM AUTHOR]

An intrusion detection system (IDS) is considered critical for detecting threats, intrusions, and unauthorized access. IDS monitors massive network traffic that includes irrelevant and extravagant features that profoundly impact the system's efficiency and slow down the classification process for accurate decisions. Its effectiveness is tested over the various techniques that comprise an enormous volume of data and heavy network traffic. Many approaches, such as machine learning algorithms, data mining, swarm intelligence, and artificial neural networks (ANN), have been implemented for adequate and improved IDSs. This paper recommends a novel feature selection method using a genetic algorithm (GA) that determines the optimal feature subsets from the NSL-KDD dataset. Further, hybrid classification has been performed using logistic regression (LR) and decision tree (DT) to achieve a better detection rate (DR) and accuracy (ACC). This research applied and compared several meta-heuristic algorithms' performance to optimize the selected optimal features. The experimental results show that the grey wolf optimization (GWO) algorithm gives the best accuracy of 99.44% and DR of 99.36% with the reduction of features (=20) out of (=41). The results of the proposed work are compared with the existing feature selection methods to verify improved performance. [Display omitted] • Novel feature selection method using a genetic algorithm (GA) that determines the optimal feature subsets. • Hybrid classification to achieve a better detection rate (DR) and accuracy (ACC). • Compared several meta-heuristic algorithms' performance to optimize the selected optimal features. • The grey wolf optimization (GWO) algorithm gives the best accuracy of 99.44% and DR of 99.36% with the reduction of features (=20) out of (=41). • Optimization for improvement of IDS. [ABSTRACT FROM AUTHOR]

Cross-view geo-localization is to localize the same geographic target in images from different perspectives, e.g., satellite-view and drone-view. The primary challenge faced by existing methods is the large visual appearance changes across views. Most previous work utilizes the deep neural network to obtain the discriminative representations and directly uses them to accomplish the geo-localization task. However, these approaches ignore that the redundancy retained in the extracted features negatively impacts the result. In this paper, we argue that the information bottleneck (IB) can retain the most relevant information while removing as much redundancy as possible. The variational self-distillation (VSD) strategy provides an accurate and analytical solution to estimate the mutual information. To this end, we propose to learn discriminative representations via variational self-distillation (dubbed LDRVSD). Extensive experiments are conducted on two widely-used datasets University-1652 and CVACT, showing the remarkable performance improvements obtained by our LDRVSD method compared with several state-of-the-art approaches. • Variational self-distillation is used for cross-view geo-localization. • Square-ring partition strategy is adopted to enrich discriminative clues. • Information bottleneck module is designed to discard redundancy information. [ABSTRACT FROM AUTHOR]

• Present a computer aided diagnosis model for skin cancer. • Propose an ODNN CADSCC model for skin cancer detection and classification. • Employ U-Net segmentation with Squeezenet feature extraction. • Introduce IWOA with deep neural network for skin cancer classification. • Validate the performance on benchmark ISIC 2019 dataset. Image-guided intervention is a medical procedure that leverages computerized systems to deliver virtual image overlays to help physicians in visualization and targeting the surgical site in an accurate manner. Computer Aided Diagnosis (CAD) models that use Deep Learning (DL) techniques are useful in achieving accurate skin cancer classification. In this background, the current research paper concentrates on the design of Optimal Deep Neural Network Driven Computer Aided Diagnosis Model for Skin Cancer Detection and Classification (ODNN CADSCC) model. The presented ODNN CADSCC model primarily applies Wiener Filtering (WF)-based pre-processing step followed by U-Net segmentation approach. In addition, SqueezeNet model is also exploited to generate a collection of feature vectors. Finally, Improved Whale Optimization Algorithm (IWOA) with DNN model is utilized for effectual skin cancer detection and classification. In this procedure, IWOA is applied to select the DNN parameters in a proficient manner. The comparative analysis results established the promising performance of the proposed ODNN CADSCC model over recent approaches with a maximum accuracy of 99.90%. [Display omitted] [ABSTRACT FROM AUTHOR]