Search

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

These instructions give guidelines for preparing papers for this publication. Presents information for authors publishing in this journal. [ABSTRACT FROM AUTHOR]

In this paper, an adaptive admittance control scheme is developed for robots to interact with time-varying environments. Admittance control is adopted to achieve a compliant physical robot–environment interaction, and the uncertain environment with time-varying dynamics is defined as a linear system. A critic learning method is used to obtain the desired admittance parameters based on the cost function composed of interaction force and trajectory tracking without the knowledge of the environmental dynamics. To deal with dynamic uncertainties in the control system, a neural-network (NN)-based adaptive controller with a dynamic learning framework is developed to guarantee the trajectory tracking performance. Experiments are conducted and the results have verified the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Dropout and DropConnect are two techniques to facilitate the regularization of neural network models, having achieved the state-of-the-art results in several benchmarks. In this paper, to improve the generalization capability of spiking neural networks (SNNs), the two drop techniques are first applied to the state-of-the-art SpikeProp learning algorithm resulting in two improved learning algorithms called SPDO (SpikeProp with Dropout) and SPDC (SpikeProp with DropConnect). In view that a higher membrane potential of a biological neuron implies a higher probability of neural activation, three adaptive drop algorithms, SpikeProp with Adaptive Dropout (SPADO), SpikeProp with Adaptive DropConnect (SPADC), and SpikeProp with Group Adaptive Drop (SPGAD), are proposed by adaptively adjusting the keep probability for training SNNs. A convergence theorem for SPDC is proven under the assumptions of the bounded norm of connection weights and a finite number of equilibria. In addition, the five proposed algorithms are carried out in a collaborative neurodynamic optimization framework to improve the learning performance of SNNs. The experimental results on the four benchmark data sets demonstrate that the three adaptive algorithms converge faster than SpikeProp, SPDO, and SPDC, and the generalization errors of the five proposed algorithms are significantly smaller than that of SpikeProp. Furthermore, the experimental results also show that the five algorithms based on collaborative neurodynamic optimization can be improved in terms of several measures. [ABSTRACT FROM AUTHOR]

We review the current literature concerned with information plane (IP) analyses of neural network (NN) classifiers. While the underlying information bottleneck theory and the claim that information-theoretic compression is causally linked to generalization are plausible, empirical evidence was found to be both supporting and conflicting. We review this evidence together with a detailed analysis of how the respective information quantities were estimated. Our survey suggests that compression visualized in IPs is not necessarily information-theoretic but is rather often compatible with geometric compression of the latent representations. This insight gives the IP a renewed justification. Aside from this, we shed light on the problem of estimating mutual information in deterministic NNs and its consequences. Specifically, we argue that, even in feedforward NNs, the data processing inequality needs not to hold for estimates of mutual information. Similarly, while a fitting phase, in which the mutual information is between the latent representation and the target increases, is necessary (but not sufficient) for good classification performance, depending on the specifics of mutual information estimation, such a fitting phase needs to not be visible in the IP. [ABSTRACT FROM AUTHOR]

The well-known backpropagation learning algorithm is probably the most popular learning algorithm in artificial neural networks. It has been widely used in various applications of deep learning. The backpropagation algorithm requires a separate feedback network to back propagate errors. This feedback network must have the same topology and connection strengths (weights) as the feed-forward network. In this article, we propose a new learning algorithm that is mathematically equivalent to the backpropagation algorithm but does not require a feedback network. The elimination of the feedback network makes the implementation of the new algorithm much simpler. The elimination of the feedback network also significantly increases biological plausibility for biological neural networks to learn using the new algorithm by means of some retrograde regulatory mechanisms that may exist in neurons. This new algorithm also eliminates the need for two-phase adaptation (feed-forward phase and feedback phase). Hence, neurons can adapt asynchronously and concurrently in a way analogous to that of biological neurons. [ABSTRACT FROM AUTHOR]

Scene graph generation (SGGen) is a challenging task due to a complex visual context of an image. Intuitively, the human visual system can volitionally focus on attended regions by salient stimuli associated with visual cues. For example, to infer the relationship between man and horse, the interaction between human leg and horseback can provide strong visual evidence to predict the predicate ride. Besides, the attended region face can also help to determine the object man. Till now, most of the existing works studied the SGGen by extracting coarse-grained bounding box features while understanding fine-grained visual regions received limited attention. To mitigate the drawback, this article proposes a region-aware attention learning method. The key idea is to explicitly construct the attention space to explore salient regions with the object and predicate inferences. First, we extract a set of regions in an image with the standard detection pipeline. Each region regresses to an object. Second, we propose the object-wise attention graph neural network (GNN), which incorporates attention modules into the graph structure to discover attended regions for object inference. Third, we build the predicate-wise co-attention GNN to jointly highlight subject’s and object’s attended regions for predicate inference. Particularly, each subject-object pair is connected with one of the latent predicates to construct one triplet. The proposed intra-triplet and inter-triplet learning mechanism can help discover the pair-wise attended regions to infer predicates. Extensive experiments on two popular benchmarks demonstrate the superiority of the proposed method. Additional ablation studies and visualization further validate its effectiveness. [ABSTRACT FROM AUTHOR]

Modulation classification is one of the key tasks for communications systems monitoring, management, and control for addressing technical issues, including spectrum awareness, adaptive transmissions, and interference avoidance. Recently, deep learning (DL)-based modulation classification has attracted significant attention due to its superiority in feature extraction and classification accuracy. In DL-based modulation classification, one major challenge is to preprocess a received signal and represent it in a proper format before feeding the signal into deep neural networks. This article provides a comprehensive survey of the state-of-the-art DL-based modulation classification algorithms, especially the techniques of signal representation and data preprocessing utilized in these algorithms. Since a received signal can be represented by either features, images, sequences, or a combination of them, existing algorithms of DL-based modulation classification can be categorized into four groups and are reviewed accordingly in this article. Furthermore, the advantages as well as disadvantages of each signal representation method are summarized and discussed. [ABSTRACT FROM AUTHOR]

In this article, the finite-time synchronization (FTSYN) of a class of quaternion-valued neural networks (QVNNs) with discrete and distributed time delays is studied. Furthermore, the FTSYN and fixed-time synchronization (FIXSYN) of the QVNNs without time delay are investigated. Different from the existing results, which used decomposition techniques, by introducing an improved one-norm, we use a direct analytical method to study the synchronization problems. Incidentally, several properties of one-norm of the quaternion are analyzed, and then, three effective controllers are proposed to synchronize the drive and response QVNNs within a finite time or fixed time. Moreover, efficient criteria are proposed to guarantee that the synchronization of QVNNs with or without mixed time delays can be realized within a finite and fixed time interval, respectively. In addition, the settling times are reckoned. Compared with the existing work, our advantages are mainly reflected in the simpler Lyapunov analytical process and more general activation function. Finally, the validity and practicability of the conclusions are illustrated via four numerical examples. [ABSTRACT FROM AUTHOR]

Neuromorphic computing is a promising technology that realizes computation based on event-based spiking neural networks (SNNs). However, fault-tolerant on-chip learning remains a challenge in neuromorphic systems. This study presents the first scalable neuromorphic fault-tolerant context-dependent learning (FCL) hardware framework. We show how this system can learn associations between stimulation and response in two context-dependent learning tasks from experimental neuroscience, despite possible faults in the hardware nodes. Furthermore, we demonstrate how our novel fault-tolerant neuromorphic spike routing scheme can avoid multiple fault nodes successfully and can enhance the maximum throughput of the neuromorphic network by 0.9%–16.1% in comparison with previous studies. By utilizing the real-time computational capabilities and multiple-fault-tolerant property of the proposed system, the neuronal mechanisms underlying the spiking activities of neuromorphic networks can be readily explored. In addition, the proposed system can be applied in real-time learning and decision-making applications, brain–machine integration, and the investigation of brain cognition during learning. [ABSTRACT FROM AUTHOR]

In traditional graph neural networks (GNNs), graph convolutional learning is carried out through topology-driven recursive node content aggregation for network representation learning. In reality, network topology and node content each provide unique and important information, and they are not always consistent because of noise, irrelevance, or missing links between nodes. A pure topology-driven feature aggregation approach between unaligned neighborhoods may deteriorate learning from nodes with poor structure-content consistency, due to the propagation of incorrect messages over the whole network. Alternatively, in this brief, we advocate a co-alignment graph convolutional learning (CoGL) paradigm, by aligning topology and content networks to maximize consistency. Our theme is to enforce the learning from the topology network to be consistent with the content network while simultaneously optimizing the content network to comply with the topology for optimized representation learning. Given a network, CoGL first reconstructs a content network from node features then co-aligns the content network and the original network through a unified optimization goal with: 1) minimized content loss; 2) minimized classification loss; and 3) minimized adversarial loss. Experiments on six benchmarks demonstrate that CoGL achieves comparable and even better performance compared with existing state-of-the-art GNN models. [ABSTRACT FROM AUTHOR]

A convolutional neural network (CNN) is one of the most significant networks in the deep learning field. Since CNN made impressive achievements in many areas, including but not limited to computer vision and natural language processing, it attracted much attention from both industry and academia in the past few years. The existing reviews mainly focus on CNN’s applications in different scenarios without considering CNN from a general perspective, and some novel ideas proposed recently are not covered. In this review, we aim to provide some novel ideas and prospects in this fast-growing field. Besides, not only 2-D convolution but also 1-D and multidimensional ones are involved. First, this review introduces the history of CNN. Second, we provide an overview of various convolutions. Third, some classic and advanced CNN models are introduced; especially those key points making them reach state-of-the-art results. Fourth, through experimental analysis, we draw some conclusions and provide several rules of thumb for functions and hyperparameter selection. Fifth, the applications of 1-D, 2-D, and multidimensional convolution are covered. Finally, some open issues and promising directions for CNN are discussed as guidelines for future work. [ABSTRACT FROM AUTHOR]

In this brief, we consider the problem of descriptors construction for the task of content-based image retrieval using deep neural networks. The idea of neural codes, based on fully connected layers’ activations, is extended by incorporating the information contained in convolutional layers. It is known that the total number of neurons in the convolutional part of the network is large and the majority of them have little influence on the final classification decision. Therefore, in this brief, we propose a novel algorithm that allows us to extract the most significant neuron activations and utilize this information to construct effective descriptors. The descriptors consisting of values taken from both the fully connected and convolutional layers perfectly represent the whole image content. The images retrieved using these descriptors match semantically very well to the query image, and also, they are similar in other secondary image characteristics, such as background, textures, or color distribution. These features of the proposed descriptors are verified experimentally based on the IMAGENET1M dataset using the VGG16 neural network. For comparison, we also test the proposed approach on the ResNet50 network. [ABSTRACT FROM AUTHOR]

Image smoothing is a prerequisite for many computer vision and graphics applications. In this article, we raise an intriguing question whether a dataset that semantically describes meaningful structures and unimportant details can facilitate a deep learning model to smooth complex natural images. To answer it, we generate ground-truth labels from easy samples by candidate generation and a screening test and synthesize hard samples in structure-preserving smoothing by blending intricate and multifarious details with the labels. To take full advantage of this dataset, we present a joint edge detection and structure-preserving image smoothing neural network (JESS-Net). Moreover, we propose the distinctive total variation loss as prior knowledge to narrow the gap between synthetic and real data. Experiments on different datasets and real images show clear improvements of our method over the state of the arts in terms of both the image cleanness and structure-preserving ability. Code and dataset are available at https://github.com/YidFeng/Easy2Hard. [ABSTRACT FROM AUTHOR]

Few-shot learning, aiming to learn novel concepts from one or a few labeled examples, is an interesting and very challenging problem with many practical advantages. Existing few-shot methods usually utilize data of the same classes to train the feature embedding module and in a row, which is unable to learn adapting to new tasks. Besides, traditional few-shot models fail to take advantage of the valuable relations of the support-query pairs, leading to performance degradation. In this article, we propose a transductive relation-propagation graph neural network (GNN) with a decoupling training strategy (TRPN-D) to explicitly model and propagate such relations across support-query pairs, and empower the few-shot module the ability of transferring past knowledge to new tasks via the decoupling training. Our few-shot module, namely TRPN, treats the relation of each support-query pair as a graph node, named relational node, and resorts to the known relations between support samples, including both intraclass commonality and interclass uniqueness. Through relation propagation, the model could generate the discriminative relation embeddings for support-query pairs. To the best of our knowledge, this is the first work that decouples the training of the embedding network and the few-shot graph module with different tasks, which might offer a new way to solve the few-shot learning problem. Extensive experiments conducted on several benchmark datasets demonstrate that our method can significantly outperform a variety of state-of-the-art few-shot learning methods. [ABSTRACT FROM AUTHOR]

In reinforcement learning (RL), function approximation errors are known to easily lead to the $Q$ -value overestimations, thus greatly reducing policy performance. This article presents a distributional soft actor–critic (DSAC) algorithm, which is an off-policy RL method for continuous control setting, to improve the policy performance by mitigating $Q$ -value overestimations. We first discover in theory that learning a distribution function of state–action returns can effectively mitigate $Q$ -value overestimations because it is capable of adaptively adjusting the update step size of the $Q$ -value function. Then, a distributional soft policy iteration (DSPI) framework is developed by embedding the return distribution function into maximum entropy RL. Finally, we present a deep off-policy actor–critic variant of DSPI, called DSAC, which directly learns a continuous return distribution by keeping the variance of the state–action returns within a reasonable range to address exploding and vanishing gradient problems. We evaluate DSAC on the suite of MuJoCo continuous control tasks, achieving the state-of-the-art performance. [ABSTRACT FROM AUTHOR]

The staggering innovations and emergence of numerous deep learning (DL) applications have forced researchers to reconsider hardware architecture to accommodate fast and efficient application-specific computations. Applications, such as object detection, image recognition, speech translation, as well as music synthesis and image generation, can be performed with high accuracy at the expense of substantial computational resources using DL. Furthermore, the desire to adopt Industry 4.0 and smart technologies within the Internet of Things infrastructure has initiated several studies to enable on-chip DL capabilities for resource-constrained devices. Specialized DL processors reduce dependence on cloud servers, improve privacy, lessen latency, and mitigate bandwidth congestion. As we reach the limits of shrinking transistors, researchers are exploring various application-specific hardware architectures to meet the performance and efficiency requirements for DL tasks. Over the past few years, several software optimizations and hardware innovations have been proposed to efficiently perform these computations. In this article, we review several DL accelerators, as well as technologies with emerging devices, to highlight their architectural features in application-specific integrated circuit (IC) and field-programmable gate array (FPGA) platforms. Finally, the design considerations for DL hardware in portable applications have been discussed, along with some deductions about the future trends and potential research directions to innovate DL accelerator architectures further. By compiling this review, we expect to help aspiring researchers widen their knowledge in custom hardware architectures for DL. [ABSTRACT FROM AUTHOR]

This study proposes the time-/event-triggered adaptive neural control strategies for the asymptotic tracking problem of a class of uncertain nonlinear systems with full-state constraints. First, we design a time-triggered strategy. The effect caused by the residuals of the estimation via radial basis function (RBF) neural networks (NNs), and the reasonable upper bounds on the first derivative of the reference signal and the derivative of each virtual control, can be eliminated by designing appropriate adaptive laws and utilizing the basic properties of RBF NNs. Moreover, the construction of the barrier Lyapunov functions (BLFs) in this work ensures the compliance of the full-state constraints and also holds the asymptotic output tracking performance. Then, based on the time-triggered strategy, we further design a relative threshold event-triggered strategy. The proposed event-triggered adaptive neural controller can solve the main control objective of this work, that is: 1) the full-state constraint requirements of the system are not violated and 2) the output signal asymptotically tracks the reference signal. Compared with the traditional method, the event-triggered strategy can improve the utilization of communication channels and resources and has greater practical significance. Finally, an example of single-link robot under the proposed two strategies illustrates the validity of the constructed controllers. [ABSTRACT FROM AUTHOR]

A novel robust adaptive neural network (NN) control scheme with prescribed performance is developed for the 3-D trajectory tracking of underactuated autonomous underwater vehicles (AUVs) with uncertain dynamics and unknown disturbances using new prescribed performance functions, an additional term, the radial basis function (RBF) NN, and the command-filtered backstepping approach. Different from the traditional prescribed performance functions, the new prescribed performance functions are innovatively proposed such that the time desired for the trajectory tracking errors of AUVs to reach and stay within the prescribed error tolerance band can be preset exactly and flexibly. The additional term with the Nussbaum function is designed to deal with the underactuation problem of AUVs. By means of RBF NN, the uncertain item lumped by the uncertain dynamics of AUVs and unknown disturbances is eventually transformed into a linearly parametric form with only a single unknown parameter. The developed control scheme ensures that all signals in the AUV 3-D trajectory tracking closed-loop control system are bounded. Simulation results with comparisons show the validity and the superiority of our developed control scheme. [ABSTRACT FROM AUTHOR]

The adaptive hinging hyperplane (AHH) model is a popular piecewise linear representation with a generalized tree structure and has been successfully applied in dynamic system identification. In this article, we aim to construct the deep AHH (DAHH) model to extend and generalize the networking of AHH model for high-dimensional problems. The network structure of DAHH is determined through a forward growth, in which the activity ratio is introduced to select effective neurons and no connecting weights are involved between the layers. Then, all neurons in the DAHH network can be flexibly connected to the output in a skip-layer format, and only the corresponding weights are the parameters to optimize. With such a network framework, the backpropagation algorithm can be implemented in DAHH to efficiently tackle large-scale problems and the gradient vanishing problem is not encountered in the training of DAHH. In fact, the optimization problem of DAHH can maintain convexity with convex loss in the output layer, which brings natural advantages in optimization. Different from the existing neural networks, DAHH is easier to interpret, where neurons are connected sparsely and analysis of variance (ANOVA) decomposition can be applied, facilitating to revealing the interactions between variables. A theoretical analysis toward universal approximation ability and explicit domain partitions are also derived. Numerical experiments verify the effectiveness of the proposed DAHH. [ABSTRACT FROM AUTHOR]

Deep learning has become the most powerful machine learning tool in the last decade. However, how to efficiently train deep neural networks remains to be thoroughly solved. The widely used minibatch stochastic gradient descent (SGD) still needs to be accelerated. As a promising tool to better understand the learning dynamic of minibatch SGD, the information bottleneck (IB) theory claims that the optimization process consists of an initial fitting phase and the following compression phase. Based on this principle, we further study typicality sampling, an efficient data selection method, and propose a new explanation of how it helps accelerate the training process of the deep networks. We show that the fitting phase depicted in the IB theory will be boosted with a high signal-to-noise ratio of gradient approximation if the typicality sampling is appropriately adopted. Furthermore, this finding also implies that the prior information of the training set is critical to the optimization process, and the better use of the most important data can help the information flow through the bottleneck faster. Both theoretical analysis and experimental results on synthetic and real-world datasets demonstrate our conclusions. [ABSTRACT FROM AUTHOR]

The neural-network (NN)-based control method is a new emerging promising technique for controller design in a power electronic circuit (PEC). However, the optimization of NN-based controllers (NNCs) has significant challenges in two aspects. The first challenge is that the search space of the NNC optimization problem is such complex that the global optimization ability of the existing algorithms still needs to be improved. The second challenge is that the training process of the NNC parameters is very computationally expensive and requires a long execution time. Thus, in this article, we develop a powerful evolutionary computation-based algorithm to find a high-quality solution and reduce computational time. First, the differential evolution (DE) algorithm is adopted because it is a powerful global optimizer in solving a complex optimization problem. This can help to overcome the premature convergence in local optima to train the NNC parameters well. Second, to reduce the computational time, the DE is extended to distribute DE (DDE) by dispatching all the individuals to different distributed computing resources for parallel computing. Moreover, a resource-aware strategy (RAS) is designed to further efficiently utilize the resources by adaptively dispatching individuals to resources according to the real-time performance of the resources, which can simultaneously concern the computing ability and load state of each resource. Experimental results show that, compared with some other typical evolutionary algorithms, the proposed algorithm can get significantly better solutions within a shorter computational time. [ABSTRACT FROM AUTHOR]

Although deep neural networks have been proved effective in many applications, they are data hungry, and training deep models often requires laboriously labeled data. However, when labeled data contain erroneous labels, they often lead to model performance degradation. A common solution is to assign each sample with a dynamic weight during optimization, and the weight is adjusted in accordance with the loss. However, those weights are usually unreliable since they are measured by the losses of corrupted labels. Thus, this scheme might impede the discriminative ability of neural networks trained on noisy data. To address this issue, we propose a novel reweighting method, dubbed self-reweighting from class centroids (SRCC), by assigning sample weights based on the similarities between the samples and our online learned class centroids. Since we exploit statistical class centers in the image feature space to reweight data samples in learning, our method is robust to noise caused by corrupted labels. In addition, even after reweighting the noisy data, the decision boundaries might still suffer distortions. Thus, we leverage mixed inputs that are generated by linearly interpolating two random images and their labels to further regularize the boundaries. We employ the learned class centroids to evaluate the confidence of our generated mixed data via measuring feature similarities. During the network optimization, the class centroids are updated as more discriminative feature representations of original images are learned. In doing so, SRCC will generate more robust weighting coefficients for noisy and mixed data and facilitates our feature representation learning in return. Extensive experiments on both the synthetic and real image recognition tasks demonstrate that our method SRCC outperforms the state of the art on learning with noisy data. [ABSTRACT FROM AUTHOR]

Cluster synchronization means that all multiagents are divided into different clusters according to the equations or roles of nodes in a complex network, and by designing an appropriate algorithm, each cluster can achieve synchronization to a certain value or an isolated node. However, the synchronization values between different clusters are different. With a feedback controller based on the calculation of the control input value and a trigger condition leading to the updating instants, this article introduces the trigger mechanism and designs a new data sampling strategy to achieve cluster synchronization of the coupled neural networks (CNNs), which reduces the number of updates of the controller, thereby reducing unnecessary waste of limited resources. In addition, an example proposes a synchronization algorithm and gives iterative procedures to calculate the trigger instants and prove the validity of the theoretical results. [ABSTRACT FROM AUTHOR]

With the aid of neural networks, this article develops two data-driven designs of fault detection (FD) for dynamic systems. The first neural network is constructed for generating residual signals in the so-called finite impulse response (FIR) filter-based form, and the second one is designed for recursively generating residual signals. By theoretical analysis, we show that two proposed neural networks via self-organizing learning can find their optimal architectures, respectively, corresponding to FIR filter and recursive observer for FD purposes. Additional contributions of this study lie in that we establish bridges that link model- and neural-network-based methods for detecting faults in dynamic systems. An experiment on a three-tank system is adopted to illustrate the effectiveness of two proposed neural network-aided FD algorithms. [ABSTRACT FROM AUTHOR]

Spike-timing-dependent plasticity (STDP) is one of the most popular and deeply biologically motivated forms of unsupervised Hebbian-type learning. In this article, we propose a variant of STDP extended by an additional activation-dependent scale factor. The consequent learning rule is an efficient algorithm, which is simple to implement and applicable to spiking neural networks (SNNs). It is demonstrated that the proposed plasticity mechanism combined with competitive learning can serve as an effective mechanism for the unsupervised development of receptive fields (RFs). Furthermore, the relationship between synaptic scaling and lateral inhibition is explored in the context of the successful development of RFs. Specifically, we demonstrate that maintaining a high level of synaptic scaling followed by its rapid increase is crucial for the development of neuronal mechanisms of selectivity. The strength of the proposed solution is assessed in classification tasks performed on the Modified National Institute of Standards and Technology (MNIST) data set with an accuracy level of 94.65% (a single network) and 95.17% (a network committee)—comparable to the state-of-the-art results of single-layer SNN architectures trained in an unsupervised manner. Furthermore, the training process leads to sparse data representation and the developed RFs have the potential to serve as local feature detectors in multilayered spiking networks. We also prove theoretically that when applied to linear Poisson neurons, our rule conserves total synaptic strength, guaranteeing the convergence of the learning process. [ABSTRACT FROM AUTHOR]

This article is devoted to investigating finite-time synchronization (FTS) for coupled neural networks (CNNs) with time-varying delays and Markovian jumping topologies by using an intermittent quantized controller. Due to the intermittent property, it is very hard to surmount the effects of time delays and ascertain the settling time. A new lemma with novel finite-time stability inequality is developed first. Then, by constructing a new Lyapunov functional and utilizing linear programming (LP) method, several sufficient conditions are obtained to assure that the Markovian CNNs achieve synchronization with an isolated node in a settling time that relies on the initial values of considered systems, the width of control and rest intervals, and the time delays. The control gains are designed by solving the LP. Moreover, an optimal algorithm is given to enhance the accuracy in estimating the settling time. Finally, a numerical example is provided to show the merits and correctness of the theoretical analysis. [ABSTRACT FROM AUTHOR]

Efficient modeling of feature interactions underpins supervised learning for nonsequential tasks, characterized by a lack of inherent ordering of features (variables). The brute force approach of learning a parameter for each interaction of every order comes at an exponential computational and memory cost (curse of dimensionality). To alleviate this issue, it has been proposed to implicitly represent the model parameters as a tensor, the order of which is equal to the number of features; for efficiency, it can be further factorized into a compact tensor train (TT) format. However, both TT and other tensor networks (TNs), such as tensor ring and hierarchical Tucker, are sensitive to the ordering of their indices (and hence to the features). To establish the desired invariance to feature ordering, we propose to represent the weight tensor through the canonical polyadic (CP) decomposition (CPD) and introduce the associated inference and learning algorithms, including suitable regularization and initialization schemes. It is demonstrated that the proposed CP-based predictor significantly outperforms other TN-based predictors on sparse data while exhibiting comparable performance on dense nonsequential tasks. Furthermore, for enhanced expressiveness, we generalize the framework to allow feature mapping to arbitrarily high-dimensional feature vectors. In conjunction with feature vector normalization, this is shown to yield dramatic improvements in performance for dense nonsequential tasks, matching models such as fully connected neural networks. [ABSTRACT FROM AUTHOR]

Neural networks have been demonstrated to be trainable even with hundreds of layers, which exhibit remarkable improvement on expressive power and provide significant performance gains in a variety of tasks. However, the prohibitive computational cost has become a severe challenge for deploying them on resource-constrained platforms. Meanwhile, widely adopted deep neural network architectures, for example, ResNets or DenseNets, are manually crafted on benchmark datasets, which hamper their generalization ability to other domains. To cope with these issues, we propose an evolutionary algorithm-based method for shallowing deep neural networks (DNNs) at block levels, which is termed as ESNB. Different from existing studies, ESNB utilizes the ensemble view of block-wise DNNs and employs the multiobjective optimization paradigm to reduce the number of blocks while avoiding performance degradation. It automatically discovers shallower network architectures by pruning less informative blocks, and employs knowledge distillation to recover the performance. Moreover, a novel prior knowledge incorporation strategy is proposed to improve the exploration ability of the evolutionary search process, and a correctness-aware knowledge distillation strategy is designed for better knowledge transferring. Experimental results show that the proposed method can effectively accelerate the inference of DNNs while achieving superior performance when compared with the state-of-the-art competing methods. [ABSTRACT FROM AUTHOR]

Perturbation has a positive effect, as it contributes to the stability of neural systems through adaptation and robustness. For example, deep reinforcement learning generally engages in exploratory behavior by injecting noise into the action space and network parameters. It can consistently increase the agent’s exploration ability and lead to richer sets of behaviors. Evolutionary strategies also apply parameter perturbations, which makes network architecture robust and diverse. Our main concern is whether the notion of synaptic perturbation introduced in a spiking neural network (SNN) is biologically relevant or if novel frameworks and components are desired to account for the perturbation properties of artificial neural systems. In this work, we first review part of the locality-sensitive hashing (LSH) of similarity search, the FLY algorithm, as recently published in Science, and propose an improved architecture, time-shifted spiking LSH (TS-SLSH), with the consideration of temporal perturbations of the firing moments of spike pulses. Experiment results show promising performance of the proposed method and demonstrate its generality to various spiking neuron models. Therefore, we expect temporal perturbation to play an active role in SNN performance. [ABSTRACT FROM AUTHOR]

The cerebellum plays a vital role in motor learning and control with supervised learning capability, while neuromorphic engineering devises diverse approaches to high-performance computation inspired by biological neural systems. This article presents a large-scale cerebellar network model for supervised learning, as well as a cerebellum-inspired neuromorphic architecture to map the cerebellar anatomical structure into the large-scale model. Our multinucleus model and its underpinning architecture contain approximately 3.5 million neurons, upscaling state-of-the-art neuromorphic designs by over 34 times. Besides, the proposed model and architecture incorporate 3411k granule cells, introducing a 284 times increase compared to a previous study including only 12k cells. This large scaling induces more biologically plausible cerebellar divergence/convergence ratios, which results in better mimicking biology. In order to verify the functionality of our proposed model and demonstrate its strong biomimicry, a reconfigurable neuromorphic system is used, on which our developed architecture is realized to replicate cerebellar dynamics during the optokinetic response. In addition, our neuromorphic architecture is used to analyze the dynamical synchronization within the Purkinje cells, revealing the effects of firing rates of mossy fibers on the resonance dynamics of Purkinje cells. Our experiments show that real-time operation can be realized, with a system throughput of up to 4.70 times larger than previous works with high synaptic event rate. These results suggest that the proposed work provides both a theoretical basis and a neuromorphic engineering perspective for brain-inspired computing and the further exploration of cerebellar learning. [ABSTRACT FROM AUTHOR]

In this article, we propose a novel model-parallel learning method, called local critic training, which trains neural networks using additional modules called local critic networks. The main network is divided into several layer groups, and each layer group is updated through error gradients estimated by the corresponding local critic network. We show that the proposed approach successfully decouples the update process of the layer groups for both convolutional neural networks (CNNs) and recurrent neural networks (RNNs). In addition, we demonstrate that the proposed method is guaranteed to converge to a critical point. We also show that trained networks by the proposed method can be used for structural optimization. Experimental results show that our method achieves satisfactory performance, reduces training time greatly, and decreases memory consumption per machine. Code is available at https://github.com/hjdw2/Local-critic-training. [ABSTRACT FROM AUTHOR]

Neural networks are widely used as a model for classification in a large variety of tasks. Typically, a learnable transformation (i.e., the classifier) is placed at the end of such models returning a value for each class used for classification. This transformation plays an important role in determining how the generated features change during the learning process. In this work, we argue that this transformation not only can be fixed (i.e., set as nontrainable) with no loss of accuracy and with a reduction in memory usage, but it can also be used to learn stationary and maximally separated embeddings. We show that the stationarity of the embedding and its maximal separated representation can be theoretically justified by setting the weights of the fixed classifier to values taken from the coordinate vertices of the three regular polytopes available in $\mathbb {R}^{d}$ , namely, the $d$ -Simplex, the $d$ -Cube, and the $d$ -Orthoplex. These regular polytopes have the maximal amount of symmetry that can be exploited to generate stationary features angularly centered around their corresponding fixed weights. Our approach improves and broadens the concept of a fixed classifier, recently proposed by Hoffer et al., to a larger class of fixed classifier models. Experimental results confirm the theoretical analysis, the generalization capability, the faster convergence, and the improved performance of the proposed method. Code will be publicly available. [ABSTRACT FROM AUTHOR]

Change detection based on heterogeneous images, such as optical images and synthetic aperture radar images, is a challenging problem because of their huge appearance differences. To combat this problem, we propose an unsupervised change detection method that contains only a convolutional autoencoder (CAE) for feature extraction and the commonality autoencoder for commonalities exploration. The CAE can eliminate a large part of redundancies in two heterogeneous images and obtain more consistent feature representations. The proposed commonality autoencoder has the ability to discover common features of ground objects between two heterogeneous images by transforming one heterogeneous image representation into another. The unchanged regions with the same ground objects share much more common features than the changed regions. Therefore, the number of common features can indicate changed regions and unchanged regions, and then a difference map can be calculated. At last, the change detection result is generated by applying a segmentation algorithm to the difference map. In our method, the network parameters of the commonality autoencoder are learned by the relevance of unchanged regions instead of the labels. Our experimental results on five real data sets demonstrate the promising performance of the proposed framework compared with several existing approaches. [ABSTRACT FROM AUTHOR]

This article is concerned with the $H_{\infty }$ state estimation problem for a class of bidirectional associative memory (BAM) neural networks with binary mode switching, where the distributed delays are included in the leakage terms. A couple of stochastic variables taking values of 1 or 0 are introduced to characterize the switching behavior between the redundant models of the BAM neural network, and a general type of neuron activation function (i.e., the sector-bounded nonlinearity) is considered. In order to prevent the data transmissions from collisions, a periodic scheduling protocol (i.e., round-robin protocol) is adopted to orchestrate the transmission order of sensors. The purpose of this work is to develop a full-order estimator such that the error dynamics of the state estimation is exponentially mean-square stable and the $H_{\infty }$ performance requirement of the output estimation error is also achieved. Sufficient conditions are established to ensure the existence of the required estimator by constructing a mode-dependent Lyapunov-Krasovskii functional. Then, the desired estimator parameters are obtained by solving a set of matrix inequalities. Finally, a numerical example is provided to show the effectiveness of the proposed estimator design method. [ABSTRACT FROM AUTHOR]

The human brain is the gold standard of adaptive learning. It not only can learn and benefit from experience, but also can adapt to new situations. In contrast, deep neural networks only learn one sophisticated but fixed mapping from inputs to outputs. This limits their applicability to more dynamic situations, where the input to output mapping may change with different contexts. A salient example is continual learning-learning new independent tasks sequentially without forgetting previous tasks. Continual learning of multiple tasks in artificial neural networks using gradient descent leads to catastrophic forgetting, whereby a previously learned mapping of an old task is erased when learning new mappings for new tasks. Herein, we propose a new biologically plausible type of deep neural network with extra, out-of-network, task-dependent biasing units to accommodate these dynamic situations. This allows, for the first time, a single network to learn potentially unlimited parallel input to output mappings, and to switch on the fly between them at runtime. Biasing units are programed by leveraging beneficial perturbations (opposite to well-known adversarial perturbations) for each task. Beneficial perturbations for a given task bias the network toward that task, essentially switching the network into a different mode to process that task. This largely eliminates catastrophic interference between tasks. Our approach is memory-efficient and parameter-efficient, can accommodate many tasks, and achieves the state-of-the-art performance across different tasks and domains. [ABSTRACT FROM AUTHOR]

Due to the huge success and rapid development of convolutional neural networks (CNNs), there is a growing demand for hardware accelerators that accommodate a variety of CNNs to improve their inference latency and energy efficiency, in order to enable their deployment in real-time applications. Among popular platforms, field-programmable gate arrays (FPGAs) have been widely adopted for CNN acceleration because of their capability to provide superior energy efficiency and low-latency processing, while supporting high reconfigurability, making them favorable for accelerating rapidly evolving CNN algorithms. This article introduces a highly customized streaming hardware architecture that focuses on improving the compute efficiency for streaming applications by providing full-stack acceleration of CNNs on FPGAs. The proposed accelerator maps most computational functions, that is, convolutional and deconvolutional layers into a singular unified module, and implements the residual and concatenative connections between the functions with high efficiency, to support the inference of mainstream CNNs with different topologies. This architecture is further optimized through exploiting different levels of parallelism, layer fusion, and fully leveraging digital signal processing blocks (DSPs). The proposed accelerator has been implemented on Intel’s Arria 10 GX1150 hardware and evaluated with a wide range of benchmark models. The results demonstrate a high performance of over 1.3 TOP/s of throughput, up to 97% of compute [multiply-accumulate (MAC)] efficiency, which outperforms the state-of-the-art FPGA accelerators. [ABSTRACT FROM AUTHOR]

Power amplifier (PA) models, such as the neural network (NN) models and the multilayer NN models, have problems with high complexity. In this article, we first propose a novel behavior model for wideband PAs, using a real-valued time-delay convolutional NN (RVTDCNN). The input data of the model is sorted and arranged as a graph composed of the in-phase and quadrature ($I/Q$) components and envelope-dependent terms of current and past signals. Then, we created a predesigned filter using the convolutional layer to extract the basis functions required for the PA forward or reverse modeling. Finally, the generated rich basis functions are input into a simple, fully connected layer to build the model. Due to the weight sharing characteristics of the convolutional model’s structure, the strong memory effect does not lead to a significant increase in the complexity of the model. Meanwhile, the extraction effect of the predesigned filter also reduces the training complexity of the model. The experimental results show that the performance of the RVTDCNN model is almost the same as the NN models and the multilayer NN models. Meanwhile, compared with the abovementioned models, the coefficient number and computational complexity of the RVTDCNN model are significantly reduced. This advantage is noticeable when the memory effects of the PA are increased by using wider signal bandwidths. [ABSTRACT FROM AUTHOR]

Deep convolutional neural networks (CNNs) have demonstrated promising performance on image classification tasks, but the manual design process becomes more and more complex due to the fast depth growth and the increasingly complex topologies of CNNs. As a result, neural architecture search (NAS) has emerged to automatically design CNNs that outperform handcrafted counterparts. However, the computational cost is immense, e.g., 22400 GPU-days and 2000 GPU-days for two outstanding NAS works named NAS and NASNet, respectively, which motivates this work. A new effective and efficient surrogate-assisted particle swarm optimization (PSO) algorithm is proposed to automatically evolve CNNs. This is achieved by proposing a novel surrogate model, a new method of creating a surrogate data set, and a new encoding strategy to encode variable-length blocks of CNNs, all of which are integrated into a PSO algorithm to form the proposed method. The proposed method shows its effectiveness by achieving the competitive error rates of 3.49% on the CIFAR-10 data set, 18.49% on the CIFAR-100 data set, and 1.82% on the SVHN data set. The CNN blocks are efficiently learned by the proposed method from CIFAR-10 within 3 GPU-days due to the acceleration achieved by the surrogate model and the surrogate data set to avoid the training of 80.1% of CNN blocks represented by the particles. Without any further search, the evolved blocks from CIFAR-10 can be successfully transferred to CIFAR-100, SVHN, and ImageNet, which exhibits the transferability of the block learned by the proposed method. [ABSTRACT FROM AUTHOR]

This article addresses the problem of fault-tolerant consensus control of a general nonlinear multiagent system subject to actuator faults and disturbed and faulty networks. By using neural network (NN) and adaptive control techniques, estimations of unknown state-dependent boundaries of nonlinear dynamics and actuator faults, which can reflect the worst impacts on the system, are first developed. A novel NN-based adaptive observer is designed for the observation of faulty transformation signals in networks. On the basis of the NN-based observer and adaptive control strategies, fault-tolerant consensus control schemes are designed to guarantee the bounded consensus of the closed-loop multiagent system with disturbed and faulty networks and actuator faults. The validity of the proposed adaptively distributed consensus control schemes is demonstrated by a multiagent system composed of five nonlinear forced pendulums. [ABSTRACT FROM AUTHOR]

MixUp is an effective data augmentation method to regularize deep neural networks via random linear interpolations between pairs of samples and their labels. It plays an important role in model regularization, semisupervised learning (SSL), and domain adaption. However, despite its empirical success, its deficiency of randomly mixing samples has poorly been studied. Since deep networks are capable of memorizing the entire data set, the corrupted samples generated by vanilla MixUp with a badly chosen interpolation policy will degrade the performance of networks. To overcome overfitting to corrupted samples, inspired by metalearning (learning to learn), we propose a novel technique of learning to a mixup in this work, namely, MetaMixUp. Unlike the vanilla MixUp that samples interpolation policy from a predefined distribution, this article introduces a metalearning-based online optimization approach to dynamically learn the interpolation policy in a data-adaptive way (learning to learn better). The validation set performance via metalearning captures the noisy degree, which provides optimal directions for interpolation policy learning. Furthermore, we adapt our method for pseudolabel-based SSL along with a refined pseudolabeling strategy. In our experiments, our method achieves better performance than vanilla MixUp and its variants under SL configuration. In particular, extensive experiments show that our MetaMixUp adapted SSL greatly outperforms MixUp and many state-of-the-art methods on CIFAR-10 and SVHN benchmarks under the SSL configuration. [ABSTRACT FROM AUTHOR]

Nature has always inspired the human spirit and scientists frequently developed new methods based on observations from nature. Recent advances in imaging and sensing technology allow fascinating insights into biological neural processes. With the objective of finding new strategies to enhance the learning capabilities of neural networks, we focus on a phenomenon that is closely related to learning tasks and neural stability in biological neural networks, called homeostatic plasticity. Among the theories that have been developed to describe homeostatic plasticity, synaptic scaling has been found to be the most mature and applicable. We systematically discuss previous studies on the synaptic scaling theory and how they could be applied to artificial neural networks. Therefore, we utilize information theory to analytically evaluate how mutual information is affected by synaptic scaling. Based on these analytic findings, we propose two flavors in which synaptic scaling can be applied in the training process of simple and complex, feedforward, and recurrent neural networks. We compare our approach with state-of-the-art regularization techniques on standard benchmarks. We found that the proposed method yields the lowest error in both regression and classification tasks compared to previous regularization approaches in our experiments across a wide range of network feedforward and recurrent topologies and data sets. [ABSTRACT FROM AUTHOR]