Search

In this paper, we experimentally and analytically evaluate the reliability of two state-of-the-art neural networks for linear regression and pattern recognition (multilayer perceptron and single-layer perceptron) implemented in a system-on-chip composed of a field-programmable gate array (FPGA) and a microprocessor. We have considered, for each neural network, three different activation function complexities, to evaluate how the implementation affects FPGAs reliability. As we show in this paper, higher complexity increases the exposed area but reduces the probability of one failure to impact the network output. In addition, we propose to distinguish between critical and tolerable errors in artificial neural networks. Experiments using a controlled heavy-ions beam show that, for both networks, only about 30% of the observed output errors actually affect the outputs correctness. We identify the causes of critical errors through fault injection, and found that faults in initial layers are more likely to significantly affect the output. [ABSTRACT FROM PUBLISHER]

Deep neural networks have proven to be able to learn rich internal representations, including for features that can also be used for different purposes than those the networks are originally developed for. In this paper, we are interested in exploring such ability and, to this aim, we propose a novel approach for investigating the internal behavior of networks trained for source code processing tasks. Using a simple autoencoder trained in the reconstruction of vectors representing programs (i.e., program embeddings), we first analyze the performance of the internal neurons in classifying programs according to different labeling policies inspired by real programming issues, showing that some neurons can actually detect different program properties. We then study the dynamics of the network from an information-theoretic standpoint, namely by considering the neurons as signaling systems and by computing the corresponding entropy. Further, we define a way to distinguish neurons according to their behavior, to consider them as formally associated with different abstract concepts, and through the application of nonparametric statistical tests to pairs of neurons, we look for neurons with unique (or almost unique) associated concepts, showing that the entropy value of a neuron is related to the rareness of its concept. Finally, we discuss how the proposed approaches for ranking the neurons can be generalized to different domains and applied to more sophisticated and specialized networks so as to help the research in the growing field of explainable artificial intelligence. [ABSTRACT FROM AUTHOR]

Multilayer neural networks are current trends in machine learning. Although complex architectures bring high performance, having sparse neurons in each layer can save memory, energy, and computational resources. In this paper, we aim to balance benefits between the complexity of architectures and the sparsity of neurons. An algorithm is proposed to prune neurons in multilayer neural networks through the global sensitivity analysis. Motivated by layer-wise training, we construct autoencoders with linear decoders, so mathematical models of multilayer neural networks can be considered as additive models. Hence, a first-order sensitivity analysis method, called random balance designs (RBD), is employed to select redundant neurons in hidden layers. This paper provides a novel framework to apply RBD in multilayer neural networks. Multiple experimental results demonstrate the generality and effectiveness of the proposed approach on structural learning of neural networks. After removing superfluous hidden neurons, higher accuracy can be obtained in most cases with less computation. [ABSTRACT FROM AUTHOR]

Current neural network based algorithmic composition methods are very different compared to human brain's composition process, while the biological plausibility of composition and generative models are essential for the future of Artificial Intelligence. To explore this problem, this paper presents a spiking neural network based on the inspiration from brain structures and musical information processing mechanisms at multiple scales. Unlike previous methods, our model has three novel characteristics: (1) Inspired by brain structures, multiple brain regions with different cognitive functions, including musical memory and knowledge learning, are simulated and cooperated to generate stylistic melodies. A hierarchical neural network is constructed to formulate musical knowledge. (2) Biologically plausible neural model is employed to construct the network and synaptic connections are modulated using spike-timing-dependent plasticity (STDP) learning rule. Besides, brain oscillation activities with different frequencies perform importantly during the learning and generating process. (3) Based on significant musical memory and knowledge learning, genre-based and composer-based melody composition can be achieved by different neural circuits, the experiments show that the model can compose melodies with different styles of composers or genres. [ABSTRACT FROM AUTHOR]

Recurrent neural networks based on reservoir computing are increasingly being used in many applications. Optimization of the topological structure of the reservoir and the internal connection weights for a given task is one of the most important problems in reservoir computing. In this paper, considering the fact that one can construct a large matrix using Kronecker products of several small-size matrices, we propose a method to optimize the reservoir. Having a small number of parameters to optimize, a gradient based algorithm is applied to optimize parameters, and consequently the reservoir. In addition to reducing the number of parameters for optimization, potentially, the method is able to control several other properties of the reservoir such as spectral radius, sparsity, weight distribution and underlying connections, i.e. connection topology. To reveal the effectiveness of the proposed optimization method, the application to the following tasks are considered: Nonlinear autoregressive moving average and multiple superimposed oscillators. Simulation results show satisfactory performance of the method. [ABSTRACT FROM PUBLISHER]

Shows how seemingly minor differences in training and evaluation procedures used in some studies of recurrent neural networks as context free language predictors can lead to significant differences in apparent network performance. Experimental methods; Results of the study; Conclusion.

Combining a deep neural network with fuzzy theory, this paper proposes an advertising click-through rate (CTR) prediction approach based on a fuzzy deep neural network (FDNN). In this approach, fuzzy Gaussian-Bernoulli restricted Boltzmann machine (FGBRBM) is first applied to input raw data from advertising datasets. Next, fuzzy restricted Boltzmann machine (FRBM) is used to construct the fuzzy deep belief network (FDBN) with the unsupervised method layer by layer. Finally, fuzzy logistic regression (FLR) is utilized for modeling the CTR. The experimental results show that the proposed FDNN model outperforms several baseline models in terms of both data representation capability and robustness in advertising click log datasets with noise. [ABSTRACT FROM AUTHOR]

Recent advances in deep learning (DL) have stimulated increasing interests in neural networks (NN). From the perspective of operation type and network architecture, deep neural networks can be categorized into full convolution-based neural network (ConvNet), recurrent neural network (RNN), and fully-connected neural network (FCNet). Different types of neural networks are usually cascaded and combined as a hybrid neural network (Hybrid-NN) to complete real-life cognitive tasks. Such hybrid-NN implementation is memory-intensive with large number of memory accesses, hence the performance of hybrid-NN is often limited by the insufficient memory bandwidth. A “3D + 2.5D” integration system, which integrates a high-bandwidth 3D stacked DRAM side-by-side with a highly-parallel neural processing unit (NPU) on a silicon interposer, overcomes the bandwidth bottleneck in hybrid-NN acceleration. However, intensive concurrent 3D DRAM accesses produced by the NPU lead to a serious thermal problem in 3D DRAM. In this paper, we propose a neural processor called Parana for hybrid-NN acceleration in consideration of thermal problem of 3D DRAM. Parana solves the thermal problem of 3D memory by optimizing both the total number of memory accesses and memory accessing behaviors. For memory accessing behaviors, Parana balances the memory bandwidth by spatial division mapping hybrid-NN onto computing resources, which efficiently avoids that masses of memory accesses are issued in a short time period. To reduce the total number of memory accesses, we design a new NPU architecture and propose a memory-oriented tiling and scheduling mechanism to exploit the maximum utilization of on-chip buffer. Experimental results show that Parana reduces the peak temperature by up to 54.72 $^\circ$ C and the steady temperature by up to 32.27 $^\circ$ C over state-of-the-art accelerators with 3D memory without performance degradation. [ABSTRACT FROM AUTHOR]

This paper deals with the global robust non-fragile Mittag-Leffler synchronization issue for uncertain fractional-order neural networks with discontinuous activations. Firstly, a new inequality, which is concerned with the fractional derivative of the variable upper limit integral for the non-smooth integrable functional, is developed, and to be applied in the main results analysis. Then, the appropriate non-fragile controller with two types of gain perturbations is designed, and the global asymptotical stability is discussed for the synchronization error dynamical system by applying Lyapunov functional approach, non-smooth analysis theory and inequality analysis technique. In addition, the robust non-fragile Mittag-Leffler synchronization conditions are addressed in terms of linear matrix inequalities. Finally, two numerical examples are given to demonstrate the feasibility of the proposed non-fragile controller and the validity of the theoretical results. [ABSTRACT FROM AUTHOR]

This paper is concerned with a class of cellular neural networks with proportional delay and impulses. First, by employing the improved Razumikhin technique and Lyapunov functions, some delay-dependent criteria are established to guarantee asymptotic stability and global stability of a class of general impulsive differential equations with proportional delay. Second, applying the obtained criteria, we get some delay-dependent sufficient conditions ensuring the existence, uniqueness and globally asymptotic stability of the equilibrium point of the cellular neural networks with proportional delay and impulses presented in this paper. Finally, three examples are presented to illustrate the effectiveness and advantages of the results obtained. [ABSTRACT FROM AUTHOR]

In this paper, an adaptive robust finite-time neural control scheme is proposed for uncertain permanent magnet synchronous motor servo system with nonlinear dead-zone input. According to the differential mean value theorem, the dead zone is represented as a linear time-varying system, and the model uncertainty including the dead zone is approximated by using a simple neural network. Then, an adaptive finite-time controller is designed based on a fast terminal sliding mode control principle, and the singularity problem in the initial TSMC is circumvented by modifying the terminal sliding manifold. Comparative experiments are conducted to validate the effectiveness and superior performance of the proposed method. [ABSTRACT FROM AUTHOR]

This paper addresses the globally stable tracking control problem of a class of uncertain multiple-input-multiple-output nonlinear systems. By employing the radial basis function neural networks to compensate for the system uncertainties, a novel switching controller is developed. The key features of the proposed control scheme are presented as follows. First, to design the desired adaptive neural controller successfully, an nth-order smoothly switching function is constructed originally. Second, the number of the neural networks and the adaptive parameters is reduced by adopting the direct adaptive approach, so a simplified controller is designed and it is easy to implement in practice. By utilizing the special properties of the affine terms of the considered systems, the singularity problem of the controller is completely avoided. Finally, the overall controller guarantees that all the signals in the closed-loop system are globally uniformly ultimately bounded and the system output converges to a small neighborhood of the reference trajectory by appropriately choosing the design parameters. A simulation example is given to illustrate the effectiveness of the proposed control scheme. [ABSTRACT FROM AUTHOR]

In this paper, the event-triggered H state estimation problem is investigated for a class of neural networks with mixed time delays and sensor saturations. The mixed time delays consist of discrete and distributed delays. The measurement outputs are subject to the sensor saturations due to the physical constraints. Through the available measurement outputs, the main purpose of the addressed problem is to design a state estimator to estimate the actual neural states. In order to improve the efficiency in resource utilization, an event-triggered mechanism is employed to decide whether the received measurement output is transmitted to the state estimator. Different from the existing event-triggering strategies, the triggering condition is given for each sensor, and the measurement output from each sensor is sent according to their separate triggering conditions. By using the Lyapunov functional approach, sufficient conditions are derived to guarantee that the estimation error dynamics is exponentially stable and the H performance requirement is satisfied. Then, the desired H state estimator is designed in terms of the solution to a linear matrix inequality that can be easily solved by the MATLAB toolboxes. Finally, one simulation example is provided to show the effectiveness of the proposed event-triggered estimation scheme. [ABSTRACT FROM AUTHOR]

This paper presents theoretical and historical backgrounds related to neural network rule extraction. It also investigates approaches for neural network rule extraction by ensemble concepts. Bologna pointed out that although many authors had generated comprehensive models from individual networks, much less work had been done to explain ensembles of neural networks. This paper carefully surveyed the previous work on rule extraction from neural network ensembles since 1988. We are aware of three major research groups i.e., Bologna' group, Zhou' group and Hayashi' group. The reason of these situations is obvious. Since the structures of previous neural network ensembles were quite complicated, the research on the efficient rule extraction algorithm from neural network ensembles was few although their learning capability was extremely high. Thus, these issues make rule extraction algorithm for neural network ensemble difficult task. However, there is a practical need for new ideas for neural network ensembles in order to realize the extremely high-performance needs of various rule extraction problems iu real life. This paper successively explain nature of artificial neural networks, origin of neural network rule extraction, incorporating fuzziness in neural network rule extraction, theoretical foundation of neural network rule extraction, computational complexity of neural network rule extraction, neuro-fuzzy hybridization, previous rule extraction from neural network ensembles and difficulties of previous neural network ensembles. Next, this paper address three principles of proposed neural network rule extraction: to increase recognition rates, to extract rules from neural network ensembles, and to minimize the use of computing resources. We also propose an ensemble-recursive-rule extraction (E-Re-RX) by two or three standard backpropagation to train multi-layer perceptrons (MLPs), which enabled extremely high recognition accuracy and the extraction of comprehensible rules. Furthermore, this enabled rule extraction that resulted in fewer rules than those in previously proposed methods. This paper summarizes experimental results of rule extraction using E-Re-RX by multiple standard backpropagation MLPs and provides deep discussions. The results make it possible for the output from a neural network ensemble to be in the form of rules, thus open the "black box" of trained neural networks ensembles. Finally, we provide valuable conclusions and as future work, three open questions on the E-Re-RX algorithm. [ABSTRACT FROM AUTHOR]

Implementing associative memory experiments with nanoscale memristors is an interesting subject, which can allow robots to mimic human thinking. This paper is concerned with a new emotion model of memristor-based neural network and its circuit implementation. The model has three inputs and two outputs, which can feel happy when it receives good news or feel sad when it receives bad news. Unknown news can be recognized through associative memory, which simulates human emotions. Associative learning and three kinds of forgetting process together make up the full-function emotion model. In addition, the Ag/AgInSbTe/Ta-based model closely related to the actual physical properties of memristors is used to design synaptic structures. The circuits of memristor-based neural networks are also simplified. Furthermore, the new emotion model is able to adjust the changing rate of emotion. The process reflects the fact that humans learn the same thing faster at the second time, compared with the first time. Finally, PSPICE is used to simulate all the circuits of the emotion model. The presented emotion model in this paper offers more possibilities for designing intelligent machines. [ABSTRACT FROM AUTHOR]

The recognition technology that recognizes or discriminates certain individuals is very important for the security that provides intelligence services. Face recognition rate can vary depending on variability of the face itself as well as other external factors such as illumination, background, angle and distance of a camera position. The paper suggests a proper method for long-distance face recognition by resolving the change in recognition rate resulting from distance change in long-distance face recognition. For the long-distance face recognition test, face images by actual distance from 1 to 9 m away were obtained directly. Actual face images taken by distance were applied to resolve the issue rising from distance change and CNN was applied to extract overall features of face. The test showed that proposed face recognition algorithm that used CNN as feature extraction and face images by actual distance for training was found to show the best performance. [ABSTRACT FROM AUTHOR]

This paper studies a general class of dynamical neural networks with lateral inhibition, exhibiting winner-take-all (WTA) behavior. These networks are motivated by a metal-oxide-semiconductor field effect transistor (MOSFET) implementation of neural networks, in which mutual competition plays a very important role. We show that for a fairly general class of competitive neural networks, WTA behavior exists. Sufficient conditions for the network to have aWTA equilibrium are obtained, and rigorous convergence analysis is carried out. The conditions for the network to have the WTA behavior obtained in this paper provide design guidelines for the network implementation and fabrication. We also demonstrate that whenever the network gets into the WTA region, it will stay in that region and settle down exponentially fast to the WTA point. This provides a speeding procedure for the decision making: as soon as it gets into the region, the winner can be declared. Finally, we show that this WTA neural network has a self-resetting property, and a resetting principle is proposed. [ABSTRACT FROM AUTHOR]

Abstract-The robust state estimation problem for a class of uncertain neural networks with time-varying delay is studied in this paper. The parameter uncertainties are assumed to be norm bounded. Based on a new bounding technique, a sufficient condition is presented to guarantee the existence of the desired state estimator for the uncertain delayed neural networks. The criterion is dependent on the size of the time-varying delay and on the size of the time derivative of the time-varying delay. It. is shown that the design of the robust state estimator for such neural networks can be achieved by solving a linear matrix inequality (LMI), which can be easily facilitated by using some standard numerical packages. Finally, two simulation examples are given to demonstrate the effectiveness of the developed approach. [ABSTRACT FROM AUTHOR]

Abstract-In this paper, neural networks are used along with two-player policy iterations to solve for the feedback strategies of a continuous-time zero-sum game that appears in L2-gain optimal control, suboptimal H∞ control, of nonlinear systems affine in input with the control policy having saturation constraints. The result is a closed-form representation, on a prescribed compact set chosen a priori, of the feedback strategies and the value function that solves the associated Hamilton-Jacobi-Isaacs (HJI) equation. The closed-loop stability, L2-gain disturbance attenuation of the neural network saturated control feedback strategy, and uniform convergence results are proven. Finally, this approach is applied to the rotational/translational actuator (RTAC) nonlinear benchmark problem under actuator saturation, offering guaranteed stability and disturbance attenuation. [ABSTRACT FROM AUTHOR]

In this paper, the recurrent neural networks (RNNs) with a generalized activation function class is proposed. In this proposed model, every component of the neuron's activation function belongs to a convex hull which is bounded by two odd symmetric piecewise linear functions that are convex or concave over the real space. All of the convex hulls are composed of generalized activation function classes. The novel activation function class is not only with a more flexible and more specific description of the activation functions than other function classes but it also generalizes some traditional activation function classes. The absolute exponential stability (AEST) of the RNN with a generalized activation function class is studied through three steps. The first step is to demonstrate the global exponential stability (GES) of the equilibrium point of original RNN with a generalized activation function being equivalent to that of RNN under all vertex functions of convex hull. The second step transforms the RNN under every vertex activation function into neural networks under an array of saturated linear activation functions. Because the GES of the equilibrium point of three systems are equivalent, the next stability analysis focuses on the GES of the equilibrium point of RNN system under an array of saturated linear activation functions. The last step is to study both the existence of equilibrium point and the GES of the RNN under saturated linear activation functions using the theory of M-matrix. In the end, a two-neuron RNN with a generalized activation function is constructed to show the effectiveness of our results. [ABSTRACT FROM AUTHOR]

The electromagnetic torque introduces ripples into the electromechanical motion system due to nonlinearities, such as uncertain changes of magnet field, load, and friction, which generate speed oscillations and deteriorate the tracking performance of servo system. Furthermore, the minimum time response and smooth trajectory tracking are cruces in servo control. In this paper, an available method is proposed to solve them by using neural networks (NNs) and a nonlinear smooth trajectory filter (STF) for the robust smoothing feedforward control of a class of general nonlinear systems. First, the online weight-tuning scheme based on Lyapunov function can guarantee the boundedness of tracking error by good performance of NNs modeling nonlinear functions. Second, a feedforward controller based on the output of nonlinear STF is designed to guarantee minimum time response and smooth trajectory tracking. Finally, as a example, the motion system can be equivalent to the two-order system under the linear closed-loop current control in view of the (d, q) mathematic model for PM synchronous motor, so that this robust smoothing control method using neutral networks can be applied into position servo control. Moreover, the validity and effectiveness of this control method are verified through simulations and experiments. [ABSTRACT FROM AUTHOR]

This paper introduces a general class of neural networks with arbitrary constant delays in the neuron interconnections, and neuron activations belonging to the set of discontinuous monotone increasing and (possibly) unbounded functions. The discontinuities in the activations are an ideal model of the situation ai where the gain of the neuron amplifiers is very high and tends to infinity, while the delay accounts for the finite switching speed of the neuron amplifiers, or the finite signal propagation speed. It is known that the delay in combination with high-gain nonlinearities is a particularly harmful source of potential instability. The goal of this paper is to single out a subclass of the considered discontinuous neural networks for which stability is instead insensitive to the presence of a delay. More precisely, conditions are given under which there is a unique equilibrium point of the neural net. work, which is globally exponentially stable for the states, with a known convergence rate. The conditions are easily testable and independent of the delay. Moreover, global convergence in finite time of the state and output is investigated. In doing so, new interesting dynamical phenomena are highlighted with respect to the case without delay, which make the study of convergence in finite time significantly more difficult. The obtained results extend previous work on global stability of delayed neural networks with Lipschitz continuous neuron activations, and neural networks with discontinuous neuron activations but without delays. [ABSTRACT FROM AUTHOR]

Artificial neural networks have been extensively applied to document analysis and recognition. Most efforts have been devoted to the recognition of isolated handwritten and printed characters with widely recognized successful results. However, many other document processing tasks, like preprocessing, layout analysis, character segmentation, word recognition, and signature verification, have been effectively faced with very promising results. This paper surveys the most significant problems in the area of offline document image processing, where connectionist-based approaches have been applied. Similarities and differences between approaches belonging to different categories are discussed. A particular emphasis is given on the crucial role of prior knowledge for the conception of both appropriate architectures and learning algorithms. Finally, the paper provides a critical analysis on the reviewed approaches and depicts the most promising research guidelines in the field. In particular, a second generation of connectionist-based models are foreseen which are based on appropriate graphical representations of the learning environment. [ABSTRACT FROM AUTHOR]

In this letter, we point out that the linear matrix inequality (LMI)-based criterion obtained in the above paper for the global exponential stability of the delayed neural networks can be simplified to a simpler but equivalent form and, thus, show that it is not necessary to have such complex form of condition in the above paper. As a result, we also answer the question raised by the author of the above paper. [ABSTRACT FROM AUTHOR]

Purpose This paper aims to present a bio-inspired neural network for improvement of information processing capability of the existing artificial neural networks.Design/methodology/approach In the network, the authors introduce a property often found in biological neural system – hysteresis – as the neuron activation function and a bionic algorithm – extreme learning machine (ELM) – as the learning scheme. The authors give the gradient descent procedure to optimize parameters of the hysteretic function and develop an algorithm to online select ELM parameters, including number of the hidden-layer nodes and hidden-layer parameters. The algorithm combines the idea of the cross validation and random assignment in original ELM. Finally, the authors demonstrate the advantages of the hysteretic ELM neural network by applying it to automatic license plate recognition.Findings Experiments on automatic license plate recognition show that the bio-inspired learning system has better classification accuracy and generalization capability with consideration to efficiency.Originality/value Comparing with the conventional sigmoid function, hysteresis as the activation function enables has two advantages: the neuron’s output not only depends on its input but also on derivative information, which provides the neuron with memory; the hysteretic function can switch between the two segments, thus avoiding the neuron falling into local minima and having a quicker learning rate. The improved ELM algorithm in some extent makes up for declining performance because of original ELM’s complete randomness with the cost of a litter slower than before. [ABSTRACT FROM AUTHOR]

This paper is concerned with the generalized neutral-type neural networks with impulses and delays. By applying the contraction mapping principle and generalized Gronwall-Bellman’s inequality, we employ a novel argument to establish new results on the existence, uniqueness and exponential stability of piecewise pseudo almost periodic solutions. Some corresponding results in the literature can be enriched and extended. Moreover, a numerical example is given to illustrate the effectiveness of our results. [ABSTRACT FROM AUTHOR]

Speech recognition (SR) has been improved significantly by artificial neural networks (ANNs), but ANNs have the drawbacks of biologically implausibility and excessive power consumption because of the nonlocal transfer of real-valued errors and weights. While spiking neural networks (SNNs) have the potential to solve these drawbacks of ANNs due to their efficient spike communication and their natural way to utilize kinds of synaptic plasticity rules found in brain for weight modification. However, existing SNN models for SR either had bad performance, or were trained in biologically implausible ways. In this paper, we present a biologically inspired convolutional SNN model for SR. The network adopts the time-to-first-spike coding scheme for fast and efficient information processing. A biological learning rule, spike-timing-dependent plasticity (STDP), is used to adjust the synaptic weights of convolutional neurons to form receptive fields in an unsupervised way. In the convolutional structure, the strategy of local weight sharing is introduced and could lead to better feature extraction of speech signals than global weight sharing. We first evaluated the SNN model with a linear support vector machine (SVM) on the TIDIGITS dataset and it got the performance of 97.5%, comparable to the best results of ANNs. Deep analysis on network outputs showed that, not only are the output data more linearly separable, but they also have fewer dimensions and become sparse. To further confirm the validity of our model, we trained it on a more difficult recognition task based on the TIMIT dataset, and it got a high performance of 93.8%. Moreover, a linear spike-based classifier—tempotron—can also achieve high accuracies very close to that of SVM on both the two tasks. These demonstrate that an STDP-based convolutional SNN model equipped with local weight sharing and temporal coding is capable of solving the SR task accurately and efficiently. [ABSTRACT FROM AUTHOR]

In the context of Arabic optical characters recognition, Arabic poses more challenges because of its cursive nature. We purpose a system for recognizing a document containing Arabic text, using a pipeline of three neural networks. The first network model predicts the font size of an Arabic word, then the word is normalized to an 18pt font size that will be used to train the next two models. The second model is used to segment a word into characters. The problem of words segmentation in the Arabic language, as in many similar cursive languages, presents a challenge to the OCR systems. This paper presents a multichannel neural network to solve the offline segmentation of machine-printed Arabic documents. The segmented characters are then fed as an input to a convolutional neural network for Arabic characters recognition. The font size prediction model produced a test accuracy of 99.1%. The accuracy of the segmentation model using one font is 98.9%, while four-font model showed 95.5% accuracy. The whole pipeline showed an accuracy of 94.38% on Arabic Transparent font of size 18pt from APTI data set. [ABSTRACT FROM AUTHOR]

Exponential raise in data and images leads to evolution of Artificial Intelligence. Deep learning is a domain of Machine Learning and a unique form of Artificial Neural Networks from AI. These peculiar features facilitated to apply Deep Learning technique on Medical images. These applications acquire deep insights from the image processing. Diagnosing, tumor detection, predictive modeling on electronic health records, classification on clinical heart failures and Blood flow Quantification and visualization are some of the real time applications in health care industry for image processing. Bio-medical images have numerous parameters and variations with respect to dimensions and channels. Hence the images require deep analysis to detect subtle dissimilarities. De-noising, spatial normalization, bias-field correction and nature of acquisition are some of the challenges while handling image datasets. These challenges are addressed by applying TensorFlow framework using deep learning methods. In this paper a review on image processing using deep learning is discussed in order to improve the efficiency of the models by adapting tensor flow framework. [ABSTRACT FROM AUTHOR]

The relevance of the topic, in particular, if to take one of the information flows, whether it is the action of a human factor or a specific object, then it is true that special processing of the machine learning language and automatic information output significantly optimize human life. With the help of neural networks and their chest radiography is one of the most accessible radiological studies for screening and diagnosis of many lung diseases a special machine learning language is to study the flow of information about it and the same object in real time using neural networks. The article describes the terminology of the problem of X-ray recognition using machine learning methods and algorithms, examines the relevance of the problem, and analyzes the current state of the problem in the field of X-ray recognition. The aspects of the problem being solved, identified during the analysis, in the form of solved problems, approaches, methods, information technologies used, tools and software solutions to the problem are noted The paper is devoted to the description of a modified method of fuzzy clustering of halftone images, which at each iteration performs a dynamic transformation of the source data based on a singular decomposition with automatic selection of the most significant columns of the matrix of left singular vectors. The results of experimental studies were obtained by processing X-ray images. As a result of testing a neural network model, in the output layer of which a sigmoidal activation function was used to activate neurons, and an algorithm was used as an optimization method, the best values of accuracy and completeness were obtained: accuracy – 94.2 During testing, the neural network showed an accuracy of pneumonia recognition equal to 94,27 %. [ABSTRACT FROM AUTHOR]

In general, it is not possible to have access to all variables that determine the behavior of a system. Once a number of measurable variables is identified, there might still exist hidden variables which influence the behavior of the system. The result is model ambiguity in the sense that, for the same (or very similar) input values, distinct outputs are obtained. In addition, the degree of ambiguity may vary across the range of input values. Therefore, to evaluate the accuracy of a model it is important to devise a method to obtain the degree of reliability for each output result. In this paper, we present such a scheme composed of two coupled neural networks, the first one computing the average predicted value and the other the reliability of the output, which is learned from the error values of the first one. As an illustration, the scheme is applied to a model for tracking slopes in a straw chamber and to a credit scoring model. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) is an extensive scientific discipline which enables computer systems to solve problems by emulating complex biological processes such as learning, reasoning and self-correction. This paper presents a comprehensive review of the application of AI techniques for improving performance of optical communication systems and networks. The use of AI-based techniques is first studied in applications related to optical transmission, ranging from the characterization and operation of network components to performance monitoring, mitigation of nonlinearities, and quality of transmission estimation. Then, applications related to optical network control and management are also reviewed, including topics like optical network planning and operation in both transport and access networks. Finally, the paper also presents a summary of opportunities and challenges in optical networking where AI is expected to play a key role in the near future. [ABSTRACT FROM AUTHOR]

The scientific interest attracted by Spiking Neural Networks (SNN) has lead to the development of tools for the simulation and study of neuronal dynamics ranging from phenomenological models to the more sophisticated and biologically accurate Hodgkin-and-Huxley-based and multi-compartmental models. However, despite the multiple features offered by neural modelling tools, their integration with environments for the simulation of robots and agents can be challenging and time consuming. The implementation of artificial neural circuits to control robots generally involves the following tasks: (1) understanding the simulation tools, (2) creating the neural circuit in the neural simulator, (3) linking the simulated neural circuit with the environment of the agent and (4) programming the appropriate interface in the robot or agent to use the neural controller. The accomplishment of the above-mentioned tasks can be challenging, especially for undergraduate students or novice researchers. This paper presents an alternative tool which facilitates the simulation of simple SNN circuits using the multi-agent simulation and the programming environment Netlogo (educational software that simplifies the study and experimentation of complex systems). The engine proposed and implemented in Netlogo for the simulation of a functional model of SNN is a simplification of integrate and fire (I&F) models. The characteristics of the engine (including neuronal dynamics, STDP learning and synaptic delay) are demonstrated through the implementation of an agent representing an artificial insect controlled by a simple neural circuit. The setup of the experiment and its outcomes are described in this work. [ABSTRACT FROM AUTHOR]

Polymeric materials are present in various industrial sectors and in daily life, presenting advantages such as low cost and durability. Several processes for manufacturing have been developed. To achieve safety and operational goals measurement methods for proper process monitoring and effective control are needed. However, in real polymer plants, measuring devices are subject to uncertainties and are not always available. Hence, this paper proposes a virtual sensor scheme based on a particle filter and artificial neural network (ANN) that is applied to a simulated polymerization reactor. This scheme reduces uncertainties and enables the observation of latent variables. The ANN is also used for predicting the final properties of the polymer. The goal is to provide controllers with more complete and improved information. The results show that the virtual sensor scheme improves the process control, providing accurate estimates and action times that are consistent with industrial sampling intervals, which highlights its potential for practical applications. [ABSTRACT FROM AUTHOR]

Classification of alcoholic electroencephalogram (EEG) signals is a challenging job in biomedical research for diagnosis and treatment of brain diseases of alcoholic people. The aim of this study was to introduce a robust method that can automatically identify alcoholic EEG signals based on time-frequency (T-F) image information as they convey key characteristics of EEG signals. In this paper, we propose a new hybrid method to classify automatically the alcoholic and control EEG signals. The proposed scheme is based on time-frequency images, texture image feature extraction and nonnegative least squares classifier (NNLS). In T-F analysis, the spectrogram of the short-time Fourier transform is considered. The obtained T-F images are then converted into 8-bit grayscale images. Co-occurrence of the histograms of oriented gradients (CoHOG) and Eig(Hess)-CoHOG features are extracted from T-F images. Finally, obtained features are fed into NNLS classifier as input for classify alcoholic and control EEG signals. To verify the effectiveness of the proposed approach, we replace the NNLS classifier by artificial neural networks, k-nearest neighbor, linear discriminant analysis and support vector machine classifier separately, with the same features. Experimental outcomes along with comparative evaluations with the state-of-the-art algorithms manifest that the proposed method outperforms competing algorithms. The experimental outcomes are promising, and it can be anticipated that upon its implementation in clinical practice, the proposed scheme will alleviate the onus of the physicians and expedite neurological diseases diagnosis and research. [ABSTRACT FROM AUTHOR]

Researchers studying complex cognition have grown increasingly interested in mapping symbolic cognitive architectures onto subsymbolic brain models. Such a mapping seems essential for understanding cognition under all but the most extreme viewpoints (namely, that cognition consists exclusively of digitally implemented rules; or instead, involves no rules whatsoever). Making this mapping reduces to specifying an interface between symbolic and subsymbolic descriptions of brain activity. To that end, we propose parameterization techniques for building cognitive models as programmable, structured, recurrent neural networks. Feedback strength in these models determines whether their components implement classically subsymbolic neural network functions (e.g., pattern recognition), or instead, logical rules and digital memory. These techniques support the implementation of limited production systems. Though inherently sequential and symbolic, these neural production systems can exploit principles of parallel, analog processing from decision-making models in psychology and neuroscience to explain the effects of brain damage on problem solving behavior. [ABSTRACT FROM PUBLISHER]

'Particularism' and 'generalism' refer to families of positions in the philosophy of moral reasoning, with the former playing down the importance of principles, rules or standards, and the latter stressing their importance. Part of the debate has taken an empirical turn, and this turn has implications for AI research and the philosophy of cognitive modeling. In this paper, Jonathan Dancy's approach to particularism (arguably one of the best known and most radical approaches) is questioned both on logical and empirical grounds. Doubts are raised over whether Dancy's brand of particularism can adequately explain the graded nature of similarity assessments in analogical arguments. Also, simple recurrent neural network models of moral case classification are presented and discussed. This is done to raise concerns about Dancy's suggestion that neural networks can help us to understand how we could classify situations in a way that is compatible with his particularism. Throughout, the idea of a surveyable standard-one with restricted length and complexity-plays a key role. Analogical arguments are taken to involve multidimensional similarity assessments, and surveyable contributory standards are taken to be attempts to articulate the dimensions of similarity that may exist between cases. This work will be of relevance both to those who have interests in computationally modeling human moral cognition and to those who are interested in how such models may or may not improve our philosophical understanding of such cognition. [ABSTRACT FROM AUTHOR]

In this paper, we consider a class of fuzzy BAM neural networks with distributed delays and variable coefficients. By using Brouwer fixed point theorem and differential inequality, we obtain some sufficient conditions for the existence and global exponential stability of periodic solutions for a class of fuzzy BAM neural networks with distributed delays and variable coefficients. In addition, we also present an illustrative example to show the effectiveness of obtained results. [ABSTRACT FROM AUTHOR]

An improved hybrid particle swarm optimization (PSO)-based wavelet neural network (WNN) for Modeling the development of Fluid Dispensing for Electronic Packaging (MFD-EP) is presented in this paper. In modeling the fluid dispensing process, it is important to understand the process behavior as well as determine the optimum operating conditions of the process for a high-yield, low-cost, and robust operation. Modeling the fluid dispensing process is a complex nonlinear problem. This kind of problem is suitable to be solved by applying a neural network. Among the different kinds of neural networks, the WNN is a good choice to solve the problem. In the proposed WNN, the translation parameters are variables depending on the network inputs. Due to the variable translation parameters, the network becomes an adaptive one that provides better performance and increased learning ability than conventional WNNs. An improved hybrid PSO is applied to train the parameters of the proposed WNN. The proposed hybrid PSO incorporates a wavelet-theory-based mutation operation. It applies the wavelet theory to enhance the PSO in more effectively exploring the solution space to reach a better solution. A case study of MFD-EP is employed to demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]