Search

The rising popularity of intelligent mobile devices and the computational cost of deep learning-based models call for efficient and accurate on-device inference schemes. We propose a novel model compression scheme that allows inference to be carried out using bit-level sparsity, which can be efficiently implemented using in-memory computing macros. In this paper, we introduce a method called BitS-Net to leverage the benefits of bit-sparsity (where the number of zeros are more than number of ones in binary representation of weight/activation values) when applied to compute-in-memory (CIM) with resistive RAM (RRAM) to develop energy efficient DNN accelerators operating in the inference mode. We demonstrate that BitS-Net improves the energy efficiency by up to 5x for ResNet models on the ImageNet dataset. [ABSTRACT FROM AUTHOR]

The integration of wind power with traditional electricity grids is challenging because of the volatile and intermittent nature of wind speed. Consequently, wind speed forecasting is an essential tool in the energy system because it supports economic, regulation, and operational efficiency issues decisions. The development of systems for accurate wind speed forecasting is a difficult task because it presents linear and nonlinear temporal patterns, and is differently influenced by climate and environmental variables. In this paper, we propose a method to search for the best combination of linear, nonlinear forecasts, and exogenous variables, aiming to increase the accuracy of wind speed forecasting. The proposed method performs: (i) the choice of the linear statistical model to forecast the time series; (ii) the error series modeling using a Long Short-Term Memory model; and (iii) the evolutionary search for the most suitable nonlinear combination of linear and nonlinear forecasts with exogenous variables. The proposed method is assessed using hourly and monthly time series of three stations located in northeast Brazil. The evaluation of the results using two traditional measures shows that the proposed method has a better performance than statistical techniques, Machine Learning models, and state-of-the-art hybrid systems in most series. [ABSTRACT FROM AUTHOR]

Event causality identification is an important research task in natural language processing. Existing methods largely focus on identifying explicit causal relations, and give poor performance in implicit causalities, especially in the document level. In this paper, we formalize event causality identification as a graph-based edge prediction problem and propose a novel document-level context-based graph inference mechanism. Specifically, we use attention-based neural networks to automatically extract document-level contextual information, and a direction-sensitive graph inference mechanism to achieve information transfer and interaction among event causalities. Experimental results on the EventStoryLine v1.5 dataset show that our approach outperforms previous methods and baseline systems by a large margin in F1-score metrics (2.45% improvement on intra-sentence causalities and 3.08% improvement on cross-sentence causalities). Further analysis demonstrates that our model can effectively capture the document-level contextual information and latent causal information among events. [ABSTRACT FROM AUTHOR]

This paper presents a novel method for learning the cyclic contents of stochastic time series: the deep time-growing neural network (DTGNN). The DTGNN combines supervised and unsupervised methods in different levels of learning for an enhanced performance. It is employed by a multiscale learning structure to classify cyclic time series (CTS), in which the dynamic contents of the time series are preserved in an efficient manner. This paper suggests a systematic procedure for finding the design parameter of the classification method for a one-versus-multiple class application. A novel validation method is also suggested for evaluating the structural risk, both in a quantitative and a qualitative manner. The effect of the DTGNN on the performance of the classifier is statistically validated through the repeated random subsampling using different sets of CTS, from different medical applications. The validation involves four medical databases, comprised of 108 recordings of the electroencephalogram signal, 90 recordings of the electromyogram signal, 130 recordings of the heart sound signal, and 50 recordings of the respiratory sound signal. Results of the statistical validations show that the DTGNN significantly improves the performance of the classification and also exhibits an optimal structural risk. [ABSTRACT FROM AUTHOR]

• Entropy based gradient compression mechanism for distributed training. • New distributed training algorithms for deep neural network training. • Experiments to examine and validate the advantages of the proposed algorithms. With the increase of volume of training data and scale of network models, distributed deep learning is becoming more and more popular, which employs multiple workers to train a single model. However, communication among workers has been always a major challenge, because it may cause large time latency and bandwidth consumption. In this paper, we propose an entropy-based gradient compression (EGC) mechanism to reduce communication overhead. EGC selects the gradients communicated based on the entropy of the gradient items, which can achieve a high compression ratio without sacrificing accuracy. More importantly, EGC is a general and flexible mechanism that can be adopted in different distributed training algorithms. Accordingly, we propose three EGC-based training algorithms for different scenarios, i.e., EGC-DSGD for decentralized training, EGC-PS for centralized training, and EGC-FL for federated training. To improve the accuracy of these algorithms, we also adopt associated mechanisms, including automatic learning rate correction, momentum correction and residuals accumulation. We prove the convergence of EGC by analysis and evaluate its performance by experiments. Eight models are trained using popular public datasets (including MNIST, CIFAR-10, Tiny ImageNet and Penn Treebank) for the tasks of image classification and language modeling. The experimental results show that, compared with existing works, the EGC based algorithms can achieve roughly 1000 times gradient compression ratio while keeping the accuracy similar or even higher. [ABSTRACT FROM AUTHOR]

In modern era, the graphical information is presented in the form of web images. As the dependency of human beings on web information is increasing day-by-day, so the spammers are injecting spam by adopting new spamming techniques. Image spam is a spamming technique that integrates spam text contents into graphical images in order to bypass conventional text-based spam filters. The spam images are of various categories, such as redirection spam, advertisement spam, fake review, and content spam. In order to detect image spam efficiently, it is important to analyze the features of the image data. However, the existing image spam detection techniques in literature focused on textual or graphic features of the image. Moreover, to extract the relevant features from the images is also a challenging task. So, to fill these gaps, in this paper, we propose a Spam Protector for Advertisement of Malicious Images (SPAMI) framework using features extraction by browsing different websites and webpages. SPAMI is a cognitive spam protector which labels the spam advertisement images by using deep learning models. Three deep learning models are used for the same, i.e., CNN, RNN, and LSTM. The regress analysis of output from these models is done in the proposed SPAMI framework. Finally, we analysed the labels (Advertisement, Suspicious, Normal) for all the 600 images collected. The accuracy obtained from these models is 95% with real-time collected images, which improved up to 97% when tested with "Image Spam Hunter" dataset. [ABSTRACT FROM AUTHOR]

Reinforcement learning has shown to be profoundly successful at learning optimal policies for simulated environments using distributed training with extensive compute capacity. Model-free reinforcement learning uses the notion of trial and error, where the error is a vital part of learning the agent to behave optimally. In mission-critical, real-world environments, there is little tolerance for failure and can cause damaging effects on humans and equipment. In these environments, current state-of-the-art reinforcement learning approaches are not sufficient to learn optimal control policies safely. On the other hand, model-based reinforcement learning tries to encode environment transition dynamics into a predictive model. The transition dynamics describes the mapping from one state to another, conditioned on an action. If this model is accurate enough, the predictive model is sufficient to train agents for optimal behavior in real environments. This paper presents the Dreaming Variational Autoencoder (DVAE) for safely learning good policies with a significantly lower risk of catastrophes occurring during training. The algorithm combines variational autoencoders, risk-directed exploration, and curiosity to train deep-q networks inside "dream" states. We introduce a novel environment, ASRS-Lab, for research in the safe learning of autonomous vehicles in grid-based warehousing. The work shows that the proposed algorithm has better sample efficiency with similar performance to novel model-free deep reinforcement learning algorithms while maintaining safety during training. [ABSTRACT FROM AUTHOR]

In this paper, we present a novel deep metric learning method to tackle the multi-label image classification problem. In order to better learn the correlations among images features, as well as labels, we attempt to explore a latent space, where images and labels are embedded via two unique deep neural networks, respectively. To capture the relationships between image features and labels, we aim to learn a two-way deep distance metric over the embedding space from two different views, i.e., the distance between one image and its labels is not only smaller than those distances between the image and its labels’ nearest neighbors but also smaller than the distances between the labels and other images corresponding to the labels’ nearest neighbors. Moreover, a reconstruction module for recovering correct labels is incorporated into the whole framework as a regularization term, such that the label embedding space is more representative. Our model can be trained in an end-to-end manner. Experimental results on publicly available image data sets corroborate the efficacy of our method compared with the state of the arts. [ABSTRACT FROM AUTHOR]

Deep neural networks bring in impressive accuracy in various applications, but the success often relies on heavy network architectures. Taking well-trained heavy networks as teachers, classical teacher–student learning paradigm aims to learn a student network that is lightweight yet accurate. In this way, a portable student network with significantly fewer parameters can achieve considerable accuracy, which is comparable to that of a teacher network. However, beyond accuracy, the robustness of the learned student network against perturbation is also essential for practical uses. Existing teacher-student learning frameworks mainly focus on accuracy and compression ratios, but ignore the robustness. In this paper, we make the student network produce more confident predictions with the help of the teacher network, and analyze the lower bound of the perturbation that will destroy the confidence of the student network. Two important objectives regarding prediction scores and gradients of examples are developed to maximize this lower bound, to enhance the robustness of the student network without sacrificing the performance. Experiments on benchmark data sets demonstrate the efficiency of the proposed approach to learning robust student networks that have satisfying accuracy and compact sizes. [ABSTRACT FROM AUTHOR]

Traditionally, deep learning algorithms update the network weights, whereas the network architecture is chosen manually using a process of trial and error. In this paper, we propose two novel approaches that automatically update the network structure while also learning its weights. The novelty of our approach lies in our parameterization, where the depth, or additional complexity, is encapsulated continuously in the parameter space through control parameters that add additional complexity. We propose two methods. In tunnel networks, this selection is done at the level of a hidden unit, and in budding perceptrons, this is done at the level of a network layer; updating this control parameter introduces either another hidden unit or layer. We show the effectiveness of our methods on the synthetic two-spiral data and on three real data sets of MNIST, MIRFLICKR, and CIFAR, where we see that our proposed methods, with the same set of hyperparameters, can correctly adjust the network complexity to the task complexity. [ABSTRACT FROM AUTHOR]

We present the global deep video representation learning to video-based person reidentification (re-ID) that aggregates local 3-D features across the entire video extent. Existing methods typically extract frame-wise deep features from 2-D convolutional networks (ConvNets) which are pooled temporally to produce the video-level representations. However, 2-D ConvNets lose temporal priors immediately after the convolutions, and a separate temporal pooling is limited in capturing human motion in short sequences. In this paper, we present global video representation learning, to be complementary to 3-D ConvNets as a novel layer to capture the appearance and motion dynamics in full-length videos. Nevertheless, encoding each video frame in its entirety and computing aggregate global representations across all frames is tremendously challenging due to the occlusions and misalignments. To resolve this, our proposed network is further augmented with the 3-D part alignment to learn local features through the soft-attention module. These attended features are statistically aggregated to yield identity-discriminative representations. Our global 3-D features are demonstrated to achieve the state-of-the-art results on three benchmark data sets: MARS, Imagery Library for Intelligent Detection Systems-Video Re-identification, and PRID2011. [ABSTRACT FROM AUTHOR]

Due to object detection’s close relationship with video analysis and image understanding, it has attracted much research attention in recent years. Traditional object detection methods are built on handcrafted features and shallow trainable architectures. Their performance easily stagnates by constructing complex ensembles that combine multiple low-level image features with high-level context from object detectors and scene classifiers. With the rapid development in deep learning, more powerful tools, which are able to learn semantic, high-level, deeper features, are introduced to address the problems existing in traditional architectures. These models behave differently in network architecture, training strategy, and optimization function. In this paper, we provide a review of deep learning-based object detection frameworks. Our review begins with a brief introduction on the history of deep learning and its representative tool, namely, the convolutional neural network. Then, we focus on typical generic object detection architectures along with some modifications and useful tricks to improve detection performance further. As distinct specific detection tasks exhibit different characteristics, we also briefly survey several specific tasks, including salient object detection, face detection, and pedestrian detection. Experimental analyses are also provided to compare various methods and draw some meaningful conclusions. Finally, several promising directions and tasks are provided to serve as guidelines for future work in both object detection and relevant neural network-based learning systems. [ABSTRACT FROM AUTHOR]

This paper proposes a new and efficient technique to regularize the neural network in the context of deep learning using correlations among features. Previous studies have shown that oversized deep neural network models tend to produce a lot of redundant features that are either the shifted version of one another or are very similar and show little or no variations, thus resulting in redundant filtering. We propose a way to address this problem and show that such redundancy can be avoided using regularization and adaptive feature dropout mechanism. We show that regularizing both negative and positive correlated features according to their differentiation and based on their relative cosine distances yields network extracting dissimilar features with less overfitting and better generalization. This concept is illustrated with deep multilayer perceptron, convolutional neural network, sparse autoencoder, gated recurrent unit, and long short-term memory on MNIST digits recognition, CIFAR-10, ImageNet, and Stanford Natural Language Inference data sets. [ABSTRACT FROM AUTHOR]

In this paper, we study a new problem in the scenario of sequences labeling. To be exact, we consider that the training data are with annotation of various degrees, namely, fully labeled, unlabeled, and partially labeled sequences. The learning with fully un/labeled sequence refers to the standard setting in traditional un/supervised learning, and the proposed partially labeling specifies the subject that the element does not belong to. The partially labeled data are cheaper to obtain compared with the fully labeled data though it is less informative, especially when the tasks require a lot of domain knowledge. To solve such a practical challenge, we propose a novel deep conditional random field (CRF) model which utilizes an end-to-end learning manner to smoothly handle fully/un/partially labeled sequences within a unified framework. To the best of our knowledge, this could be one of the first works to utilize the partially labeled instance for sequence labeling, and the proposed algorithm unifies the deep learning and CRF in an end-to-end framework. Extensive experiments show that our method achieves state-of-the-art performance in two sequence labeling tasks on some popular data sets. [ABSTRACT FROM AUTHOR]

Convolutional autoencoders (CAEs) have shown their remarkable performance in stacking to deep convolutional neural networks (CNNs) for classifying image data during the past several years. However, they are unable to construct the state-of-the-art CNNs due to their intrinsic architectures. In this regard, we propose a flexible CAE (FCAE) by eliminating the constraints on the numbers of convolutional layers and pooling layers from the traditional CAE. We also design an architecture discovery method by exploiting particle swarm optimization, which is capable of automatically searching for the optimal architectures of the proposed FCAE with much less computational resource and without any manual intervention. We test the proposed approach on four extensively used image classification data sets. Experimental results show that our proposed approach in this paper significantly outperforms the peer competitors including the state-of-the-art algorithms. [ABSTRACT FROM AUTHOR]

The softmax function is widely used in deep neural networks (DNNs), its hardware performance plays an important role in the training and inference of DNN accelerators. However, due to the complexity of the traditional softmax, the existing hardware architectures are resource-consuming or have low precision. In order to address the challenges, we study a base-2 softmax function in terms of its suitability for neural network training and efficient hardware implementation. Compared to the classical base- $e$ softmax function, the base-2 softmax function is a new softmax function that uses 2 as the exponential base instead of $e$. From the aspects of mathematical derivation and software simulation, we first demonstrate the feasibility and good accuracy of the base-2 softmax function in the application of neural network training. Then, we use the symmetric-mapping lookup table (SM-LUT) method to design a low-complexity architecture but with high precision to implement it. Under TSMC 28nm CMOS technology, an example design of our architecture has the area of $5676 ~\mu m^{2}$ and the power consumption of 13.12 mW for circuit synthesis at the frequency of 3 GHz. Compared with the latest works, our architecture achieves the best performance and efficiency. [ABSTRACT FROM AUTHOR]

Multi-bit-width convolutional neural network (CNN) maintains the balance between network accuracy and hardware efficiency, thus enlightening a promising method for accurate yet energy-efficient edge computing. In this work, we develop state-of-the-art multi-bit-width accelerator for NAS Optimized deep learning neural networks. To efficiently process the multi-bit-width network inferencing, multi-level optimizations have been proposed. Firstly, differential Neural Architecture Search (NAS) method is adopted for the high accuracy multi-bit-width network generation. Secondly, hybrid Booth based multi-bit-width multiply-add-accumulation (MAC) unit is developed for data processing. Thirdly, vector systolic array is proposed for effectively accelerating the matrix multiplications. With vector-style systolic dataflow, both the processing time and logic resources consumption can be reduced when compared with the classical systolic array. Finally, The proposed multi-bit-width CNN acceleration scheme has been practically deployed on FPGA platform of Xilinx ZCU102. Average performance on accelerating the full NAS optimized VGG16 network is 784.2 GOPS, and peek performance of the convolutional layer can reach as high as 871.26 GOPS for INT8, 1676.96 GOPS for INT4, and 2863.29 GOPS for INT2 respectively, which is among the best results in previous CNN accelerator benchmarks. [ABSTRACT FROM AUTHOR]

Keyword spotting (KWS) is a crucial function enabling the interaction with the many ubiquitous smart devices in our surroundings, either activating them through wake-word or directly as a human-computer interface. For many applications, KWS is the entry point for our interactions with the device and, thus, an always-on workload. Many smart devices are mobile and their battery lifetime is heavily impacted by continuously running services. KWS and similar always-on services are thus the focus when optimizing the overall power consumption. This work addresses KWS energy-efficiency on low-cost microcontroller units (MCUs). We combine analog binary feature extraction with binary neural networks. By replacing the digital preprocessing with the proposed analog front-end, we show that the energy required for data acquisition and preprocessing can be reduced by $29\times $ , cutting its share from a dominating 85% to a mere 16% of the overall energy consumption for our reference KWS application. Experimental evaluations on the Speech Commands Dataset show that the proposed system outperforms state-of-the-art accuracy and energy efficiency, respectively, by 1% and $4.3\times $ on a 10-class dataset while providing a compelling accuracy-energy trade-off including a 2% accuracy drop for a $71\times $ energy reduction. [ABSTRACT FROM AUTHOR]

Batch normalization (BN) has recently become a standard component for accelerating and improving the training of deep neural networks (DNNs). However, BN brings in additional calculations, consumes more memory, and significantly slows down the training iteration. Furthermore, the nonlinear square and sqrt operations in the normalization process impede low bit-width quantization techniques, which draw much attention to the deep learning hardware community. In this paper, we propose an $L1$ -norm BN (L1BN) with only linear operations in both forward and backward propagations during training. L1BN is approximately equivalent to the conventional $L2$ -norm BN (L2BN) by multiplying a scaling factor that equals $({\pi }/{2})^{1/2}$. Experiments on various convolutional neural networks and generative adversarial networks reveal that L1BN can maintain the same performance and convergence rate as L2BN but with higher computational efficiency. In real application-specified integrated circuit synthesis with reduced resources, L1BN achieves 25% speedup and 37% energy saving compared to the original L2BN. Our hardware-friendly normalization method not only surpasses L2BN in speed but also simplifies the design of deep learning accelerators. Last but not least, L1BN promises a fully quantized training of DNNs, which empowers future artificial intelligence applications on mobile devices with transfer and continual learning capability. [ABSTRACT FROM AUTHOR]

Online peer-to-peer (P2P) lending is expected to benefit both investors and borrowers due to their low transaction cost and the elimination of expensive intermediaries. From the lenders’ perspective, maximizing their return on investment is an ultimate goal during their decision-making procedure. In this paper, we explore and address a fundamental problem underlying such a goal: how to represent the two competing risks, charge-off and prepayment, in funded loans. We propose to model both potential risks simultaneously, which remains largely unexplored until now. We first develop a hierarchical grading framework to integrate two risks of loans both qualitatively and quantitatively. Afterward, we introduce an end-to-end deep learning approach to solve this problem by breaking it down into multiple binary classification subproblems that are amenable to both feature representation and risks learning. Particularly, we leverage deep neural networks to jointly solve these subtasks, which leads to the in-depth exploration of the interaction involved in these tasks. To the best of our knowledge, this is the first attempt to characterize competing risks for loans in P2P lending via deep neural networks. The comprehensive experiments on real-world loan data show that our methodology is able to achieve an appealing investment performance by modeling the competition within and between risks explicitly and properly. The feature analysis based on saliency maps provides useful insights into payment dynamics of loans for potential investors intuitively. [ABSTRACT FROM AUTHOR]

An efficient deep learning requires a memory-efficient construction of a neural network. This paper introduces a layerwise tensorized formulation of a multilayer neural network, called LTNN, such that the weight matrix can be significantly compressed during training. By reshaping the multilayer neural network weight matrix into a high-dimensional tensor with a low-rank approximation, significant network compression can be achieved with maintained accuracy. An according layerwise training is developed by a modified alternating least-squares method with backward propagation for fine-tuning only. LTNN can provide the state-of-the-art results on various benchmarks with significant compression. For MNIST benchmark, LTNN shows $64 \times $ compression rate without accuracy drop. For Imagenet12 benchmark, our proposed LTNN achieves $35.84 \times $ compression of the neural network with around 2% accuracy drop. We have also shown $1.615 \times $ faster on inference speed than the existing works due to the smaller tensor core ranks. [ABSTRACT FROM AUTHOR]

Computerized microscopy image analysis plays an important role in computer aided diagnosis and prognosis. Machine learning techniques have powered many aspects of medical investigation and clinical practice. Recently, deep learning is emerging as a leading machine learning tool in computer vision and has attracted considerable attention in biomedical image analysis. In this paper, we provide a snapshot of this fast-growing field, specifically for microscopy image analysis. We briefly introduce the popular deep neural networks and summarize current deep learning achievements in various tasks, such as detection, segmentation, and classification in microscopy image analysis. In particular, we explain the architectures and the principles of convolutional neural networks, fully convolutional networks, recurrent neural networks, stacked autoencoders, and deep belief networks, and interpret their formulations or modelings for specific tasks on various microscopy images. In addition, we discuss the open challenges and the potential trends of future research in microscopy image analysis using deep learning. [ABSTRACT FROM AUTHOR]

5G millimeter-wave (mmWave) communication systems enable exciting new applications by significantly reducing the latency and increasing the data rate. However, this comes at a large computational cost, which results in long latency and large energy consumption. In this work, we aim to address this challenge in the problem of channel estimation of such systems through a set of algorithm-hardware co-optimizations. First of all, we employed a model-based neural network to improve the rate of convergence. We also optimized the neural network and achieved improved loss while using approximately the same number of operations. Furthermore, we were able to reduce the computational complexity through the use of sparsity inherent in mmWave channels. The proposed neural network for the channel estimation scales the computational complexity by more than two orders. Based on these innovations, we implemented a channel estimation subsystem on Zynq 7020 FPGA. The subsystem obtains an improvement in latency of up to ~10X and an improvement in energy consumption of up to ~300X over CPU and GPU based systems. [ABSTRACT FROM AUTHOR]

Deep reinforcement learning (RL) comprehensively uses the psychological mechanisms of “trial and error” and “reward and punishment” in RL as well as powerful feature expression and nonlinear mapping in deep learning. Currently, it plays an essential role in the fields of artificial intelligence and machine learning. Since an RL agent needs to constantly interact with its surroundings, the deep Q network (DQN) is inevitably faced with the need to learn numerous network parameters, which results in low learning efficiency. In this paper, a multisource transfer double DQN (MTDDQN) based on actor learning is proposed. The transfer learning technique is integrated with deep RL to make the RL agent collect, summarize, and transfer action knowledge, including policy mimic and feature regression, to the training of related tasks. There exists action overestimation in DQN, i.e., the lower probability limit of action corresponding to the maximum Q value is nonzero. Therefore, the transfer network is trained by using double DQN to eliminate the error accumulation caused by action overestimation. In addition, to avoid negative transfer, i.e., to ensure strong correlations between source and target tasks, a multisource transfer learning mechanism is applied. The Atari2600 game is tested on the arcade learning environment platform to evaluate the feasibility and performance of MTDDQN by comparing it with some mainstream approaches, such as DQN and double DQN. Experiments prove that MTDDQN achieves not only human-like actor learning transfer capability, but also the desired learning efficiency and testing accuracy on target task. [ABSTRACT FROM AUTHOR]

Although there have been many decades of research and commercial presence on high performance general purpose processors, there are still many applications that require fully customized hardware architectures for further computational acceleration. Recently, deep learning has been successfully used to learn in a wide variety of applications, but their heavy computation demand has considerably limited their practical applications. This paper proposes a fully pipelined acceleration architecture to alleviate high computational demand of an artificial neural network (ANN) which is restricted Boltzmann machine (RBM) ANNs. The implemented RBM ANN accelerator (integrating $1024\times 1024$ network size, using 128 input cases per batch, and running at a 303-MHz clock frequency) integrated in a state-of-the art field-programmable gate array (FPGA) (Xilinx Virtex 7 XC7V-2000T) provides a computational performance of 301-billion connection-updates-per-second and about 193 times higher performance than a software solution running on general purpose processors. Most importantly, the architecture enables over 4 times (12 times in batch learning) higher performance compared with a previous work when both are implemented in an FPGA device (XC2VP70). [ABSTRACT FROM AUTHOR]

Developments in deep learning have seen the use of layerwise unsupervised learning combined with supervised learning for fine-tuning. With this layerwise approach, a deep network can be seen as a more modular system that lends itself well to learning representations. In this paper, we investigate whether such modularity can be useful to the insertion of background knowledge into deep networks, whether it can improve learning performance when it is available, and to the extraction of knowledge from trained deep networks, and whether it can offer a better understanding of the representations learned by such networks. To this end, we use a simple symbolic language—a set of logical rules that we call confidence rules—and show that it is suitable for the representation of quantitative reasoning in deep networks. We show by knowledge extraction that confidence rules can offer a low-cost representation for layerwise networks (or restricted Boltzmann machines). We also show that layerwise extraction can produce an improvement in the accuracy of deep belief networks. Furthermore, the proposed symbolic characterization of deep networks provides a novel method for the insertion of prior knowledge and training of deep networks. With the use of this method, a deep neural–symbolic system is proposed and evaluated, with the experimental results indicating that modularity through the use of confidence rules and knowledge insertion can be beneficial to network performance. [ABSTRACT FROM AUTHOR]

Broad Learning System (BLS) that aims to offer an alternative way of learning in deep structure is proposed in this paper. Deep structure and learning suffer from a time-consuming training process because of a large number of connecting parameters in filters and layers. Moreover, it encounters a complete retraining process if the structure is not sufficient to model the system. The BLS is established in the form of a flat network, where the original inputs are transferred and placed as “mapped features” in feature nodes and the structure is expanded in wide sense in the “enhancement nodes.” The incremental learning algorithms are developed for fast remodeling in broad expansion without a retraining process if the network deems to be expanded. Two incremental learning algorithms are given for both the increment of the feature nodes (or filters in deep structure) and the increment of the enhancement nodes. The designed model and algorithms are very versatile for selecting a model rapidly. In addition, another incremental learning is developed for a system that has been modeled encounters a new incoming input. Specifically, the system can be remodeled in an incremental way without the entire retraining from the beginning. Satisfactory result for model reduction using singular value decomposition is conducted to simplify the final structure. Compared with existing deep neural networks, experimental results on the Modified National Institute of Standards and Technology database and NYU NORB object recognition dataset benchmark data demonstrate the effectiveness of the proposed BLS. [ABSTRACT FROM AUTHOR]

Under the background of the wanton spread of the coronavirus disease (COVID-19), the pandemic is changing and hitting lives all over the world. Fortunately, Spatio-temporal processes bear essential importance in many applied scientific fields. And the disease pandemic can be viewed as Spatio-temporal dynamics processes. Generally, partial differential equations (PDEs) have been widely used to investigate interfacial dynamic processes. In this work, we use a physical-constraint neural network learning the Spatio-temporal mean-field dynamics to control the propagation of epidemics. Also, we use the AI-based algorithm physical-constraint learning (PCL) to solve the minimization problems of the mean-field game (MFG) and control (MFC) problems instead of the traditional computational method. In PCL, the PDEs are encoded into the loss function, where partial derivatives can be obtained through automatic differentiation (AD). We demonstrate how they can be applied in practice by considering the problem of controlling the propagation of epidemics. Numerical experiments on different input data are implemented to demonstrate the effectiveness and superiority of the proposed models compared to the state-of-the-art approach and illustrate how to separate the infected patients in a spatial domain effectively. [ABSTRACT FROM AUTHOR]

In this paper, we introduce a novel form of association rules (ARs) that do not require discretization of continuous variables or the use of intervals in either sides of the rule. This rule form captures nonlinear relationships among variables, and provides an alternative pattern representation for mining essential relations hidden in a given data set. We refer to the new rule form as a functional AR (FAR). A new neural network-based, co-operative, coevolutionary algorithm is presented for FAR mining. The algorithm is applied to both synthetic and real-world data sets, and its performance is analyzed. The experimental results show that the proposed mining algorithm is able to discover valid and essential underlying relations in the data. Comparison experiments are also carried out with the two state-of-the-art AR mining algorithms that can handle continuous variables to demonstrate the competitive performance of the proposed method. [ABSTRACT FROM PUBLISHER]

Deep hierarchical representations of the data have been found out to provide better informative features for several machine learning applications. In addition, multilayer neural networks surprisingly tend to achieve better performance when they are subject to an unsupervised pretraining. The booming of deep learning motivates researchers to identify the factors that contribute to its success. One possible reason identified is the flattening of manifold-shaped data in higher layers of neural networks. However, it is not clear how to measure the flattening of such manifold-shaped data and what amount of flattening a deep neural network can achieve. For the first time, this paper provides quantitative evidence to validate the flattening hypothesis. To achieve this, we propose a few quantities for measuring manifold entanglement under certain assumptions and conduct experiments with both synthetic and real-world data. Our experimental results validate the proposition and lead to new insights on deep learning. [ABSTRACT FROM AUTHOR]

Machine learning inference is of broad interest, increasingly in energy-constrained applications. However, platforms are often pushed to their energy limits, especially with deep learning models, which provide state-of-the-art inference performance but are also computationally intensive. This has motivated algorithmic co-design, where flexibility in the model and model parameters, derived from training, is exploited for hardware energy efficiency. This work extends a model-training algorithm referred to as Stochastic Data-Driven Hardware Resilience (S-DDHR) to enable statistical models of computations, amenable for energy/throughput aggressive hardware operating points as well as emerging variation-prone device technologies. S-DDHR itself extends the previous approach of DDHR by incorporating the statistical distribution of hardware variations for model-parameter learning, rather than a sample of the distributions. This is critical to developing accurate and composable abstractions of computations, to enable scalable hardware-generalized training, rather than hardware instance-by-instance training. S-DDHR is demonstrated and evaluated for a bit-scalable MRAM-based in-memory computing architecture, whose energy/throughput trade-offs explicitly motivate statistical computations. Using foundry data to model MRAM device variations, S-DDHR is shown to preserve high inference performance for benchmark datasets (MNIST, CIFAR-10, SVHN) as variation parameters are scaled to high levels, exhibiting less than 3.5% accuracy drop at $10 \times $ the nominal variation level. [ABSTRACT FROM AUTHOR]

Automatic modulation recognition (AMR) detects the modulation scheme of the received signals for further signal processing without needing prior information, and provides the essential function when such information is missing. Recent breakthroughs in deep learning (DL) have laid the foundation for developing high-performance DL-AMR approaches for communications systems. Comparing with traditional modulation detection methods, DL-AMR approaches have achieved promising performance including high recognition accuracy and low false alarms due to the strong feature extraction and classification abilities of deep neural networks. Despite the promising potential, DL-AMR approaches also bring concerns to complexity and explainability, which affect the practical deployment in wireless communications systems. This paper aims to present a review of the current DL-AMR research, with a focus on appropriate DL models and benchmark datasets. We further provide comprehensive experiments to compare the state of the art models for single-input-single-output (SISO) systems from both accuracy and complexity perspectives, and propose to apply DL-AMR in the new multiple-input-multiple-output (MIMO) scenario with precoding. Finally, existing challenges and possible future research directions are discussed. [ABSTRACT FROM AUTHOR]

This paper presents a novel change detection approach for synthetic aperture radar images based on deep learning. The approach accomplishes the detection of the changed and unchanged areas by designing a deep neural network. The main guideline is to produce a change detection map directly from two images with the trained deep neural network. The method can omit the process of generating a difference image (DI) that shows difference degrees between multitemporal synthetic aperture radar images. Thus, it can avoid the effect of the DI on the change detection results. The learning algorithm for deep architectures includes unsupervised feature learning and supervised fine-tuning to complete classification. The unsupervised feature learning aims at learning the representation of the relationships between the two images. In addition, the supervised fine-tuning aims at learning the concepts of the changed and unchanged pixels. Experiments on real data sets and theoretical analysis indicate the advantages, feasibility, and potential of the proposed method. Moreover, based on the results achieved by various traditional algorithms, respectively, deep learning can further improve the detection performance. [ABSTRACT FROM PUBLISHER]

In this paper, we propose a hybrid architecture that combines the image modeling strengths of the bag of words framework with the representational power and adaptability of learning deep architectures. Local gradient-based descriptors, such as SIFT, are encoded via a hierarchical coding scheme composed of spatial aggregating restricted Boltzmann machines (RBM). For each coding layer, we regularize the RBM by encouraging representations to fit both sparse and selective distributions. Supervised fine-tuning is used to enhance the quality of the visual representation for the categorization task. We performed a thorough experimental evaluation using three image categorization data sets. The hierarchical coding scheme achieved competitive categorization accuracies of 79.7% and 86.4% on the Caltech-101 and 15-Scenes data sets, respectively. The visual representations learned are compact and the model's inference is fast, as compared with sparse coding methods. The low-level representations of descriptors that were learned using this method result in generic features that we empirically found to be transferrable between different image data sets. Further analysis reveal the significance of supervised fine-tuning when the architecture has two layers of representations as opposed to a single layer. [ABSTRACT FROM AUTHOR]

We present a novel deep learning model for a neural network that reduces both computation and data storage overhead. To do so, the proposed model proposes and combines a binary-weight neural network (BNN) training, a storage reuse technique, and an incremental training scheme. The storage requirements can be tuned to meet the desired classification accuracy, storing more parameters on an on-chip memory, and thereby reducing off-chip data storage accesses. Our experiments show 4–6 $\times $ reduction in weight storage footprint when training binary deep neural network models. On the FPGA platform, this results in a reduced amount of off-chip accesses, enabling our model to train a neural network in $14\times $ shorter latency, as compared to the conventional BNN training method. [ABSTRACT FROM AUTHOR]

A deep neural network (DNN) online learning processor is proposed with high throughput and low power consumption to achieve real-time object tracking in mobile devices. Four key features enable a low-power DNN online learning. First, a proposed processor is designed with a unified core architecture and it achieves $1.33\times $ higher throughput than the previous state-of-the-art DNN learning processor. Second, the new algorithms, binary feedback alignment (BFA), and dynamic fixed-point based run-length compression (RLC), are proposed and reduce power consumption through the reduction of external memory accesses (EMA). The BFA and dynamic fixed-point-based RLC reduce the EMA by 11.4% and 32.5%, respectively. Third, the new data feeding units, including an integral RLC (iRLC) decoder and a transpose RLC (tRLC) decoder, are co-designed to maximize throughput alongside the proposed algorithms. Finally, a dropout controller in this processor reduces redundant power consumption coming from the unified core and the data feeding architecture by the proposed dynamic clock-gating scheme. This enables the proposed processor to operate DNN online learning with 38.1% lower power consumption. Implemented with 65 nm CMOS technology, the 3.52 mm2 DNN online learning processor shows 126 mW power consumption and the processor achieves 30.4 frames-per-second throughput in the object tracking application. [ABSTRACT FROM AUTHOR]

As a result of the increasing demand for deep neural network (DNN)-based services, efforts to develop hardware accelerators for DNNs are growing rapidly. However, while highly efficient accelerators on convolutional DNNs (Conv-DNNs) have been developed, less progress has been made with regards to fully-connected DNNs. Based on analysis of bit-level SRAM errors, we propose memory adaptive training with in-situ canaries (MATIC), a methodology that enables aggressive voltage scaling of accelerator weight memories to improve the energy-efficiency of DNN accelerators. To enable accurate operation with voltage overscaling, MATIC combines characteristics of SRAM bit failures with the error resilience of neural networks in a memory-adaptive training (MAT) process. Furthermore, PVT-related voltage margins are eliminated using bit-cells from synaptic weights as in-situ canaries to track runtime environmental variation. Demonstrated on a low-power DNN accelerator fabricated in 65 nm CMOS, MATIC enables up to $3.3\times $ energy reduction versus the nominal voltage, or $18.6\times $ application error reduction. We also perform a simulation study that extends MAT to Conv-DNNs, and characterize the accuracy impact of bit failure statistics. Finally, we develop a weight refinement algorithm to improve the performance of MAT, and show that it improves absolute accuracy by 0.8–1.3% or reduces training time by 5– $10\times $. [ABSTRACT FROM AUTHOR]

Convolutional neural network (CNN) offers significant accuracy in image detection. To implement image detection using CNN in the Internet of Things (IoT) devices, a streaming hardware accelerator is proposed. The proposed accelerator optimizes the energy efficiency by avoiding unnecessary data movement. With unique filter decomposition technique, the accelerator can support arbitrary convolution window size. In addition, max-pooling function can be computed in parallel with convolution by using separate pooling unit, thus achieving throughput improvement. A prototype accelerator was implemented in TSMC 65-nm technology with a core size of 5 mm2. The accelerator can support major CNNs and achieve 152GOPS peak throughput and 434GOPS/W energy efficiency at 350 mW, making it a promising hardware accelerator for intelligent IoT devices. [ABSTRACT FROM PUBLISHER]