Search

This paper investigates the periodicity and synchronization of inertial memristive neural networks with supremums and time delays. The analysis in this paper employs the results from the theory of differential equations with discontinuous right-hand side as introduced by Filippov. Further, by employing the second order differential inclusion theory and then choosing suitable variable transformation, the original system can be transformed into first order differential equations. By using the matrix measure method and Halany inequality techniques, the sufficient condition that guarantee the global exponential synchronization of drive-response system of coupled inertial memristive neural networks via the state feedback controller is derived. Furthermore, the global exponential periodicity of the addressed neural networks is also discussed. Finally, numerical examples and simulations are given to demonstrate the validation of the proposed results. [ABSTRACT FROM AUTHOR]

This paper investigates the problem of discrete-time stochastic recurrent neural networks with multiple time-varying delays in the leakage terms and impulses. A new set of sufficient conditions are obtained by constructing an appropriate Lyapunov-Krasovskii functional combining with linear matrix inequality technique and free weighting matrix method. The obtained delay-dependent stability conditions are expressed in terms of linear matrix inequalities and it can be solved via some available software packages. Up to now, the asymptotic stability problem is studied for discrete-delay in the leakage terms. For the first time in our paper, we have considered distributed delays and impulses for such kind of networks. In addition, we have provided a numerical example to demonstrate the effectiveness of our obtained stability results for the theoretical section. [ABSTRACT FROM AUTHOR]

This paper proposes a neural network based adaptive constrained control scheme for a solid oxide fuel cell (SOFC). First, a radial basis function (RBF) neural network is designed for the identification of SOFC dynamic model. The Jacobian information can be obtained through the identified RBF model. Then, a back propagation (BP) neural network based PID controller is designed to tune the parameters that BP neural network has strong self-learning and adaptive capabilities. At same time, in order to solve the control input saturation and fuel utilization problems of SOFC, a dynamic anti-windup compensator is proposed for accommodating the reference. Moreover this paper theoretically proves the stability of the proposed method based on Lyapunov stability analysis. Finally, the simulation results for SOFC are provided to demonstrate the effectiveness of the proposed constrained control approach. [ABSTRACT FROM AUTHOR]

In this paper, the problem of adaptive synchronization for time-delay neutral-type multi-slave neural networks with Markovian jumping parameters and stochastic noise is researched. The adaptive synchronization model which contains a drive system and multiple response systems is presented. By using of the generalized It o ^ ’ s formula and the M-matrix method, the sufficient condition is obtained to guarantee that the error control system is stable, and the update law of the feedback controller is designed to ensure that every slave system synchronizes with master system. Finally, a numerical example is given to illustrate the effectiveness of the method and result obtained in this paper. [ABSTRACT FROM AUTHOR]

While supervised learning approaches show great vitality and effectiveness in video object segmentation, most of them require large amounts of annotations which are expensive and time-consuming. Recently, self-supervised learning has attracted great attention by benefiting from unlabeled video sequences. However, current patch-based self-supervised video object segmentation methods only discriminate the patch from the entire image without distinguishing the object of interest from meaningless backgrounds or even occlusion. These disturbances deteriorate the extracted features and hinder the robustness of tracking when applied to real-world video sequences. In this paper, we propose a novel model named Tracker With Integration-Augmented Attention (TWIAA) to achieve both label-free and prominent performance. Specifically, we integrate both spatial and channel dimensions by introducing a feature spatial enhancement module and a two-stream channel module. With the combination of the two modules, the network can focus on exploring the discriminative object and suppressing the irrelevant part to improve the tracking robustness. Moreover, unlike other methods that calculate features separately on the search branch and template branch, the two designed modules coupled with the Siamese network compute the respective features of the search branch and the template branch jointly to augment the interdependence of the two branches. Such interdependence is injected into both spatial and channel dimensions. So that our approach establishes richer and more discriminative associations to identify the object more accurately. In addition, our method takes full advantage of cycle-consistency information in consecutive frames, which uses coherence as the learning signal to acquire object-oriented relationships. Extensive experiments and ablation studies are conducted on large VOS benchmarks, including DAVIS-2017, YouTube-VOS-2018, and YouTube-VOS-2019. The results verify that our proposed framework has both strong feature representation and competitive performance compared with supervised and self-supervised models. [ABSTRACT FROM AUTHOR]

Object detection based on deep learning typically relies on a large number of training data, which may be very labor-consuming to prepare. In this paper, we attempt to tackle the problem by addressing the One-Shot Object Detection (OSOD) task. Given a novel image denoted as the query image whose category label is not included in the training data, OSOD aims to detect objects of the same class in a complex scene denoted as the target image. The performance of recent OSOD methods is much weaker than general object detection. We find that one of the reasons behind this limited performance is that more false positives (i.e., false detections) are generated. Therefore, we argue that it is important to reduce the number of false positives generated in OSOD task to improve performance. To this end, we present a Focus On Classification One-Shot Object Detection (FOC OSOD) network. Specifically, we design the network from two perspectives: (1) how to obtain the effective similarity feature between the query image and target image; (2) how to classify the similarity feature effectively. To solve the above two challenges, firstly, we propose a Classification Feature Deformation-and-Attention (CFDA) module to obtain the high-quality query feature and target feature, so we can further generate effective similarity feature between them. Secondly, we present a Split Iterative Head (SIH) to improve the ability to classify the similarity feature. Extensive experiments on two public datasets (i.e., PASCAL VOC and COCO) demonstrate that the proposed framework achieves superior performance which outperforms other state-of-the-art methods with a considerable margin. [ABSTRACT FROM AUTHOR]

Weakly Supervised Object Localization (WSOL) is a technique for obtaining the object location from attention maps of the classification network, without using bounding box annotations. Existing WSOL approaches lack the modeling of the correlations between different regions of the target object. Hence they can only locate some discriminative attentions, which are small and sparse. Besides, they introduce too many background attentions when mining more object parts. In this paper, we propose a novel Adaptive Attention Augmentor (A3) to adaptively augment the target object attentions on class attention maps. It can supplement object attentions by discovering the semantic correspondence between different regions and dynamically suppress background attentions through the proposed Focal Dice loss. Extensive experiments demonstrate the effectiveness of our approach. On the ILSVRC dataset, A3 achieves a new state-of-the-art localization performance. On the fine-grained datasets including CUB-200–2011 and Cars-196, it also achieves very competitive results. [ABSTRACT FROM AUTHOR]

The biology-inspired intelligent computing system for the neuromorphic hardware implementation is useful in high-speed parallel information processing. However, the traditional Von Neumann computer architecture and the unsatisfactory signal transmission approach have jointly limited the overall performance of the specific hardware implementation. In this paper, a compact extreme learning machine (ELM) architecture synthesized with the spintronic memristor-based synaptic circuit, the biasing circuit, and the activation function circuit is presented. Notably, due to the threshold characteristic of the memristive device, the synaptic circuit has a bimodal behavior. Namely, it is capable to provide the constant and adjustable network weights between the adjacent layers in the ELM. Furthermore, two major limitations (process variations and sneak path issue) are taken into account for the detailed robustness analysis of the whole network. Finally, the entire scheme is verified with case studies in single image super-resolution (SR) reconstruction. [ABSTRACT FROM AUTHOR]

Person re-identification (Re-ID) is of important capability for artificial intelligence and human–computer interaction. The main challenge of person Re-ID lies in limited data to precisely capture a wide range of appearance variations over multiple viewpoints. Furthermore, compared to the Re-ID between the single person, the person groups contain information about the relationship between pedestrians that can potentially help identify certain identities. The Re-ID combining groups and individuals remain to be a promising task under rare study. In this paper, we propose a harmonious attention network for person re-identification, in which we jointly consider the complementarity between person groups and individuals. Concretely, first we propose a two-stream attentive network (TSAN) to respectively learn the information from the person groups and individuals. TSAN consists of a spatial–temporal fusion network for the group Re-ID, as well as a deep network for the traditionally individual person Re-ID. To jointly consider the contributions of the groups and individuals, then we propose a novel re-ranking algorithm (GIRK) based on the learned features to associate the group and individual information. We also propose a new group Re-ID dataset DukeGroupVid to evaluate the performance of our approach. Comprehensive experimental results on the proposed dataset and other Re-ID datasets demonstrate the effectiveness of our model. [ABSTRACT FROM AUTHOR]

It has been proved that deep neural networks are facing a new threat called backdoor attacks, where the adversary can inject backdoors into the neural network model through poisoning the training dataset. When the input containing some special pattern called the backdoor trigger, the model with backdoor will carry out malicious task such as misclassification specified by adversaries. In text classification systems, backdoors inserted in the models can cause spam or malicious speech to escape detection. Previous work mainly focused on the defense of backdoor attacks in computer vision, little attention has been paid to defense method for RNN backdoor attacks regarding text classification. In this paper, through analyzing the changes in inner LSTM neurons, we proposed a defense method called Backdoor Keyword Identification (BKI) to mitigate backdoor attacks which the adversary performs against LSTM-based text classification by data poisoning. This method can identify and exclude poisoning samples crafted to insert backdoor into the model from training data without a verified and trusted dataset. We evaluate our method on four different text classification datset: IMDB, DBpedia ontology, 20 newsgroups and Reuters-21578 dataset. It all achieves good performance regardless of the trigger sentences. [ABSTRACT FROM AUTHOR]

Artificial neural networks (ANNs) are known to be sensitive to the initial setting of parameters and the network architecture, such as the number of hidden nodes in multilayer perceptron (MLP). In this paper, we focus on a network structure which can help to find the proper number of hidden nodes in MLP. In this structure, so called Multilayer Perceptrons with Hidden Multipliers (MLPHM), each of the hidden nodes is associated with a tunable "gate" multiplier. With a specific regularization term, each gate tends to be opened or closed completely at the end of the training, and finally a pruned network is obtained. To study the fault tolerance and to improve the generalization of MLPHM, a noise-injected training scheme is proposed, with both multiplicative noise and additive noise taken into consideration. The objective functions and convergence theorems of the noise-injected training algorithms are obtained, and the latter have been verified by simulations. Applications to several UCI datasets have demonstrated that the proposed algorithms have efficient pruning ability and superior generalization ability. [ABSTRACT FROM AUTHOR]

DeepLab—one of the most critical deep neural network models for image segmentation—has achieved the most advanced performance in the field of image semantic understanding. However, the rich parameters and calculation of deep convolutional neural networks (DCNNs) demand further research on neuro-inspired computing chips specifically designed for hardware acceleration and AI applications. This motivates memristive solutions—the memristor integrates a range of merits, such as fast speed and nanoscale device, to provide an ideal implementation for building low-power neural computing chips and ultra-high-density non-volatile memory. Here, this paper proposes a memristive DeepLab (MDeepLab) system with software-hardware co-design, in which atrous convolution is used to enlarge the receptive field of neurons, the measure to avoid the gridding effect is adopted. Also, the weights are stored by leveraging the unit of a single crossbar array and the constant-term circuit, which significantly reduces the number of memristor devices, doubles the density, and nearly halves the power consumption conventional crossbar array. Finally, the effectiveness of the proposed scheme is verified by a series of experimental simulations and result analysis, showing the superiority of MDeepLab. This study is expected to provide a new solution for low-power consumption and real-time image processing of edge devices. [ABSTRACT FROM AUTHOR]

Dropout is an important training method for deep neural networks, because it can help avoid over-fitting. Traditional dropout methods and many extended dropout methods, omit some of the neurons' activation values according to the probabilities. These methods calculate the activation probability of neurons using the designed formula, without providing a plausible explanation of the calculation method. This paper proposes an adaptive sparse dropout (AS-Dropout) method for neural network training. The algorithm maps the neurons' activation values in a layer to a relative linear range of a sigmoid function, determines the ratio of active neurons by a probability calculation process, and drops most of neurons according to the probabilities. The probability calculation depends on the activation values of the neurons. The selection of active neurons is according to the probabilities. Therefore, AS-Dropout learns both the certainty and uncertainty in deep neural networks. Additionally, since only a small number of neurons are active, AS-Dropout increases the sparsity of the network. We applied AS-Dropout in different neural network structures. When evaluated on MNIST, COIL-100, and Caltech-101 datasets, the experimental results demonstrated that, overall, AS-Dropout substantially outperformed the traditional dropout and some improved dropout methods. [ABSTRACT FROM AUTHOR]

At present, multi-oriented text detection methods based on deep neural network have achieved promising performances on various benchmarks. Nevertheless, there are still some difficulties for arbitrary shape text detection, especially for a simple and proper representation of arbitrary shape text instances. In this paper, a pixel-based text detector is proposed to facilitate the representation and prediction of text instances with arbitrary shapes in a simple manner. Firstly, to alleviate the influence of the target vertex sorting and achieve the direct regression of arbitrary shape text instances, the starting-point-independent coordinates regression loss is proposed. Furthermore, to predict more accurate text instances, the text instance accuracy loss is proposed as an assistant task to refine the predicted coordinates under the guidance of IoU. To evaluate the effectiveness of our detector, extensive experiments have been carried on public benchmarks which contain arbitrary shape text instances and multi-oriented text instances. We obtain 84.8% of F-measure on Total-Text benchmark. The results show that our method can reach state-of-the-art performance. [ABSTRACT FROM AUTHOR]

Data visualization plays an important role in data analysis by displaying data to observers in an interpretable way. Visualizing multidimensional data requires projecting the data into a low-dimensional space that is visible to humans. In this paper, we propose a neural network model that can generate such projections while preserving the topology relationships within data points, which is named Visible Self Organizing Incremental Neural Network (V-SOINN). V-SOINN is able to construct a topology preserving visible network automatically and classify visible nodes to different classes in the low-dimensional space. The thought of topology preserving visualization stems from Self-Organizing Map (SOM). Compared to SOM, the main advantage of V-SOINN is that it does not need prior decision of network structure, including the number of nodes and grid in the output layer. V-SOINN can show the density distribution of datasets by using the activation counts of datasets. V-SOINN is able to depict the number of classes in the low-dimensional space as well. We perform experiments on artificial and real-world datasets, and V-SOINN outperforms PCA, MDS, t-SNE, Neural Gas and SOM on the datasets. Experiments show that V-SOINN can preserve the topology and V-SOINN can produce the correct classification result when the number of samples is small. [ABSTRACT FROM AUTHOR]

The existed methods for electroencephalograph (EEG) emotion recognition always train the models based on all the EEG samples indistinguishably. However, some of the source (training) samples may lead to a negative influence because they are significant dissimilar with the target (test) samples. So it is necessary to give more attention to the EEG samples with strong transferability rather than forcefully training a classification model by all the samples. Furthermore, for an EEG sample, from the aspect of neuroscience, not all the brain regions of an EEG sample contain emotional information that can transferred to the test data effectively. Even some brain region data will make strong negative effect for learning the emotional classification model. Considering these two issues, in this paper, we propose a transferable attention neural network (TANN) for EEG emotion recognition, which learns the emotional discriminative information by highlighting the transferable EEG brain regions data and samples adaptively through local and global attention mechanism. This can be implemented by measuring the outputs of multiple brain-region-level discriminators and one single sample-level discriminator. Extensive experiments on EEG emotion recognition demonstrate that the proposed TANN is superior to those state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Fixation as representation of one viewer's attention are very intuitive to reflect the viewer's observation procedure. The viewer's observation behavior can be further revealed by analyzing fixations features. In this paper, we propose a fixation based personalized salient object segmentation method involving personal observation behavior learning. Concretely, we design three neural networks and deploy a meta-learning method. The first network is a base segmentation network that can be converted into a meta-segmentation network by meta-learning. The meta- segmentation network can learn one viewer's observation behavior from only one sample and then generates the viewer's segmentation network to segment the other samples. Moreover, a fusion network plays an important role in alleviating an unsuitable transmission problem and generating a final segmentation result. The experimental results demonstrate the reasonability of our observation behavior learning and the effectiveness of the three proposed neural networks. [ABSTRACT FROM AUTHOR]

Click-through rate (CTR) prediction is essential for targeted advertising and recommendation. At present, machine learning models are widely used to build CTR estimators, including logistic regression (LR), factorization machine (FM), and deep neural network (DNN). Unfortunately, these models adopt the single structure that only considers either low-order feature interactions (such as LR and FM) or high-order feature interactions (such as DNN), not sufficient for CTR prediction. Therefore, the joint learning models are proposed, such as Wide & Deep and DeepFM, which can exploit both high- and low-order feature interactions to predict CTR by combining two different models. In this paper, we first analyze the typical CTR prediction models' structures and performance, and then summarize the general form and design rules of CTR estimators. Based on the general form, we further design a new joint learning model that combines two different residual networks to explore the feature interactions automatically. Compared with the widely adopted feed-forward neural network, the residual network is more capable of exploring complex feature interactions at different layers. Additionally, we introduce a neural attention network to learn the importance of each second-order interaction of features from various fields. Finally, we evaluate the prediction performance of the proposed model based on two real-world datasets (i.e., Criteo and Avazu) in terms of LogLoss and AUC metrics. The extensive experimental results demonstrate our model performs the best compared to the state-of-the-art baselines. [ABSTRACT FROM AUTHOR]

Recent studies have found that text can be represented in Hilbert space through a neural network driven by quantum probability, which provides a unified representation of texts with different granularities without losing the performance of downstream tasks. However, these quantum probability-inspired methods only focus on intra-document semantics and lack modeling global structural information. In this paper, we explore the potential of combining quantum probability with graph neural network, and propose a quantum probability-inspired graph neural network model to capture global structural information of interaction between documents for document representation and classification. We build a document interaction graph for a given corpus based on document word relation and frequency information, then learn a graph neural network driven by quantum probability on the defined graph. First, the proposed model represents each document node in the graph as a superposition state in a Hilbert space. Then the proposed model further computes density matrix representations for nodes to encode document interaction as mixed states. Finally, the model computes classification probability by performing quantum measurement on the mixed states. Experiments on four document classification benchmarks show that the proposed model outperforms a variety of classical neural network models and the previous quantum probability-inspired model with much smaller parameter size. Extended analyses also demonstrate the robustness of the proposed model with limited training data and its ability to learn semantically distinguishable document representation. [ABSTRACT FROM AUTHOR]

Skeleton based human action recognition has attracted more and more attentions recently thanks to the accessibility of depth sensors and the development of pose estimation techniques. Conventional approaches such as convolutional neural networks usually model skeletons with grid-shaped representations, which cannot explicitly explore the dependency between two correlated joints. In this paper, we treat the skeleton as a single graph with joints as nodes and bones as edges. Based on the skeleton graph, the improved graph convolutional network called adaptive multi-view graph convolutional networks(AMV-GCNs) is proposed to deal with skeleton based action recognition. We firstly construct a novel skeleton graph and two kinds of graph nodes are defined to model the spatial configuration and temporal dynamics respectively. Then the generated graphs along with feature vectors on graph nodes are fed into AMV-GCNs. In AMV-GCNs, an adaptive view transformation module is designed to reduce the impact of view diversity. Proposed module can automatically determine suitable viewpoints and transform skeletons to new representations under those viewpoints for better recognition. Further, we employ multiple GCNs based streams to utilize and learn action information from different viewpoints. Finally, the classification scores from multiple streams are fused to provide the recognition result. Extensive experimental evaluations on four challenging datasets, NTU RGB+D 60, NTU RGB+D 120, Northwestern-UCLA and UTD-MHAD, demonstrate the superiority of our proposed network. [ABSTRACT FROM AUTHOR]

Traffic congestion prediction is a vital part of Intelligent Transportation Systems in smart cities. Effective methods for traffic congestion prediction can help people make travel plans reasonably with Advanced Traveler Information Systems. Most of the existing methods for traffic congestion prediction was designed for a specific location. The parameters need to be modified when applying these methods to different locations. Other studies on the traffic network require sophisticated data pre-processing such as map matching. In this paper, we build a model named Relative Position Congestion Tensor and propose a Predictor for Position Congestion Tensor for traffic congestion prediction. First, we design a novel approach to construct congestion matrix on region traffic networks using the concept of relative locations for road nodes and convert matrices into three-dimensional spatio-temporal tensors. Then, we propose a method based on convolutional long-short term memory network to predict congestion at all locations of the road network in the near future. The experiments show that in all locations where congestion often occurs, the proposed method significantly outperforms baseline models including Linear Regression, Autoregressive Integrated Moving Average, Support Vector Regression, Random Forest, Gradient Boosting Regression, Long-Short Term Memory and generally outperforms the Convolution-based deep Neural Network modeling Periodic traffic data. Furthermore, we study the internal structure of the Predictor for Position Congestion Tensor model to analyze the interpretability of the model for congestion prediction. The results show that the proposed model can accurately capture the temporal and spatial characteristics of traffic. [ABSTRACT FROM AUTHOR]

Neural learning plays an important role in many applications. In this paper, we derive a new learning paradigm for neural networks. Most existing neural models train network parameters including connecting weights and biases via optimizing a loss or energy function. Inspired from the associative learning in brain, we propose to associate different patterns via modeling joint distribution of them based on a hierarchical architecture. We first define an energy function based on the distance between hierarchical features of different patterns. Then a Gibbs typological distribution is constructed according to the energy field. To optimize the model, it is necessary to estimate the gradient expectation via sampling. Different from the simple architecture of existing probabilistic neural models such as restricted Boltzmann machine, the difficulty of optimizing this model lies in uncomputable probability for sampling from the distribution. Then we propose an optimization based sampling method. After learning, conditional probability can be derived and the unknown pattern can be generated via sampling as well. Compared with existing neural learning models, the proposed deep associative learning can associate different patterns and can be directly applied to many learning problems. Experiments on problems of classification, image transformation, and image change detection verify the effectiveness of the proposed learning paradigm. [ABSTRACT FROM AUTHOR]

Gradient-based algorithms are commonly used for training radial basis function neural network (RBFNN). However, one of the challenges in the training process is determining how to avoid vanishing gradient. To solve this problem, an accelerated gradient algorithm (AGA) is designed to improve the learning performance of RBFNN in this paper. First, an indirect detection mechanism, based on the instantaneous gradient decay rate (IGDR) and instantaneous convergence rate (ICR), is developed to identify the vanishing gradient in learning process. Second, an amplification gradient strategy (AGS), which can increase the gradient value of learning parameters, is designed to accelerate the learning speed of RBFNN. Third, the analysis of AGA-based RBFNN (AGA-RBFNN) is given to guarantee the successful application. Finally, some benchmark and real problems are used to illustrate the effectiveness of AGA-RBFNN. The results demonstrate the effectiveness of AGA-RBFNN. [ABSTRACT FROM AUTHOR]

Spiking Neural Networks (SNNs) are brain-inspired, event-driven machine learning algorithms that have been widely recognized in producing ultra-high-energy-efficient hardware. Among existing SNNs, unsupervised SNNs based on synaptic plasticity, especially Spike-Timing-Dependent Plasticity (STDP), are considered to have great potential in imitating the learning process of the biological brain. Nevertheless, the existing STDP-based SNNs have limitations in constrained learning capability and/or slow learning speed. Most STDP-based SNNs adopted a slow-learning Fully-Connected (FC) architectures and used a sub-optimal vote-based scheme for spike decoding. In this paper, we overcome these limitations with: 1) a design of high-parallelism network architecture, inspired by the Inception module in Artificial Neural Networks (ANNs); 2) use of a Vote-for-All (VFA) decoding layer as a replacement to the standard vote-based spike decoding scheme, to reduce the information loss in spike decoding and, 3) a proposed adaptive repolarization (resetting) mechanism that accelerates SNNs' learning by enhancing spiking activities. Our experimental results on two established benchmark datasets (MNIST/EMNIST) show that our network architecture resulted in superior performance compared to the widely used FC architecture and a more advanced Locally-Connected (LC) architecture, and that our SNN achieved competitive results with state-of-the-art unsupervised SNNs (95.64%/80.11% accuracy on the MNIST/EMNISE dataset) while having superior learning efficiency and robustness against hardware damage. Our SNN achieved great classification accuracy with only hundreds of training iterations, and random destruction of large numbers of synapses or neurons only led to negligible performance degradation. [ABSTRACT FROM AUTHOR]

[Display omitted] • We summarize the state-of-the-art of convolutional neural architecture search. • We model the convolutional neural network design as a bi-level optimization problem. • We develop BLOP-CNN as a new solution approach to our bi-level model. • We evaluate the performance of our proposal with respect to relevant existing works. During the last decade, deep neural networks have shown a great performance in many machine learning tasks such as classification and clustering. One of the most successful networks is the CNN (Convolutional Neural Network), which has been applied in many application domains such as pattern recognition, medical diagnosis, and signal processing. Despite the very interesting performance of CNNs, their architecture design is still so far a major challenge for researchers and practitioners. Several works have been proposed in the literature with the aim to find optimized architectures such as ResNet and VGGNet. Unfortunately, most of these architectures are either manually defined by experts or automatically designed by greedy induction algorithms. Recent works suggest the use of Evolutionary Algorithms (EAs) thanks to their ability to escape locally-optimal architectures. Despite the fact that EAs have shown interesting performance, researchers in this direction have considered the design task as a single-level optimization problem; which represents the main research gap we tackle in this paper. The main contribution behind our work consists in the fact that CNN architecture design has a hierarchical nature and thus could be seen as a Bi-Level Optimization Problem (BLOP) where: (1) the upper level minimizes the network complexity defined by the number of blocks and the number of nodes per block; and (2) the lower level optimizes the convolution block 'graphs' topologies by maximizing the classification accuracy. Motivated by the originality of our observation with respect to the state of the art, we frame for the first time the CNN architecture design problem as a BLOP and then solve it using an adapted version of an existing efficient bi-level EA; through the definition of the solution encoding, the fitness function, and the variation operators at each level. The adapted EA is named BLOP-CNN and is assessed on the image classification task using the commonly employed CIFAR-10 and CIFAR-100 benchmark data sets. The analysis of our experimental results show the merits of our proposed method in providing the user with optimized architectures that outperform many recent and prominent architectures coming from the three different approaches, namely: manual design, reinforcement learning-based generation, and evolutionary optimization. Moreover, to show the applicability of our approach, we have conducted a case study on the detection of the COVID-19 using a set of benchmark chest X-ray and Computed Tomography (CT) images. [ABSTRACT FROM AUTHOR]

Saliency detection based on the complementary information from RGB images and depth maps has recently gained great popularity. In this paper, we propose C omplementary A ttention and A daptive I ntegration Net work (CAAI-Net), a novel RGB-D saliency detection model that integrates complementary attention based feature concentration and adaptive cross-modal feature fusion into a unified framework for accurate saliency detection. Specifically, we propose a context-aware complementary attention (CCA) module, which consists of a feature interaction component, a complementary attention component, and a global-context component. The CCA module first utilizes the feature interaction component to extract rich local context features. The resulting features are then fed into the complementary attention component, which employs the complementary attention generated from adjacent levels to guide the attention at the current layer so that the mutual background disturbances are suppressed and the network focuses more on the areas with salient objects. Finally, we utilize a specially-designed adaptive feature integration (AFI) module, which sufficiently considers the low-quality issue of depth maps, to aggregate the RGB and depth features in an adaptive manner. Extensive experiments on six challenging benchmark datasets demonstrate that CAAI-Net is an effective saliency detection model and outperforms nine state-of-the-art models in terms of four widely-used metrics. In addition, extensive ablation studies confirm the effectiveness of the proposed CCA and AFI modules. [ABSTRACT FROM AUTHOR]

Facial image inpainting is a challenging task since the facial images lose much content causing blur and unnaturalness. In this paper, we propose facial image inpainting using attention-based multi-level generative network. We adopt multi-level feature processing to reduce the training and testing time while enhancing the performance by reducing the number of channels of each convolutional layer in the generator. We combine attention with multi-level feature processing to maintain soft correlation with the surrounding content. For network optimization, we utilize two kinds of loss functions: content and texture. Content loss consists of mean absolute error (MAE) and edge-preserving losses, and produces inpainting results close to the ground truth. Texture loss consists of adversarial and perceptual losses to fine-tune the texture synthesis. Besides, we use the edge-preserving loss to take edge and patch similarity. Comparative experiments indicate that the proposed method produces photo-realistic and plausible inpainting results in random masks as well as outperforms the state-of-the-art ones in terms of quantitative measurements and subjective evaluations. [ABSTRACT FROM AUTHOR]

Siamese networks have achieved great success in object tracking due to the balance of precision and speed. However, Siamese trackers usually utilize the local feature of the last layer, which may degrade tracking performance in some difficult scenarios. In this paper, we propose a novel Channel and Spatial Attention-guided Residual learning framework for Tracking, referred to as CSART, which can improve feature representation of Siamese networks by exploiting self-attention mechanism to capture powerful contextual information. Specifically, to be efficient and seamless integration, different kinds of self-attention are appended on the template and search branches of Siamese networks respectively, that model global semantic inter-dependencies in channel and spatial dimensions. To avoid representation degradation, we consider to adaptively aggregate basic feature and its attention-weighted features with residual learning. Furthermore, a joint loss consisting of classic logistic loss as well as focal softmax loss is designed to emphasize difficult samples and guide the learning process of the whole model. Benefiting from the above scheme, CSART alleviates the over-fitting problem to some extent and enhances the discriminability. Extensive experiments on six popular tracking datasets indicate that the proposed tracker achieves better performance with a speed of 65 fps than other state-of-the-art trackers. [ABSTRACT FROM AUTHOR]

• Residual attention mechanisms adjust the flow of sentiment information. • The other aspect term module reduces the effect from other aspect terms' sentiment. • Experimental results show our model achieves state-of-the-art performance. Aspect-based sentiment analysis (ABSA) is a fine-grained sentiment analysis task designed to predict the sentiment polarity of each aspect term in a text. Recent research mainly uses neural networks to model text and utilizes attention mechanisms to interact for associate aspect terms and context to obtain more effective feature representation. However, the general attention mechanism is easy to lose the original information. Besides, in the multi-aspect text, the sentiment information of other aspect terms interferes with the sentiment analysis of the current aspect term likely. In this paper, we propose two models named RA-CNN and RAO-CNN for ABSA tasks. In RA-CNN, we apply CNN to model the aspect term and utilize a specially designed residual attention mechanism to interact with the text. Based on the RA-CNN, RAO-CNN adds other aspect terms module, which can reduce interference of sentiment information related to other aspect terms in the multi-aspect text. To verify the proposed models' effectiveness, we conduct a large number of experiments and comparisons on seven public datasets. Experimental results show that our proposed models are useful and achieve state-of-the-art results. [ABSTRACT FROM AUTHOR]

Person re-identification plays an important role in searching for a specific person in a camera network with non-overlapping cameras. The most critical problem for re-identification is feature representation. In this paper, a multi-level cross-view consistent feature learning framework is proposed for person re-identification. First, local deep, LOMO and SIFT features are extracted to form multi-level features. Specifically, local features from the lower and higher layers of a convolutional neural network (CNN) are extracted, these features complement each other as they extract apparent and semantic properties. Second, an ID-based cross-view multi-level dictionary learning (IDB-CMDL) is carried out to obtain sparse and discriminant feature representation. Third, a cross-view consistent word learning is performed to get the cross-view consistent BoVW histograms from sparse feature representation. Finally, a multi-level metric learning fuses multiple BoVW histograms, and learns the sample distance in the subspace for ranking. Experiments on the public CUHK03, Market1501, and DukeMTMC-ReID datasets show results that are superior to many state-of-the-art methods for person re-identification. [ABSTRACT FROM AUTHOR]

• Emotion-related EEG signal segmentation strategy based on facial landmarks. • SNN-based framework for subject-independent short-term emotion recognition from EEG. • SNN hyperparameter optimisation related to spike encoding and output representation. • The emerged SNN connectivity helps in understanding the neural mechanisms of emotion. Emotion recognition still poses a challenge lying at the core of the rapidly growing area of affective computing and is crucial for establishing a successful human–computer interaction. Identification and understanding of emotions are achieved through various measures, such as subjective self-reports, face-tracking, voice analysis, gaze-tracking, as well as the analysis of autonomic and central neurophysiological measurements. Current approaches to emotion recognition based on electroencephalography (EEG) mostly rely on various handcrafted features extracted over relatively long time windows of EEG during participants exposure to appropriate affective stimuli. In this paper, we present a short-term emotion recognition framework based on spiking neural network (SNN) modelling of spatio-temporal EEG patterns. Our method relies on EEG signal segmentation based on detection of short-term changes in facial landmarks, and as such includes no computation of handcrafted EEG features. Differences between participants' EEG properties are taken into account via subject-dependent spike encoding in the formulated subject-independent emotion recognition task. We test our methods on the publicly available DEAP and MAHNOB-HCI databases due to the availability of both EEG and frontal face video data. Through an exhaustive hyperparameter optimisation strategy, we show that the proposed SNN-based representation of EEG spiking patterns provides valuable information for short- term emotion recognition. The obtained accuracies are 78.97% and 79.39% in arousal classification, and 67.76% and 72.12% in valence classification, on the DEAP and MAHNOB-HCI datasets, respectively. Furthermore, through the application of a brain-inspired SNN model, this study provides novel insight and helps in the understanding of the neural mechanisms involved in emotional processing in the context of audiovisual stimuli, such as affective videos. The presented results encourage the use of the proposed EEG processing methodology as a complement to existing features and methods commonly used for EEG-based emotion recognition, especially for short-term arousal recognition. [ABSTRACT FROM AUTHOR]

• Dynamic compositional model extended to consider additional structural tag information. • First to use a hypernetwork only based on part-of-speech tags for dynamic parameter generation. • Extensive experiments show that our proposed approach is effective and efficient. Tree-structured neural networks, such as TreeLSTM and its variants, have proven effective for learning semantic representations of sentences, which are useful for a variety of tasks in natural language processing such as text categorisation, text semantic matching and machine translation. These neural network models take as inputs parse trees of sentences, which are generated by a language parser. However, most existing tree-structured neural network models lack the ability of distinguishing different syntactic compositions, thus the expressive power of these models is limited. Moreover, the syntactic knowledge provided by Part-of-Speech tags in a parse tree has not been fully utilised in existing tree-structured neural network models. It is expected that such syntactic knowledge should help distinguish syntactic compositions, so should result in better semantic representation. This paper proposes a novel neural network model, TagHyperTreeLSTM, which contains two components, a tag-aware hypernetwork and a sentence encoder. The tag-aware hypernetwork, which accepts tags as inputs, generates the parameters of the sentence encoder dynamically in order to distinguish different syntactic compositions. The sentence encoder, which accepts words as inputs, generates the final sentence representation. Experimental results show that the proposed model achieves superior or competitive performance in text classification and text semantic matching based on six benchmark datasets when compared against previous tree-structured models. [ABSTRACT FROM AUTHOR]

Machine learning (ML) employs classification algorithms such as artificial neural networks to make automated decisions. While the proposed solutions have made significant contributions toward improving decision-making efficiency, their effectiveness to serve our multifaceted and complex society has recently come under question. This paper attempts to theorize an answer to why ML lacks effectiveness by studying the assumptions that are made to facilitate efficient pattern recognition. Specifically, this study recognizes five assumptions and investigates their influence on the effectiveness of decision-making for three well-known case studies. The results suggest including the assumptions needed for metric-optimizing supervised learning can only be justifiable and lead to decision-making effectiveness for cases in which a fair and equitable definition of success can be formulated as an objective function. In contrast, the results show using unsupervised learning or non-metric-optimizing supervised learning leads to a more reasonable balance of effectiveness and efficiency when the formulation of a fair and equitable definition of success is not possible. Moreover, the results demonstrate that the current ML approaches that employ supervised learning can improve their efficacy by rethinking the assumptions made at the stage of pattern recognition. [ABSTRACT FROM AUTHOR]

• The different kinds of outliers in drilling process are all detected, which need to be handled differently. • The causes of anomalies are discriminated intelligently and effectively by the DTW-FCM. • While improving the data quality, the data with drilling accident characteristics is reserved. • The performance of PNN-based monitoring models are improved by the discrimination and correction. During the drilling process, the data quality influences the reliability of condition monitoring results. However, the actual drilling data may encounter various kinds of abnormal data, and different kinds of them need to be handled differently. The abnormal data caused by external factors such as sensor failure, storage errors, etc., should be corrected, while the abnormal data caused by drilling accidents should be protected. Therefore, not only the anomaly detection is needed, but the causes of anomalies should be further discriminated. This paper proposes a method for discrimination and correction of abnormal data in the drilling process. First, the local outlier factor anomaly detection algorithm is developed to detect all kinds of the abnormal data. Then, the dynamic time warping and fuzzy c-means are combined for the discrimination of causes of anomalies. Finally, the discriminated abnormal data caused by external factors are corrected with the k nearest neighbor interpolation. Simulation results involving actual data illustrate that the causes of anomalies can be discriminated effectively, and the monitoring results of monitoring models based on neural network improve after using the proposed method, which verify the necessity of anomaly discrimination. [ABSTRACT FROM AUTHOR]

Neural network pruning is a critical field aimed at reducing the infrastructure costs of neural networks by removing parameters. Traditional methods follow a fixed paradigm including pretraining, pruning, and fine-tuning. Despite the close relationship among these three stages, most pruning methods treat them as independent processes. In this paper, we propose a novel two-stage pruning method, which includes pretraining a network that is instructive for subsequent pruning, and a unified optimization model that integrates pruning and fine-tuning. Specifically, in the first stage, we design a group sparse regularized model for pretraining. This model not only safeguards the network from irreversible damage but also offers valuable insights for the pruning process. In the second stage, we introduce an element-wise sparse regularization into pruning model. This model enables us to pinpoint sparse weights more precisely than pretrained network. It automatically derives effective pruning criteria, and omits the step of fine-tuning. To implement the two-stage process in practice, we utilize stochastic gradient algorithm for the pretraining and design a threshold algorithm for pruning stage. Extensive experiments confirm the competitive performance of our proposed method in terms of both accuracy and memory cost when compared to various benchmarks. Furthermore, ablation experiments validate the effectiveness of the proposed pretraining model's guidance for the pruning process. • We propose SOGP to enhance the connection between pretraining and pruning. • We propose a novel non-convex group sparse regularization GT l 1 . • We construct a unified optimization model to integrate pruning and fine-tuning. • Extensive experiments validate the good pruning and accuracy performance of SOGP. [ABSTRACT FROM AUTHOR]

Recently, various normalization layers have been proposed to stabilize the training of deep neural networks. Among them, group normalization is a generalization of layer normalization and instance normalization by allowing a degree of freedom in the number of groups it uses. However, to determine the optimal number of groups, trial-and-error-based hyperparameter tuning is required, and such experiments are time-consuming. In this study, we discuss a reasonable method for setting the number of groups. First, we find that the number of groups influences the gradient behavior of the group normalization layer. Based on this observation, we derive the ideal number of groups, which calibrates the gradient scale to facilitate gradient descent optimization. This paper is the first to propose an optimal number of groups that is theoretically grounded, architecture-aware, and can provide a proper value in a layer-wise manner for all layers. The proposed method exhibited improved performance over existing methods in numerous neural network architectures, tasks, and datasets. • We propose a method to determine the number of groups of group normalization. • A theoretical analysis of group normalization with activation function is provided. • The proposed method is validated against various practical problems. [ABSTRACT FROM AUTHOR]

Model pruning is a commonly used technique for compressing DNNs and reducing computation requirements to accelerate inference. However, the required granularity of pruning varies across different application scenarios, making it difficult and cumbersome to customize different pruning methods for each hardware or computing platform. Therefore, a unified framework is necessary to accommodate various levels of granularity of pruning. Furthermore, some available methods require additional fine-tuning or model retraining to restore accuracy, which can result in significant time costs. With this motivation, this paper proposes a Multi-Granularity Pruning Framework, namely MGPF, to obtain sparse models of different granularity without fine-tuning the remaining connections. Specifically, a deformable soft mask is introduced as the pruning initiator to achieve different levels of pruning granularity, such as weight pruning, channel pruning, and filter pruning, etc. The model parameters and soft masks are jointly trained, and we just apply L 1 regularization on soft masks for sparsity to ensure that the model can be repaired during training without fine-tuning or retraining. After pruning, the soft masks are absorbed into the model parameters in the form of element product without changing the model structures. Experimental results on three image classification benchmarks CIFAR-10/100 and ImageNet-1K demonstrate the effectiveness of our method for various CNN architectures, datasets, and pruning rates. Particularly, for ResNet-50 on ImageNet-1K, we achieve a higher accuracy under the pruning rate of 98% for unstructured pruning which leads the advanced method by 12%. [ABSTRACT FROM AUTHOR]

Traditional knowledge distillation approaches are typically designed for specific tasks, as they primarily distilling deep features from intermediate layers of a neural network, generally with ingeniously designed knowledge representations, which increase the difficulty of model development and interpretation. In contrast, we empirically show that a soft logits knowledge distillation is enough to significantly narrow the teacher–student model performance gap. In this paper, the S oft H ybrid K nowledge D istillation (SHKD) is proposed to generate representative features from the input and output location of models, and applicable to various teacher and student architectures. Specifically, the additional mixup task is first applied to the same batch of training images, which enriches the high-order feature information transferred by the model. And then, a top-k guided specification filter is added to the teacher model, to reduce the bias induced throughout the knowledge transfer process. Finally, the SHKD utilizes the same and different teacher–student model pairs of the deep neural network, which can be easily adjusted to the computational resources available in the current world. Extensive experiments are conducted to examine the efficiency, and show that the SHKD outperforms state-of-the-art knowledge distillation methods. With the SHKD, the high performance of models will be maintained, and the gap between the teacher model and the student model can be reduced, making the construction of deep neural networks more convenient to the unknown scenarios. The source code is available at https://github.com/lambett/SHKD. [ABSTRACT FROM AUTHOR]

Inference centers need more data to have a more comprehensive and beneficial learning model, and for this purpose, they need to collect data from data providers. On the other hand, data providers are cautious about delivering their datasets to inference centers in terms of privacy considerations. In this paper, by modifying the structure of the autoencoder, we present a method that manages the utility-privacy trade-off well. To be more precise, the data is first compressed using the encoder, then confidential and non-confidential features are separated and uncorrelated using the classifier. The confidential feature is appropriately combined with noise, and the non-confidential feature is enhanced, and at the end, data with the original data format is produced by the decoder. The suggested architecture additionally enables data providers to modify the degree of privacy needed for private features and the level of utility for non-private features. The proposed method has been examined for both image and categorical databases, and the results show a significant performance improvement compared to previous methods. [ABSTRACT FROM AUTHOR]

As the deep neural networks (DNNs) with satisfied expression ability usually require large scale to gain adequate performance, and deploying large-scale DNNs on resource-limited environments is still a challenge, neural network compression becomes a hot topic nowadays. Among the multiple compression methods, tensor decomposition reveals many specific advantages, such as regular data structure comes from linear algebra, convenient approach of training from scratch, ideal compression ratio, etc. Nevertheless, for some compact neural modules such as two-dimensional/three-dimensional (2D/3D) convolutional kernels, traditional tensor decomposition in the way of approximation has encountered intractable obstacles. Fortunately, some works utilize the tensor mapping approach, which just regards the data structure of tensor decomposition as neural layers, to reconstruct the convolutional kernel into a new lightweight module with several thinner convolutional kernels. The only two flies in the ointment are there lack necessary theories of tensor mapping, and there is still no tensor mapping way to compress high-order three-dimensional convolutional neural networks (3DCNNs). In this paper, we first deeply analyze the tensor mapping theory including convergence and precision, which separately establishes the rationality of tensor mapping and its superiority over the traditional tensor approximation, according to the Lottery Ticket Hypothesis. Then we propose an efficient method termed as 3D-KCPNet, to compress 3DCNNs based on the novel Kronecker canonical polyadic (KCP) tensor decomposition. The proposed method can not only compress the 3D convolutional kernels, but also convert a 3D convolution to efficient 1 × 1 × 1 and 2D depthwise convolutions. The experiments on the video recognition datasets VIVA Challenge, UCF11, UCF50, and UCF101 show that the accuracy of 3D-KCPNet can surpass its original baseline model and the corresponding tensor approximation model. [ABSTRACT FROM AUTHOR]

This paper, proposes a novel artificial neural network, called self-adjusting feature map (SAM), and develop its unsupervised learning ability with self-adjusting mechanism. The trained network structure of representative connected neurons not only displays the spatial relation of the input data distribution but also quantizes the data well. The SAM can automatically isolate a set of connected neurons, in which, the used number of the sets may indicate the number of clusters. The idea of self-adjusting mechanism is based on combining of mathematical statistics and neurological advantages and retreat of waste. In the training process, for each representative neuron has are three phases, growth, adaptation and decline. The network of representative neurons, first create the necessary neurons according to the local density of the input data in the growth phase. In the adaption phase, it adjusts neighborhood neuron pair׳s connected/disconnected topology constantly according to the statistics of input feature data. Finally, the unnecessary neurons of the network are merged or remove in the decline phase. In this paper, we exploit the SAM to handle some peculiar cases that cannot be handled easily by classical unsupervised learning networks such as self-organizing map (SOM) network. The remarkable characteristics of the SAM can be seen on various real world cases in the experimental results. [ABSTRACT FROM AUTHOR]

This paper is concerned with the problem of μ -stability analysis of memristor-based recurrent neural networks with the effects of bounded and unbounded leakage time-varying delays. A new idea of μ -stability analysis of neural networks is given first, then by means of the linear matrix inequality (LMI) approach, stability criteria are presented. Two types of memductance functions is used to derive the proposed stability results. Obviously, the memristive neural network with different memductance functions is a state-dependent switched system or a state-dependent continuous system, which is the generalization of those for conventional artificial neural networks. Under the framework of Filippov solutions, μ -stability analysis of solutions for functional differential inclusions and memristor-based neural networks can be guaranteed by constructing suitable Lyapunov–Krasovskii functionals, suitable inequalities and LMIs. The dynamic analysis in this paper utilizes the theory of set-valued maps and functional differential equations with discontinuous right-hand sides of Filippov type. Leakage time-varying delay is considered in this paper to be bounded, unbounded and differentiable. Taking into account of the information of the neuron activation functions and unbounded time-varying delays, several improved results have been obtained in terms of LMIs and these are tested by MATLAB LMI toolbox. The model based on the memristor widens the relevance scope for the design of neural networks, and the new effective results also enrich the toolbox for the qualitative analysis of neural networks. Finally, numerical examples are provided to demonstrate the effectiveness of the proposed theoretical results. [ABSTRACT FROM AUTHOR]

This study is focused on the issue of adaptive exponential synchronization for a general class of memristive neural networks (MNNs) with mixed time-varying delays. A new and simple adaptive controller with feedback control law is designed to achieve exponential synchronization by using Lyapunov functional method. The adaptive controller proposed in the paper possesses a powerful adaptive capability that it can be utilized for various MNNs with different mathematical definitions of memristor. In addition, no excessive calculations such as solving linear matrix inequality or computing algebraic conditions are required in our synchronization criteria. We also present two synchronization conditions for a special class of MNNs that part or even all of the system׳s right-hand is reduced to be continuous when activation functions are zero at the neuron׳s switching points. And two lemmas are introduced to modify a misunderstanding in this situation in some previous papers. Finally, an example with numerical simulations is presented to illustrate the efficiency and accuracy of the theoretical results. [ABSTRACT FROM AUTHOR]

This paper deals with the problem of the global exponential stability of a class of uncertain neural networks with discontinuous Lurie-type activation and mixed delays. By establishing a new sufficient condition, we first prove the existence of the equilibrium point by using the Leray–Schauder alternative theorem. Then, by employing a new Lyapunov functional, we obtain the global exponential stability of the equilibrium point of the uncertain neural network. In the end, some comparisons and numerical examples are given to show the improvement of the conclusions in this paper. [ABSTRACT FROM AUTHOR]

In this paper, we propose a generalized regression neural networks (GRNNS) with K -fold cross-validation (GRNNSK) method for predicting the displacement of landslide. Furthermore, correlation analysis is used to find the potential input variables for this predicting model, such as Pearson cross-correlation coefficients (PCC) and mutual information (MI) are applied in this paper. Tests on two case studies of Liangshuijing (LSJ) and Baishuihe (BSH) landslide in the Three Gorges reservoir area of China demonstrate the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

This paper deals with the design of an adaptive observer based both on a Multi-Layer Perceptron Neural Network (MLPNN) and a sliding mode technique. Its main objective is to construct the complete state of a given exoskeleton worn by a human subject. The observer we propose in this paper can be used for any application: rehabilitation, assistance, etc. The dynamic model of the global system composed of the exoskeleton and the human is complex and supposed completely unknown. The MLPNN chosen for its characteristic of universal approximation has been used here to identify the unknown dynamic. Its parameters have been adjusted by taking into account the structure of the dynamic model of the considered system and the closed-loop stability based on Lyapunov׳s approach. A Taylor series expansion allows resolving the non-linearity problem present in the MLPNN. Besides the fact that the proposed adaptive observer can be integrated in a control scheme, it also allows us to study the behavior of the exoskeleton before any application on the human subject. The proposed study has been validated both in simulation and in experimentation. The obtained results show the effectiveness of the proposed adaptive approach. [ABSTRACT FROM AUTHOR]

This paper is concerned with the problem of passivity analysis of neural networks with an interval time-varying delay. Unlike existing results in the literature, the time-delay considered in this paper is subjected to interval time-varying without any restriction on the rate of change. Based on novel refined Jensen inequalities and by constructing an improved Lyapunov–Krasovskii functional (LKF), which fully utilizes information of the neuron activation functions, new delay-dependent conditions that ensure the passivity of the network are derived in terms of tractable linear matrix inequalities (LMIs) which can be effectively solved by various computational tools. The effectiveness and improvement over existing results of the proposed method in this paper are illustrated through numerical examples. [ABSTRACT FROM AUTHOR]

In this paper, the H ∞ filtering problem is investigated for a class of discrete-time arbitrary switched neural networks with missing measurements, stochastic perturbations, and communication delays. Based on the average dwell time approach and a set of Kronecker delta functions, a unified measurement model is established to represent the phenomena of missing measurements, time delays and nonlinearities. The aim of this paper is to design an H ∞ filter such that the filter error dynamics is exponentially mean-square stable and the H ∞ performance requirement is satisfied simultaneously. By using the Lyapunov stability theory and the matrix technology, the design method of the desired filter is given in terms of a matrix inequality which can be solved by using the available software. Finally, a numerical example is provided to illustrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

The passivity problem has been further researched for a class of memristive neural networks with probabilistic time-varying delays in this paper. Based on an effective Lyapunov functional and the Wirtinger-type inequality, sufficient conditions are presented to guarantee the passive performance of the memristive models. By establishing a stochastic variable with Bernoulli distribution, the information of probabilistic time-varying delays are considered, which were transformed into one with deterministic time-varying delay and stochastic parameters. Moreover, the range of the delays as well as the probability distribution of its variation are all taken into consideration, thus, the results derived in this paper are more reasonable. Finally, the advantages of the proposed techniques are demonstrated by two numerical examples. [ABSTRACT FROM AUTHOR]

In this paper a people recognition algorithm is proposed, in which the color processing mechanism is inspired by the biological visual system. The algorithm is constructed in two parts. In the first part a spiking neural network is proposed to extract the color features of the people images which are captured from videos, and a set of new features is generated by fusing the color features and color moments. In the second part, after a feature reduction, a Support Vector Machine is trained and then used to recognize a specific people. The algorithm has been successfully applied to recognize people in CASIA Database with a high recognition rate. In order to evaluate performance and analyze characteristics of people recognition algorithms in multi-camera scenes, Multi-Camera Video (MCV) dataset is made in this paper. It is used to evaluate and analyze the proposed method and a set of characteristics of the proposed algorithms are obtained. Experimental results demonstrate that the algorithm is comparable to state-of-the-art approaches in terms of accuracy and the direction for further improvement of the proposed algorithm is provided. [ABSTRACT FROM AUTHOR]