Search

In this paper, we address the adaptive neural tracking control problem for a class of uncertain switched stochastic nonlinear pure-feedback systems with nonlower triangular form. The significant design difficulty is the completely unknown nonlinear functions with all state variables that can neither be directly estimated by radial basis function (RBF) neural networks (NNs) nor be eliminated by the traditional backstepping technique. To achieve the control objective of this paper, a common state-feedback controller for all subsystems is first systematically constructed by using the common coordinate transformation, the variable separation technique, and the universal approximation capability of RBF NNs. Then the stability analysis shows that the semi-global bounded in probability of the whole closed-loop switched system can be obtained and the desired tracking performance can also be insured under a class of switching signals with the average dwell time property. Finally, simulation results are given to demonstrate the effectiveness of the obtained control scheme. [ABSTRACT FROM AUTHOR]

Most of the existing blind omnidirectional image quality assessment (BOIQA) methods are based on data-driven approach where the end-to-end neural network or deep learning tools are mainly used for feature extraction. However, it usually lacks interpretability and is difficult to discover the perceptual mechanism behind. In this paper, from the perspective of perception modeling, we propose a novel multi-angle projection based BOIQA (MP-BOIQA) method. Considering the omnibearing and near eye display characteristics with head mounted display, multiple color cubemap projection images with respect to different viewpoints are grouped as the color omnidirectional distortion (COD) units so as to simulate the user’s viewing behavior in subjective quality assessment. In the designed multi-angle projection based feature extractor, tensor decomposition is implemented on each COD unit for dimensionality reduction, and piecewise exponential fitting is used to get the distribution of mean subtracted contrast normalized coefficients of the unit’s feature matrices in tensor domain. Finally, the extracted features are pooled with random forest. The experimental results on three omnidirectional image quality datasets show that the MP-BOIQA method can deliver highly competitive performance compared with some representative full-reference quality assessment methods, as well as some state-of-the-art BOIQA methods. [ABSTRACT FROM AUTHOR]

Positron Emission Tomography (PET) has become a preferred imaging modality for cancer diagnosis, radiotherapy planning, and treatment responses monitoring. Accurate and automatic tumor segmentation is the fundamental requirement for these clinical applications. Deep convolutional neural networks have become the state-of-the-art in PET tumor segmentation. The normalization process is one of the key components for accelerating network training and improving the performance of the network. However, existing normalization methods either introduce batch noise into the instance PET image by calculating statistics on batch level or introduce background noise into every single pixel by sharing the same learnable parameters spatially. In this paper, we proposed an attentive transformation (AT)-based normalization method for PET tumor segmentation. We exploit the distinguishability of breast tumor in PET images and dynamically generate dedicated and pixel-dependent learnable parameters in normalization via the transformation on a combination of channel-wise and spatial-wise attentive responses. The attentive learnable parameters allow to re-calibrate features pixel-by-pixel to focus on the high-uptake area while attenuating the background noise of PET images. Our experimental results on two real clinical datasets show that the AT-based normalization method improves breast tumor segmentation performance when compared with the existing normalization methods. [ABSTRACT FROM AUTHOR]

Convolutional Neural Networks (CNN) and Graph Neural Networks (GNN), such as Graph Attention Networks (GAT), are two classic neural network models, which are applied to the processing of grid data and graph data respectively. They have achieved outstanding performance in hyperspectral images (HSIs) classification field, which have attracted great interest. However, CNN has been facing the problem of small samples and GNN has to pay a huge computational cost, which restrict the performance of the two models. In this paper, we propose Weighted Feature Fusion of Convolutional Neural Network and Graph Attention Network (WFCG) for HSI classification, by using the characteristics of superpixel-based GAT and pixel-based CNN, which proved to be complementary. We first establish GAT with the help of superpixel-based encoder and decoder modules. Then we combined the attention mechanism to construct CNN. Finally, the features are weighted fusion with the characteristics of two neural network models. Rigorous experiments on three real-world HSI data sets show WFCG can fully explore the high-dimensional feature of HSI, and obtain competitive results compared to other state-of-the art methods. [ABSTRACT FROM AUTHOR]

A sum-product network (SPN) is a probabilistic model, based on a rooted acyclic directed graph, in which terminal nodes represent probability distributions and non-terminal nodes represent convex sums (weighted averages) and products of probability distributions. They are closely related to probabilistic graphical models, in particular to Bayesian networks with multiple context-specific independencies. Their main advantage is the possibility of building tractable models from data, i.e., models that can perform several inference tasks in time proportional to the number of edges in the graph. They are somewhat similar to neural networks and can address the same kinds of problems, such as image processing and natural language understanding. This paper offers a survey of SPNs, including their definition, the main algorithms for inference and learning from data, several applications, a brief review of software libraries, and a comparison with related models. [ABSTRACT FROM AUTHOR]

Despite encouraging results have been achieved in human pose estimation in recent years, it remains challenging problems. When the background is similar to the human body parts, and there are small persons with low-resolution in the image, the performance may degrade dramatically. This paper addresses problems in background-inference and small-person pose estimation. To achieve this, a novel pose estimation algorithm is proposed on the basis of person semantic segmentation deep neural network. Different from most previous methods with a single pose estimation model, we generate mixture models with pose estimation and semantic segmentation. We introduce novel generative adversarial model and auxiliary model to realize the semantic segmentation network, which can handle the confusion of the similar regions in the background. In addition, to address the problem of the scale differences between big and small persons’ keypoints, we add additional position and channel attention modules to the first two stages of OpenPose. We conduct extensive experiments on COCO and VOC datasets. And we compare the proposed method with the most popular state-of-the-art human pose estimation and semantic segmentation frameworks, including MultiPoseNet, Deterton2 and DeepLab V3. Our experimental results show that the proposed method is more accurate than the state-of-the-art algorithms and performs effectively in tackling the complex situations. [ABSTRACT FROM AUTHOR]

Swarm foraging is a common test case application for multi-robot systems. In this paper RepAtt algorithm is used for improving coordination of a robot swarm by selectively broadcasting repulsion and attraction signals. This is a chemotaxis-inspired search behaviour where robots use the temporal gradients of these signals to navigate towards more advantageous areas. Hardware experiments were used to model and validate realistic, noisy sound communication and vision system. We then show through extensive simulation studies that RepAtt significantly improves swarm foraging time and robot efficiency under realistic communication and vision models. Note to Practitioners—This research developed a swarm foraging algorithm that takes into consideration the vision and communication sensing noise levels faced by robots in real world applications. The algorithm, known as RepAtt, was developed with the aim of emphasizing algorithmic simplicity and limiting the hardware requirements for the robots in the swarm. In this paper, we have focused on the problem of deploying swarm robots to forage litter in an environment such as a park. The communication model of the robots was based on the physics of sound, while their vision system was modelled using experiments with deep neural networks based object detectors. The results show that the RepAtt algorithm is robust to different distributions of targets (or litter) in the search space, exhibits good swarm efficiency with changes in swarm population and is robust to noise in its communication and vision systems. Apart from the RepAtt algorithm, other contributions made by this research include modelling of robot vision system to aid extensive study of the impact of communication and vision noise on swarm coordination. This will be relevant for extensive testing and validation before deployment to swarm robots hardware. The sound communication used in this research limits the kinds of environment the robots can be deployed in. Echoes within an enclosed environment and bandwidth limitation for communication frequency and public disturbance due to sound emitted by the robots can all contribute to this limitation. Thus, this research can be improved by investing in the development of a communication technology with similar physics. Other areas of improvement include adopting better obstacle avoidance algorithms and implementing suitable manipulators for handling litter objects. The algorithm can be extended to make it applicable for solving other problems such as search and rescue operations where foraging targets could be disaster survivors; demining and hazardous waste cleanup, where targets are the mines or waste material; and planetary exploration, where targets could be interesting features of the planets are the targets searched for by the robots. [ABSTRACT FROM AUTHOR]

Over the past decade, the high penetration of renewable-based distributed generation (DG) units has witnessed a considerable rise in electrical networks. In this context, direct current (DC) microgrids based on DGs are being preferred due to having less complexity for the establishment and control. At the same time, they offer higher efficiency and reliability compared to their alternating current (AC) counterparts. This paper proposes a new model predictive control (MPC)-trained artificial neural network (ANN) control strategy being an ANN-MPC instead of conventional cascaded-proportional-integral (PI)-trained ANN control for dynamic damping of photovoltaic (PV)-battery-based grid-connected DC microgrids. Unlike traditional controllers, the proposed control approach more rapidly attains generation-load power balancing under variable climate input (meteorological sensor data) and output (load demand), hence achieving quick DC-bus voltage damping. The proposed ANN-MPC scheme is examined under different operating conditions, and the results are compared with the ANN-based conventional PI controller. The results show the proposed control strategy's efficacy to lessen the instability issues and achieve effective attenuation of oscillations in DC microgrids. [ABSTRACT FROM AUTHOR]

In this paper, we propose a binarized neural network learning method (BiDet) for efficient object detection. Conventional network binarization methods directly quantize the weights and activations in one-stage or two-stage detectors with constrained representational capacity, so that the information redundancy in the networks causes numerous false positives and degrades the performance significantly. On the contrary, our BiDet fully utilizes the representational capacity of the binary neural networks by redundancy removal, through which the detection precision is enhanced with alleviated false positives. Specifically, we generalize the information bottleneck (IB) principle to object detection, where the amount of information in the high-level feature maps is constrained and the mutual information between the feature maps and object detection is maximized. Meanwhile, we learn sparse object priors so that the posteriors are concentrated on informative detection prediction with false positive elimination. Since BiDet employs a fixed IB trade-off to balance the total and relative information contained in the high-level feature maps, the information compression leads to ineffective utilization of the network capacity or insufficient redundancy removal for input in different complexity. To address this, we further present binary neural networks with automatic information compression (AutoBiDet) to automatically adjust the IB trade-off for each input according to the complexity. Moreover, we further propose the class-aware sparse object priors by assigning different sparsity to objects in various classes, so that the false positives are alleviated more effectively without recall decrease. Extensive experiments on the PASCAL VOC and COCO datasets show that our BiDet and AutoBiDet outperform the state-of-the-art binarized object detectors by a sizable margin. [ABSTRACT FROM AUTHOR]

This paper presents an innovative method to adaptively optimize traffic signal plans based on the estimation of traffic situation achieved from the information of various penetration rates of Connected Vehicles (CVs). The network-wide signal control problem is formulated as a linear optimization problem. Moreover, we develop a Kalman filter (KF) and Neural Network (NN) algorithms to predict and update the traffic situation under mixed non-connected and connected vehicles environment. To capture the dynamic of the traffic flow, we employ the cell transmission model synched with the Vissim traffic simulator. The methodology is tested using a challenging network of six intersections. We test our model for various Penetration Rates (PR) of the CV to provide a comparative analysis. The performance of the method is also compared with a conventional actuated-coordinated traffic signal plan. The results show that with a bare minimum PR (say more than 30%), our proposed methodology outperforms the actuated traffic signal plan. (note that the minimum PR is subject to further ongoing research in the literature, to the extent that lower PRs might be plausible). Though a 100% PR is highly desirable, our method can fetch the maximum benefit just by 60% PR. [ABSTRACT FROM AUTHOR]

Nowadays, Neural network (NN) has been proved to be an effective solution for nonlinear equalization in short reach optical systems. However, recent research has mainly focused on implementing more powerful NNs for equalization, while ignoring their adaptability to equalization tasks. In this paper, we propose Multi-Symbol Output (MSO)-Neural Networks (NN) for nonlinear equalization in high-speed short reach optical interconnects. The results show that the proposed MSO design works well on Deep Neural Networks (DNN), Long Short-Term Memory neural networks (LSTM) and Gate Recurrent Unit (GRU), which are the recent NN-based equalization structures. By increasing output symbols of the NNs, the number of slide windows in equalization can be sharply reduced, and so the complexity is reduced. By the same time, more information is brought to the MSO-NNs in back-propagations, therefore performance gain achieved. A 212-Gb/s 1-km Pulse Amplitude Modulation (PAM)-4 optical link is experimentally demonstrated as the target system, and the proposed MSO-NN equalizers are used to compare with traditional equalization algorithms including Volterra Nonlinear Equalizer (VNE) and single-symbol output NNs. Experimental results show that MSO design could help reduce the complexity of NN required for nonlinear equalization in the target system by around 2/3, and the proposed MSO-LSTM performs much better than VNE and 1 dB better than SSO-LSTM at the same time. Based on the proposed MSO-LSTM, transmission with BER under HD-FEC over 1 km NZDSF is achieved with a ROP at -2 dBm. Our work is well expandable and the proposed MSO design can be extended to other NN-based equalizers, which can help reduce complexity and learn more info from the training data and gain performance. The proposed MSO-NNs further enhance the performance and reduces the complexity of NN equalizers, provides assurance for future real-time high-speed short-range optical systems, and brings new ideas to NN-based equalizer design. [ABSTRACT FROM AUTHOR]

The performance of train communication network (TCN) is one of the most important factors to ensure the safety of train operation, and there is a lack of research for the TCN performance currently. To master the state of TCN, a combined method to evaluate and predict the TCN performance is proposed in the paper. First, we select appropriate evaluation indexes and determine their comprehensive weights. Then, the evaluation model of TCN performance is established based on a multidimensional normal cloud model combined with weights. According to the evaluation results, the artificial neural network (ANN) is used to predict the TCN performance to realize the real-time evaluation. Finally, the effectiveness of the evaluation and prediction model of TCN performance is verified by the simulation result, which is obtained from the simulation platform of TCN. The results show that service quality for TCN has been effectively improved, which provides a theoretical reference for evaluation and prediction of TCN performance. [ABSTRACT FROM AUTHOR]

Short-term load forecasts are needed for the efficient management of power systems. Although weather-based modeling is common, univariate models can be useful when the lead time of interest is less than one day. A class of univariate methods that has performed well with intraday data is exponential smoothing. This paper considers five recently developed exponentially weighted methods that have not previously been used for load forecasting. These methods include several exponential smoothing formulations, as well as methods using discount weighted regression, cubic splines, and singular value decomposition (SVD). In addition, this paper presents a new SVD-based exponential smoothing formulation. Using British and French half-hourly load data, these methods are compared for point forecasting up to one day ahead. Although the new SVD-based approach showed some potential, the best performing method was a previously developed exponential smoothing method. A second empirical study showed the better of the univariate methods outperforming a weather-based method up to about five hours ahead, with a combination of these methods producing the best results overall. [ABSTRACT FROM PUBLISHER]

A general deep learning (DL) mechanism for a multiple hidden layer feed-forward neural network contains two parts, i.e., 1) an unsupervised greedy layer-wise training and 2) a supervised fine-tuning which is usually an iterative process. Although this mechanism has been demonstrated in many fields to be able to significantly improve the generalization of neural network, there is no clear evidence to show which one of the two parts plays the essential role for the generalization improvement, resulting in an argument within the DL community. Focusing on this argument, this paper proposes a new DL approach to train multilayer feed-forward neural networks. This approach uses restricted Boltzmann machine (RBM) as the layer-wise training and uses the generalized inverse of a matrix as the supervised fine-tuning. Different from the general deep training mechanism like back-propagation (BP), the proposed approach does not need to iteratively tune the weights, and therefore, has many advantages such as quick training, better generalization, and high understandability, etc. Experimentally, the proposed approach demonstrates an excellent performance in comparison with BP-based DL and the traditional training method for multilayer random weight neural networks. To a great extent, this paper demonstrates that the supervised part plays a more important role than the unsupervised part in DL, which provides some new viewpoints for exploring the essence of DL. [ABSTRACT FROM AUTHOR]

Recently, researchers have shown the powerful ability of deep methods with multilayers to extract high-level features and to obtain better performance for hyperspectral image classification. However, a common problem of traditional deep models is that the learned deep models might be suboptimal because of the limited number of training samples, especially for the image with large intraclass variance and low interclass variance. In this paper, novel convolutional neural networks (CNNs) with multiscale convolution (MS-CNNs) are proposed to address this problem by extracting deep multiscale features from the hyperspectral image. Moreover, deep metrics usually accompany with MS-CNNs to improve the representational ability for the hyperspectral image. However, the usual metric learning would make the metric parameters in the learned model tend to behave similarly. This similarity leads to obvious model’s redundancy and, thus, shows negative effects on the description ability of the deep metrics. Traditionally, determinantal point process (DPP) priors, which encourage the learned factors to repulse from one another, can be imposed over these factors to diversify them. Taking advantage of both the MS-CNNs and DPP-based diversity-promoting deep metrics, this paper develops a CNN with multiscale convolution and diversified metric to obtain discriminative features for hyperspectral image classification. Experiments are conducted over four real-world hyperspectral image data sets to show the effectiveness and applicability of the proposed method. Experimental results show that our method is better than original deep models and can produce comparable or even better classification performance in different hyperspectral image data sets with respect to spectral and spectral–spatial features. [ABSTRACT FROM AUTHOR]

This paper investigates a deep reinforcement learning (DRL)-based MAC protocol for heterogeneous wireless networking, referred to as a Deep-reinforcement Learning Multiple Access (DLMA). Specifically, we consider the scenario of a number of networks operating different MAC protocols trying to access the time slots of a common wireless medium. A key challenge in our problem formulation is that we assume our DLMA network does not know the operating principles of the MACs of the other networks—i.e., DLMA does not know how the other MACs make decisions on when to transmit and when not to. The goal of DLMA is to be able to learn an optimal channel access strategy to achieve a certain pre-specified global objective. Possible objectives include maximizing the sum throughput and maximizing $\alpha $ -fairness among all networks. The underpinning learning process of DLMA is based on DRL. With proper definitions of the state space, action space, and rewards in DRL, we show that DLMA can easily maximize the sum throughput by judiciously selecting certain time slots to transmit. Maximizing general $\alpha $ -fairness, however, is beyond the means of the conventional reinforcement learning (RL) framework. We put forth a new multi-dimensional RL framework that enables DLMA to maximize general $\alpha $ -fairness. Our extensive simulation results show that DLMA can maximize sum throughput or achieve proportional fairness (two special classes of $\alpha $ -fairness) when coexisting with TDMA and ALOHA MAC protocols without knowing they are TDMA or ALOHA. Importantly, we show the merit of incorporating the use of neural networks into the RL framework (i.e., why DRL and not just traditional RL): specifically, the use of DRL allows DLMA (i) to learn the optimal strategy with much faster speed and (ii) to be more robust in that it can still learn a near-optimal strategy even when the parameters in the RL framework are not optimally set. [ABSTRACT FROM AUTHOR]

A new equivalent dipole model hybrid with artificial neural network (ANN) is proposed in this paper for electromagnetic interference (EMI) estimation. Equivalent dipole method, based on the free-space Green’s function, is usually used to model unknown EMI sources on printed circuit boards. For high-speed and dense circuits, there may be multi-reflection and/or diffraction between the EMI source and its nearby components. Traditional dipole model usually omits such effects and leads to an inaccurate result in some cases. In our proposed method, the Green’s function of dipole is taken as input and the radiated EMI field is taken as the output of ANN. We use the powerful mapping ability of ANN to modify the matrix–vector multiplication between free-space Green’s function and dipole moments in the traditional dipole model, so that a new mapping between equivalent dipoles and their radiated fields is established. The near field of the EMI source is obtained by planar scanning, and is used for ANN training. After training, the ANN is used to predict the EMI field at the region of interest. Both numerical example and measurement example are given to show the effectiveness of the proposed ANN method. This paper provides a novel source reconstruction solution for the EMI problems. [ABSTRACT FROM AUTHOR]

High-voltage discharge-based nonthermal plasma (NTP) treatment for diesel exhaust is a laboratorial proven efficient technique. A prophecy of the treatment results based on the knowledge of its parameters would be a step forward toward bringing it into real-time applications of pollution control. In this paper, artificial neural networks (ANNs) and fuzzy logic model (FLM) have been used to model the NO $_{ X }$ (sum of NO and NO2) concentrations as a function of parameters of the NTP process. A data set of 4032 input–output pairs has been collected by conducting experiments, in which 70% of the data are used for the training of the models derived. The performances of all the considered models have been evaluated by testing them for the remaining 30% of the data, which is novel for the models. Furthermore, a comparison of the models has been made based on the root-mean-square error (RMSE) and mean relative error (MRE), where the FLM has been found to be the better compared to the ANN-based models, i.e., ANN, multilayer perceptrons (MLP), and functional link ANN (FLANN). The RMSE of FLM is 2.53 ppm for a test data of 1210 sets. It can be said from these results that the NO $_{ X }$ concentrations can be predicted using FLM with a good accuracy. [ABSTRACT FROM AUTHOR]

This paper presents a novel approach on motor current signature analysis (MCSA) for broken bar fault detection of induction motors (IMs), using as input the current signal measured from one of the three motor phases. Independent component analysis (ICA) is used over the Fourier-domain spectral signals obtained from the input and its autocorrelation function. The standard deviation of spectral components within a region of interest (ROI) of an ICA signal output was found to exhibit substantial differences between damaged and healthy motors. Separation of the ROI in one, two, and three sectors leads to an improved extraction of feature vectors, which are further fed into a neural network for classification purposes. The assessment of the proposed method is carried out through several experiments using two damage levels (broken bar and half broken bar) and two load motor conditions (50% and 75%), with a classification accuracy ranging from 90% to 99%. The contribution of this paper lies in a new technique of signal processing for ICA-based feature extraction in a 3-D feature space for IM fault diagnosis. [ABSTRACT FROM AUTHOR]

This paper proposes energy-efficient approximate multipliers based on the Mitchell's log multiplication, optimized for performing inferences on convolutional neural networks (CNN). Various design techniques are applied to the log multiplier, including a fully-parallel LOD, efficient shift amount calculation, and exact zero computation. Additionally, the truncation of the operands is studied to create the customizable log multiplier that further reduces energy consumption. The paper also proposes using the one's complements to handle negative numbers, as an approximation of the two's complements that had been used in the prior works. The viability of the proposed designs is supported by the detailed formal analysis as well as the experimental results on CNNs. The experiments also provide insights into the effect of approximate multiplication in CNNs, identifying the importance of minimizing the range of error.The proposed customizable design at $w$w = 8 saves up to 88 percent energy compared to the exact fixed-point multiplier at 32 bits with just a performance degradation of 0.2 percent for the ImageNet ILSVRC2012 dataset. [ABSTRACT FROM AUTHOR]

This paper is devoted to solving a full-wave inverse scattering problem (ISP), which is aimed at retrieving permittivities of dielectric scatterers from the knowledge of measured scattering data. ISPs are highly nonlinear due to multiple scattering, and iterative algorithms with regularizations are often used to solve such problems. However, they are associated with heavy computational cost, and consequently, they are often time-consuming. This paper proposes the convolutional neural network (CNN) technique to solve full-wave ISPs. We introduce and compare three training schemes based on U-Net CNN, including direct inversion, backpropagation, and dominant current schemes (DCS). Several representative tests are carried out, including both synthetic and experimental data, to evaluate the performances of the proposed methods. It is demonstrated that the proposed DCS outperforms the other two schemes in terms of accuracy and is able to solve typical ISPs quickly within 1 s. The proposed deep-learning inversion scheme is promising in providing quantitative images in real time. [ABSTRACT FROM AUTHOR]

In this paper, we present a two-speed, radix-4, serial-parallel multiplier for accelerating applications such as digital filters, artificial neural networks, and other machine learning algorithms. Our multiplier is a variant of the serial–parallel (SP) modified radix-4 Booth multiplier that adds only the nonzero Booth encodings and skips over the zero operations, making the latency dependent on the multiplier value. Two subcircuits with different critical paths are utilized so that throughput and latency are improved for a subset of multiplier values. The multiplier is evaluated on an Intel Cyclone V field-programmable gate array against standard parallel–parallel and SP multipliers across four different process–voltage–temperature corners. We show that for bit widths of 32 and 64, our optimizations can result in a $1.42\times $ – $3.36\times $ improvement over the standard parallel Booth multiplier in terms of area–time depending on the input set. [ABSTRACT FROM AUTHOR]

Over the last few decades, pattern classification has become one of the most important fields of artificial intelligence because it constitutes an essential component in many different real-world applications. Artificial neural networks and fuzzy logic are two most widely used models in pattern classification. To build an efficient and powerful model, researchers have introduced hybrid models that combine both fuzzy logic and artificial neural networks. Among the hybrid models, the fuzzy min–max (FMM) neural network has been proven to be a premier model for undertaking pattern classification problems. While FMM is useful in terms of its capability of online learning, it suffers from several limitations in the learning procedure. Therefore, over the past years, researchers have proposed numerous improvements to overcome the limitations of the original FMM model. This paper carries out a comprehensive survey of the developments conducted on the FMM model for pattern classification. In order to assist recent researchers in selecting the most suitable FMM variant and to provide proper guidance for future developments, this study divides the variants of FMM into two main board categories, namely FMM variants with and without contraction. This division facilitates understanding of the developments conducted by researchers on the original FMM neural network, as well as provides the scope to identify the limitations that still exist in the FMM models. This paper also summarizes the use of FMM and its variants in solving different benchmark and real-world problems. Finally, the possible future trends are highlighted. [ABSTRACT FROM AUTHOR]

In this paper, we developed a deep convolutional neural network (CNN) for the classification of malignant and benign masses in digital breast tomosynthesis (DBT) using a multi-stage transfer learning approach that utilized data from similar auxiliary domains for intermediate-stage fine-tuning. Breast imaging data from DBT, digitized screen-film mammography, and digital mammography totaling 4039 unique regions of interest (1797 malignant and 2242 benign) were collected. Using cross validation, we selected the best transfer network from six transfer networks by varying the level up to which the convolutional layers were frozen. In a single-stage transfer learning approach, knowledge from CNN trained on the ImageNet data was fine-tuned directly with the DBT data. In a multi-stage transfer learning approach, knowledge learned from ImageNet was first fine-tuned with the mammography data and then fine-tuned with the DBT data. Two transfer networks were compared for the second-stage transfer learning by freezing most of the CNN structures versus freezing only the first convolutional layer. We studied the dependence of the classification performance on training sample size for various transfer learning and fine-tuning schemes by varying the training data from 1% to 100% of the available sets. The area under the receiver operating characteristic curve (AUC) was used as a performance measure. The view-based AUC on the test set for single-stage transfer learning was 0.85 ± 0.05 and improved significantly (${p} < 0.05$) to 0.91 ± 0.03 for multi-stage learning. This paper demonstrated that, when the training sample size from the target domain is limited, an additional stage of transfer learning using data from a similar auxiliary domain is advantageous. [ABSTRACT FROM AUTHOR]

The abnormal rise in contact temperature in switchgear may lead to overheating or even explosion. It is necessary to monitor the rise in real contact temperature to ensure the safety and stability of switchgear. The accurate measurement of real contact temperature is still a problem, and the continuous state of switchgear cannot be assessed accurately by the temperature monitor. In this paper, a particular installation scheme of temperature sensor is put forward and a calculation algorithm for temperature rise of tulip contacts by the measured temperature of bus bar in real time based on artificial neural network (ANN) is established. An electromagnetic-thermal-fluid coupled simulation model for 10-kV/4000-A switchgears was built, and the thermal characteristic was studied taking the influential factors such as load current, ambient temperature, and contact resistance into consideration. The effectiveness of simulation model is discussed by the comparison to temperature rise experiment. Then, an ANN with three layers of structure is established to calculate the temperature of tulip contacts. The sensitivity and accuracy of network are calculated and verified. The results of this paper may provide a novel method for the monitor of switchgear temperature and assess continuous condition of high-current switchgear. [ABSTRACT FROM AUTHOR]

In this paper, we analyze the exponential stability, passivity, and $\boldsymbol {(\mathfrak {Q},\mathfrak {S},\mathfrak {R})}$ - $\boldsymbol {\gamma }$ -dissipativity of generalized neural networks (GNNs) including mixed time-varying delays in state vectors. Novel exponential stability, passivity, and $\boldsymbol {(\mathfrak {Q},\mathfrak {S},\mathfrak {R})}$ - $\boldsymbol {\gamma }$ -dissipativity criteria are developed in the form of linear matrix inequalities for continuous-time GNNs by constructing an appropriate Lyapunov-Krasovskii functional (LKF) and applying a new weighted integral inequality for handling integral terms in the time derivative of the established LKF for both single and double integrals. Some special cases are also discussed. The superiority of employing the method presented in this paper over some existing methods is verified by numerical examples. [ABSTRACT FROM AUTHOR]

The trend of pushing inference from cloud to edge due to concerns of latency, bandwidth, and privacy has created demand for energy-efficient neural network hardware. This paper presents a mixed-signal binary convolutional neural network (CNN) processor for always-on inference applications that achieves 3.8 $\mu \text{J}$ /classification at 86% accuracy on the CIFAR-10 image classification data set. The goal of this paper is to establish the minimum-energy point for the representative CIFAR-10 inference task, using the available design tradeoffs. The BinaryNet algorithm for training neural networks with weights and activations constrained to +1 and −1 drastically simplifies multiplications to XNOR and allows integrating all memory on-chip. A weight-stationary, data-parallel architecture with input reuse amortizes memory access across many computations, leaving wide vector summation as the remaining energy bottleneck. This design features an energy-efficient switched-capacitor (SC) neuron that addresses this challenge, employing a 1024-bit thermometer-coded capacitive digital-to-analog converter (CDAC) section for summing pointwise products of CNN filter weights and activations and a 9-bit binary-weighted section for adding the filter bias. The design occupies 6 mm2 in 28-nm CMOS, contains 328 kB of on-chip SRAM, operates at 237 frames/s (FPS), and consumes 0.9 mW from 0.6 V/0.8 V supplies. The corresponding energy per classification (3.8 $\mu \text{J}$) amounts to a 40 $\times $ improvement over the previous low-energy benchmark on CIFAR-10, achieved in part by sacrificing some programmability. The SC neuron array is 12.9 $\times $ more energy efficient than a synthesized digital implementation, which amounts to a 4 $\times $ advantage in system-level energy per classification. [ABSTRACT FROM AUTHOR]

There has been a resurgence of research in the design of deep architecture models and learning algorithms, i.e., methods that rely on the extraction of a multilayer representation of the data. Often referred to as deep learning, this topic of research has been building on and contributing to many different research topics, such as neural networks, graphical models, feature learning, unsupervised learning, optimization, pattern recognition, and signal processing. Deep learning is also motivated and inspired by neuroscience and has had a tremendous impact on various applications such as computer vision, speech recognition, and natural language processing. The clearly multidisciplinary nature of deep learning led to a call for papers for a special issue dedicated to learning deep architectures. [ABSTRACT FROM PUBLISHER]

Due to additive operation’s dominated computation and simplified network in binary convolutional neural network (BCNN), it is promising for Internet of Things scenarios which demand ultralow power consumption. By means of fully exploiting the in-memory computing advantages and low current consumption design using multilevel cell (MLC) spin-toque transfer magnetic random access memory (STT-MRAM), this paper proposes an MLC-STT-computing in-memory-based computing in-memory architecture to achieve convolutional operation for BCNN to further reduce the power consumption. Simulation results show that compared with the resistive random access memory (RRAM)- and spin orbit torque-STT-MRAM-based counterparts, the architecture proposed in this paper reduces power consumption by ~ $35{\times}$ and 59% in Modified National Institute of Standards and Technology data set, respectively. [ABSTRACT FROM AUTHOR]

This paper presents a novel nonintrusive protection scheme for fault classification of power transmission networks in a wide-area measurement system using fault information for decision making. The protection scheme is a noncommunication without global positioning system as it depends completely on locally measured currents for the nonintrusive fault monitoring (NIFM) using the power-spectrum-based hyperbolic S-transform (PS-HST). In this work, the HST is used to extract the high-frequency components of the current signals generated by an electric fault. To effectively select the HST coefficients (HSTCs) representing fault transient signals with increasing performance, a power spectrum of the HSTCs in different scales calculated by Parseval's theorem is proposed in this paper. Finally, back-propagation artificial neural networks and PS-HST are used to identify fault classes in power transmission networks. The proposed method is tested for different breaker on/off conditions by simulations using electromagnetic transients program software. The results obtained have proved that the proposed method is promising and demonstrate a high success rates and reliability for considering different fault resistances and inception times in NIFM applications. [ABSTRACT FROM AUTHOR]

This paper presents an adaptive neural control approach for nonstrict-feedback nonlinear systems in presence of unmodeled dynamics, unknown control directions and input dead-zone nonlinearity. To handle the difficulty due to uncertain control directions, Nussbaum gain functions are applied. Based on the structural characteristic of radial basis function neural networks, a backstepping-based adaptive neural control algorithm is developed. The main contributions of this paper lie in the fact that a backstepping-based neural control algorithm is developed for nonstrict-feedback nonlinear systems with unmodeled dynamics, unknown control directions and actuator dead-zone, and the total number of adaptive laws is not greater than the order of control system. As a beneficial result, the controller is much easier to be implemented in practice with less computational burden. A simulation example is given to reveal the viability of the presented approach. It is demonstrated by both theoretical analysis and simulation study that the presented control strategy ensures the semiglobally uniform ultimate boundedness of all closed-loop system signals. [ABSTRACT FROM AUTHOR]

This paper investigates the singular perturbation (SP) theory-based composite learning control of a flexible-link manipulator using neural networks (NNs) and disturbance observer (DOB). For the dynamics, the system states are separated into fast and slow variables in terms of time scale. For the multi-input–multi-output slow dynamics, the intelligent control is designed where NNs are used for system uncertainty approximation and the DOB is used for compound disturbance estimation. The main contribution is that a novel controller using NN and DOB is constructed to deal with unknown dynamics and time-varying disturbances while the composite learning algorithm is proposed with prediction error. For the fast dynamics, sliding mode control is employed. The boundedness of the tracking error is proved via Lyapunov approach. The simulation results show that the DOB-based composite neural control can greatly improve the tracking precision. [ABSTRACT FROM AUTHOR]

This paper investigates the gene circuit control design problem of time-delayed genetic regulatory networks. In the genetic regulatory networks, the time delays are unknown constants, and the genetic regulatory is not conventional SUM regulatory logic and can be modeled to be an unknown nonlinear function of the time-delayed states of the other genes in a cell. By Lyapunov stability, a novel adaptive gene circuit control design approach is proposed for the genetic regulatory networks, where the unknown time delays are estimated online by adaptive algorithms and the unknown regulatory functions are approximated by neural networks. The design approach in this paper is delay-dependent and has less conservatism than the delay-independent approach. From theoretical analysis, the closed-loop system is asymptotically stable and all the signals in the system converge to an adjustable neighborhood of the origin. Finally, a numerical example is given to show the effectiveness of the new design approach. [ABSTRACT FROM AUTHOR]

This paper proposes a novel facial landmark detection (FLD) algorithm for use in real driving situations. The proposed algorithm is based on an ensemble of local weighted random forest regressor (WRFR) with random sampling consensus (RANSAC) and explicit global shape models, and considers the dynamic and irregular characteristics of driving. In this paper, to estimate the offset distance from a landmark and a reference point, we first detect the nose region as a reference point. Next, we propose a local WRFR to maintain the generality with a small number of regression trees. With the WRFR, we adopt RANSAC instead of the averaging or median of offsets to handle the problem of sensitivity to outlier offsets, and we estimate the accurate 2D offset vector. To identify the erroneous positions of local landmarks and rearrange the overall landmark layout, we adopt the global face models based on the spatial relation between landmarks. Using the unified framework of the proposed FLD, our proposed algorithm is robust to large head poses and partial occlusions caused by a driver’s hair or sunglasses. For a benchmark data set considering real driving situations, we construct a data set called a face alignment data set used in driving (FADID) using a near-infrared camera for FLD under real driving situations. We apply the proposed algorithm to various driving sequences in FADID, and the results show that its FLD performance is better than that of other state-of-the-art methods, while the computational speed is high for real-time applications such as driver-state monitoring systems. [ABSTRACT FROM AUTHOR]

With the development of video surveillance in public safety field, there is an increasing research on person re-identification (re-id). In this paper, we address the image-to-video person re-id, in which the probe is an image and the gallery is consists of videos captured by nonoverlapping cameras. Compared with image, video sequence contains more temporal information that can be explored to improve the performance of re-identification system. However, it is challenging to model temporal information in the matching process of image-to-video person re-id. In this paper, we proposed a novel temporally memorized similarity learning neural network for this problem. In specific, the proposed network mainly consisted of two parts, including feature representation sub-network and similarity sub-network. In the first part, we adopted a convolutional neural network (CNN) to extract features from the input image. Given a video sequence of a person, features were first extracted from each its frame by using CNN and further forward to a long shot term memory (LSTM) network to encode the temporal information of video sequence. The outputs of LSTM were concatenated together as the feature vector of video sequences. Finally, the feature vectors of probe image and the video sequence were further forward to the similarity sub-network for distance metric learning. In the proposed framework, the feature representation and the similarity metric learning can be learned and optimized simultaneously. We evaluated the proposed framework on three public person re-id data sets, and the experimental results showed that the proposed approach is effective for the image-to-video person re-id. [ABSTRACT FROM AUTHOR]

Modeling images/videos with covariance matrices has attracted increasing attentions in various vision tasks, especially in visual classification. For covariances-based visual classification, measuring the distances between covariances is one of the key issues and has been studied for decades. Since the space of covariances is a Riemannian manifold, the geometrical structure of covariances should be favorably considered when designing distance metrics. Although this problem has been widely studied, designing an effective and efficient metric between high-dimensional covariances (HDCOV) for scalable classification is still an open problem. In this paper, we present an information geometry-based distance (IGBD) to tackle this challenge from the perspective of information geometry. Our idea is based on the fact that each covariance can be viewed as a zero-mean Gaussian distribution, and thus the distances between covariances are measured by those between the corresponding Gaussian distributions. The core of our method is to project each distribution, in the form of a set of random samples, to a vector on the tangent space of a common, known distribution on the statistical manifold, based on Fisher information metric and maximum likelihood method. On the tangent space, the Euclidean norm can be used to measure the distances between those sets of projection vectors (or equivalently distributions). The proposed IGBD for HDCOV is computationally efficient and easily combined with a linear support vector machine, suitable for scalable visual classification. The experiments are conducted on various kinds and sizes of benchmarks, and results show the proposed method is efficient and the combination of HDCOV can achieve very competitive performance. [ABSTRACT FROM AUTHOR]

This paper deals with the issue of adaptive observer-based stabilization for a more general class of uncertain switched stochastic nonlower triangular nonlinear systems with average dwell-time (ADT) switching. The completely unknown nonlinear functions possessing the entire states are handled using the neural networks universal approximation property and the variable separation technique. The main novelty of this paper is that it presents a universal formula for constructing an adaptive output-feedback controller of the considered system with only one adaptive parameter by designing a switched state observer. The theoretical analysis shows that the Lyapunov stability of the overall closed-loop system is ensured in the presence of ADT switching. Finally, the effectiveness and practicability of the obtained design scheme are illustrated by two simulation examples. [ABSTRACT FROM AUTHOR]

We recently proposed a convolutional neural network (CNN) for remote sensing image pansharpening obtaining a significant performance gain over the state of the art. In this paper, we explore a number of architectural and training variations to this baseline, achieving further performance gains with a lightweight network that trains very fast. Leveraging on this latter property, we propose a target-adaptive usage modality that ensures a very good performance also in the presence of a mismatch with respect to the training set and even across different sensors. The proposed method, published online as an off-the-shelf software tool, allows users to perform fast and high-quality CNN-based pansharpening of their own target images on general-purpose hardware. [ABSTRACT FROM AUTHOR]

Domain adaptation manages to build an effective target classifier or regression model for unlabeled target data by utilizing the well-labeled source data but lying different distributions. Intuitively, to address domain shift problem, it is crucial to learn domain invariant features across domains, and most existing approaches have concentrated on it. However, they often do not directly constrain the learned features to be class discriminative for both source and target data, which is of vital importance for the final classification. Therefore, in this paper, we put forward a novel feature learning method for domain adaptation to construct both domain invariant and class discriminative representations, referred to as DICD. Specifically, DICD is to learn a latent feature space with important data properties preserved, which reduces the domain difference by jointly matching the marginal and class-conditional distributions of both domains, and simultaneously maximizes the inter-class dispersion and minimizes the intra-class scatter as much as possible. Experiments in this paper have demonstrated that the class discriminative properties will dramatically alleviate the cross-domain distribution inconsistency, which further boosts the classification performance. Moreover, we show that exploring both domain invariance and class discriminativeness of the learned representations can be integrated into one optimization framework, and the optimal solution can be derived effectively by solving a generalized eigen-decomposition problem. Comprehensive experiments on several visual cross-domain classification tasks verify that DICD can outperform the competitors significantly. [ABSTRACT FROM AUTHOR]

It is important to predict the hidden behavior of a complex system. In the existing models for predicting the hidden behavior, the hidden belief rule base (HBRB) is an effective model which can use qualitative knowledge and quantitative data. However, the frame of discernment (FoD) of HBRB which is composed of some states or propositions and the universal set including all states or propositions is not complete. The global ignorance and local ignorance cannot be considered at the same time, which may lead to the inaccurate forecasting results. To solve the problems, a new HBRB model named as PHBRB in which the hidden behavior is described on the FoD of the power set is proposed in this correspondence paper. Furthermore, by using the evidential reasoning rule as the inference tool of PHBRB, a new projection covariance matrix adaption evolution strategy is developed to optimize the parameters of PHBRB so that more accurate prediction results can be obtained. A case study of network security situation prediction is conducted to demonstrate the effectiveness of the newly proposed method. [ABSTRACT FROM AUTHOR]

Inspired by security applications in the industrial Internet of things, this paper focuses on the usage of impulsive neural network (NN) synchronization technique for intelligent image protection against illegal swiping and abuse. A class of nonlinear interconnected NNs with transmission delay and random impulse effect is first formulated and analyzed in this paper. In order to make network protocols more flexible, a randomized broadcast impulsive coupling scheme is integrated into the protocol design. Impulsive synchronization criteria are then derived for the chaotic NNs in the presence of nonlinear protocol and random broadcast impulse, with the impulse effect discussed. Illustrative examples are provided to verify the developed impulsive synchronization results and to show its potential application in image encryption and decryption. [ABSTRACT FROM AUTHOR]

The weights of a multilayer perceptron (MLP) may be altered by multiplicative and/or additive noises if it is implemented in hardware. Moreover, if an MLP is implemented using analog circuits, it is prone to stuck-at 0 faults, i.e., link failures. In this paper, we have proposed a methodology for making an MLP robust with respect to link failures, multiplicative noise, and additive noise. This is achieved by penalizing the system error with three regularizing terms. To train the system we use a weighted sum of the following four terms: 1) mean squared error (MSE); 2) ${l^{2}}$ norm of the weight vector; 3) sum of squares of the first order derivatives of MSE with respect to weights; and 4) sum of squares of the second order derivatives of MSE with respect to weights. The proposed approach has been tested on ten regression and ten classification tasks with link failure, multiplicative noise, and additive noise scenarios. Our experimental results demonstrate the effectiveness of the proposed regularization to achieve robust training of an MLP. [ABSTRACT FROM AUTHOR]