Search

Due to the diversity of power consumption patterns, the false positive rate (FPR) of data-driven electricity theft detection (ETD) methods is too high to meet practical needs, which severely restricts the engineering application of data-based methods. To reduce FPR of ETD methods based on advanced metering infrastructure (AMI), a deep neural network with low FPR (LFPR-DNN) is proposed in this paper. First, a deep model is constructed based on one-dimensional convolution and residual network, which can automatically extract features from consumption data. Then, a two-stage training scheme is used to train the network. In the first stage, the conventional gradient descent algorithm is used to update the network weights. To minimize the impact of data imbalance on detection performance, focal loss is used. Besides, grid search is used to optimize the hyper-parameters of the model. In the second stage, with FPR as the optimization objective, the particle swarm optimization (PSO) algorithm is used to train the network. Finally, the proposed LFPR-DNN is verified by using the open Irish data set. Compared to other state-of-the-art classifiers, LFPR-DNN has the lowest FPR with 0.29% and the highest AUC with 99.42%. The FPR is reduced by an order of magnitude, which verifies the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

In conventional multi-user multiple-input multiple-output (MU-MIMO) systems with frequency division duplexing (FDD), channel acquisition and precoder optimization processes have been designed separately although they are highly coupled. This paper studies an end-to-end design of downlink MU-MIMO systems which include pilot sequences, limited feedback, and precoding. To address this problem, we propose a novel deep learning (DL) framework which jointly optimizes the feedback information generation at users and the precoder design at a base station (BS). Each procedure in the MU-MIMO systems is replaced by intelligently designed multiple deep neural networks (DNN) units. At the BS, a neural network generates pilot sequences and helps the users obtain accurate channel state information. At each user, the channel feedback operation is carried out in a distributed manner by an individual user DNN. Then, another BS DNN collects feedback information from the users and determines the MIMO precoding matrices. A joint training algorithm is proposed to optimize all DNN units in an end-to-end manner. In addition, a training strategy which can avoid retraining for different network sizes for a scalable design is proposed. Numerical results demonstrate the effectiveness of the proposed DL framework compared to classical optimization techniques and other conventional DNN schemes. [ABSTRACT FROM AUTHOR]

While early AutoML frameworks focused on optimizing traditional ML pipelines and their hyperparameters, a recent trend in AutoML is to focus on neural architecture search. In this paper, we introduce Auto-PyTorch, which brings together the best of these two worlds by jointly and robustly optimizing the network architecture and the training hyperparameters to enable fully automated deep learning (AutoDL). Auto-PyTorch achieves state-of-the-art performance on several tabular benchmarks by combining multi-fidelity optimization with portfolio construction for warmstarting and ensembling of deep neural networks (DNNs) and common baselines for tabular data. To thoroughly study our assumptions on how to design such an AutoDL system, we additionally introduce a new benchmark on learning curves for DNNs, dubbed LCBench, and run extensive ablation studies of the full Auto-PyTorch on typical AutoML benchmarks, eventually showing that Auto-PyTorch performs better than several state-of-the-art competitors. [ABSTRACT FROM AUTHOR]

Online fault detection of insulators is a necessary requirement for the development of a smart grid, which directly affects the safety and reliability of power system operations. Breakage is an important factor causing insulator abnormalities, and a large number of image samples are often required for training to complete defect identification with the help of AI deep learning. In this paper, we propose a convolutional neural network ensemble implemented using small samples, data augmentation and transfer learning for insulator breakage detection. We used data augmentation to expand the small sample of 200 insulator images by 15 times to 3000 images, built a VGG16 transfer network, pre-trained the VGG16 network convolutional basis weights, and improved the training performance and efficiency of the model with freeze training and fine-tuning training. The training on 2700 images and testing results on 300 images show that the proposed method can achieve an accuracy of 0.9871 for insulator breakage recognition. [ABSTRACT FROM AUTHOR]

The analysis of over-parameterized neural networks has drawn significant attention in recent years. It was shown that such systems behave like convex systems under various restricted settings, such as for two-layer neural networks, and when learning is only restricted locally in the so-called neural tangent kernel space around specialized initializations. However, there is a lack of powerful theoretical techniques that can analyze fully trained deep neural networks under general conditions. This paper considers this fundamental problem by investigating such overparameterized deep neural networks when fully trained. Specifically, we characterize a deep neural network by its features’ distributions and propose a metric to intuitively measure the usefulness of feature representations. Under certain regularizers that bounds the metric, we show deep neural networks can be reformulated as a convex optimization and the system can guarantee effective feature representations in terms of the metric. Our new analysis is more consistent with empirical observations that deep neural networks are capable of learning efficient feature representations. Empirical studies confirm that predictions of our theory are consistent with results observed in practice. [ABSTRACT FROM AUTHOR]

Point cloud based retrieval for place recognition is still a challenging problem since the drastic appearance changes of scenes due to seasonal or artificial changes in the environments. Existing deep learning based global descriptors for the retrieval task usually consume a large amount of computational resources ($e.g$., memory), which may not be suitable for the cases of limited hardware resources. In this paper, we develop an efficient point cloud learning network (EPC-Net) to generate global descriptors of point clouds for place recognition. While obtaining good performance, it can greatly reduce computational memory and inference time. First, we propose a lightweight but effective neural network module, called ProxyConv, to aggregate the local geometric features of point clouds. We leverage the adjacency matrix and proxy points to simplify the original edge convolution for lower memory consumption. Then, we design a lightweight grouped VLAD network to form global descriptors for retrieval. Compared with the original VLAD network, we propose a grouped fully connected layer to decompose the high-dimensional vectors into a group of low-dimensional vectors, which can reduce the number of parameters of the network and maintain the discrimination of the feature vector. Finally, we further develop a simple version of EPC-Net, called EPC-Net-L, which consists of two ProxyConv modules and one max pooling layer to aggregate global descriptors. By distilling the knowledge from EPC-Net, EPC-Net-L can obtain discriminative global descriptors for retrieval. Extensive experiments on the Oxford dataset and three in-house datasets demonstrate that our method achieves good results with lower parameters, FLOPs, GPU memory, and shorter inference time. Our code is available at https://github.com/fpthink/EPC-Net. [ABSTRACT FROM AUTHOR]

Deep learning-based approaches to delineating 3D structure depend on accurate annotations to train the networks. Yet in practice, people, no matter how conscientious, have trouble precisely delineating in 3D and on a large scale, in part because the data is often hard to interpret visually and in part because the 3D interfaces are awkward to use. In this paper, we introduce a method that explicitly accounts for annotation inaccuracies. To this end, we treat the annotations as active contour models that can deform themselves while preserving their topology. This enables us to jointly train the network and correct potential errors in the original annotations. The result is an approach that boosts performance of deep networks trained with potentially inaccurate annotations. [ABSTRACT FROM AUTHOR]

A system to perform video analytics is proposed using a dynamically tuned convolutional network. Videos are fetched from cloud storage, preprocessed, and a model for supporting classification is developed on these video streams using cloud-based infrastructure. A key focus in this paper is on tuning hyper-parameters associated with the deep learning algorithm used to construct the model. We further propose an automatic video object classification pipeline to validate the system. The mathematical model used to support hyper-parameter tuning improves performance of the proposed pipeline, and outcomes of various parameters on system’s performance is compared. Subsequently, the parameters that contribute toward the most optimal performance are selected for the video object classification pipeline. Our experiment-based validation reveals an accuracy and precision of 97% and 96%, respectively. The system proved to be scalable, robust, and customizable for a variety of different applications. [ABSTRACT FROM AUTHOR]

Neural networks (NNs) have become the go-to tool for solving many real-world recognition and classification tasks with massive and complex data sets. These networks require large data sets for training, which is usually performed on GPUs and CPUs in either a cloud or edge computing setting. No matter where the training is performed, it is subject to tight power/energy and data storage/transfer constraints. While these issues can be mitigated by replacing SRAM/DRAM with nonvolatile memories (NVMs) which offer near-zero leakage power and high scalability, the massive weight updates performed during training shorten NVM endurance and engender high write energy. In this paper, an NVM-friendly NN training approach is proposed. Weight update is redesigned to reduce bit flips in NVM cells. Moreover, two techniques, namely, filter exchange and bitwise rotation, are proposed to respectively balance writes to different weights and to different bits of one weight. The proposed techniques are integrated and evaluated in Caffe. Experimental results show significant power savings and endurance improvements, while maintaining high inference accuracy. [ABSTRACT FROM AUTHOR]

Limited angle reconstruction is a typical ill-posed problem in computed tomography (CT). Given incomplete projection data, images reconstructed by conventional analytical algorithms and iterative methods suffer from severe structural distortions and artifacts. In this paper, we proposed a self-augmented multi-stage deep-learning network (Sam’s Net) for end-to-end reconstruction of limited angle CT. With the merit of the alternating minimization technique, Sam’s Net integrates multi-stage self-constraints into cross-domain optimization to provide additional constraints on the manifold of neural networks. In practice, a sinogram completion network (SCNet) and artifact suppression network (ASNet), together with domain transformation layers constitute the backbone for cross-domain optimization. An online self-augmentation module was designed following the manner defined by alternating minimization, which enables a self-augmented learning procedure and multi-stage inference manner. Besides, a substitution operation was applied as a hard constraint for the solution space based on the data fidelity and a learnable weighting layer was constructed for data consistency refinement. Sam’s Net forms a new framework for ill-posed reconstruction problems. In the training phase, the self-augmented procedure guides the optimization into a tightened solution space with enriched diverse data distribution and enhanced data consistency. In the inference phase, multi-stage prediction can improve performance progressively. Extensive experiments with both simulated and practical projections under 90-degree and 120-degree fan-beam configurations validate that Sam’s Net can significantly improve the reconstruction quality with high stability and robustness. [ABSTRACT FROM AUTHOR]

Fast and accurate MRI image reconstruction from undersampled data is crucial in clinical practice. Deep learning based reconstruction methods have shown promising advances in recent years. However, recovering fine details from undersampled data is still challenging. In this paper, we introduce a novel deep learning based method, Pyramid Convolutional RNN (PC-RNN), to reconstruct images from multiple scales. Based on the formulation of MRI reconstruction as an inverse problem, we design the PC-RNN model with three convolutional RNN (ConvRNN) modules to iteratively learn the features in multiple scales. Each ConvRNN module reconstructs images at different scales and the reconstructed images are combined by a final CNN module in a pyramid fashion. The multi-scale ConvRNN modules learn a coarse-to-fine image reconstruction. Unlike other common reconstruction methods for parallel imaging, PC-RNN does not employ coil sensitive maps for multi-coil data and directly model the multiple coils as multi-channel inputs. The coil compression technique is applied to standardize data with various coil numbers, leading to more efficient training. We evaluate our model on the fastMRI knee and brain datasets and the results show that the proposed model outperforms other methods and can recover more details. The proposed method is one of the winner solutions in the 2019 fastMRI competition. [ABSTRACT FROM AUTHOR]

When imaging through a semi-reflective medium such as glass, the reflection of another scene can often be found in the captured images. It degrades the quality of the images and affects their subsequent analyses. In this paper, a novel deep neural network approach for solving the reflection problem in imaging is presented. Traditional reflection removal methods not only require long computation time for solving different optimization functions, their performance is also not guaranteed. As array cameras are readily available in nowadays imaging devices, we first suggest in this paper a multiple-image based depth estimation method using a convolutional neural network (CNN). The proposed network avoids the depth ambiguity problem due to the reflection in the image, and directly estimates the depths along the image edges. They are then used to classify the edges as belonging to the background or reflection. Since edges having similar depth values are error prone in the classification, they are removed from the reflection removal process. We suggest a generative adversarial network (GAN) to regenerate the removed background edges. Finally, the estimated background edge map is fed to another auto-encoder network to assist the extraction of the background from the original image. Experimental results show that the proposed reflection removal algorithm achieves superior performance both quantitatively and qualitatively as compared to the state-of-the-art methods. The proposed algorithm also shows much faster speed compared to the existing approaches using the traditional optimization methods. [ABSTRACT FROM AUTHOR]

Edge-preserving image filtering is an essential task in computational photography and imaging. In this paper, we propose a simple yet effective global edge-preserving filter based on soft clustering, and we propose a novel soft clustering algorithm based on a restricted Gaussian mixture model. Given specified parameters, the soft clustering process is firstly performed on the image to derive the partition matrix, from which the affinity matrix is then constructed for filtering. The filtering output is calculated as the weighted average of the pixels in the local window, so the proposed filter could suppress the intensity shift artifacts that impede most global filters. Besides, the weights in the proposed filter are derived by clustering, which properly separates dissimilar pixels, so the proposed filter could handle the halo artifacts that haunt many local filters. Moreover, our filter provides flexible control over the amount of smoothing that is deficient in the deep learning-based filters. Besides the efficacy in smoothing, the proposed filter naturally has low computational complexity. Qualitative and quantitative results suggest that the proposed filter benefits various applications, including edge-preserving smoothing, image enhancing, flash/non-flash fusion, HDR tone mapping, and dehazing. [ABSTRACT FROM AUTHOR]

For vehicle-to-network communications, handover (HO) management enables vehicles to maintain the connection with the network while transiting through coverage areas of different base stations (BSs). However, the high mobility of vehicles means shorter connection periods with each BS that leads to frequent HOs, hence raises the necessity for optimal HO decision making for high quality infotainment services. Machine learning is capable of capturing underlying pattern via data driven methods to find optimal solutions to complex problems, and much learning-based HO optimization research has been conducted focusing on specific network setups. However, attention still needs to be paid to the actual deployment aspect and standardized datasets or simulation environments for evaluation. This paper proposes a deep reinforcement learning-based HO algorithm using the input parameters that are configurable in the existing measurement report of cellular networks. The performance of the proposed algorithm is evaluated using the well-known network simulator ns-3 with its official LTE module. A realistic network setup in the city center of Glasgow (U.K.) is configured with vehicle trajectories generated by the routes mobility model using Google Maps Directions API. Evaluation results reveal that the proposed algorithm significantly outperforms the A3 RSRP baseline with an average of 25.72% packet loss reduction per HO, suggesting significant improvement in quality of service of phone call and video streaming, etc. The proposed algorithm also has a small implementation cost compared to some state-of-the-art and should be deployed by a software update to a local BS controller. [ABSTRACT FROM AUTHOR]

Robust visual tracking is one of the most challenging problems in computer vision applications. However, the limited training data and the computational complexity have severely affected tracking performance. In this paper, we propose an auto-encoder pair model for visual tracking which is composed of source domain network and target domain network to help a more accurate localization. We adopt the dense circular samples of the object state to increase the number of training samples and prevent model overfitting. Meanwhile, a difference regularization term is also introduced into our framework to penalize the large appearance variations of the object in two domains. The alternating optimization is used to solve the optimization problems. Furthermore, our method alleviates the model update problem and improves the tracking speed by using long-term and short-term updating scheme. In addition, the target domain filter is updated by introducing the updated source domain filter to avoid the object drift. Comprehensive experiments on some challenging benchmarks demonstrate that our approach concurrently improves both tracking accuracy and speed. [ABSTRACT FROM AUTHOR]

Demand-side management (DSM) increases the complexity of forecasting environment, which makes traditional forecasting methods difficult to meet the firm's need for predictive accuracy. Since deep learning can comprehensively consider various factors to improve prediction results, this paper improves the deep belief network from three aspects of input data, model and performance, and uses it to solve the short-term load forecasting problem in DSM. In the data optimization stage, the Hankel matrix is constructed to increase the input weight of DSM data, and the gray relational analysis is used to select strongly correlated data from the data set. In the model optimization stage, the Gauss-Bernoulli restricted Boltzmann machine is used as the first restricted Boltzmann machine of the deep network to convert the continuity feature of input data into binomial distribution feature. In the performance optimization stage, a pre-training method combining error constraint and unsupervised learning is proposed to provide good initial parameters, and the global fine-tuning of network parameters is realized based on the genetic algorithm. Based on the actual data of Tianjin Power Grid in China, the experimental results show that the proposed method is superior to other methods. [ABSTRACT FROM AUTHOR]

Recent years have witnessed the promising future of hashing in the industrial applications for fast similarity retrieval. In this paper, we propose a novel supervised hashing method for large-scale cross-media search, termed self-supervised deep multimodal hashing (SSDMH), which learns unified hash codes as well as deep hash functions for different modalities in a self-supervised manner. With the proposed regularized binary latent model, unified binary codes can be solved directly without relaxation strategy while retaining the neighborhood structures by the graph regularization term. Moreover, we propose a new discrete optimization solution, termed as binary gradient descent, which aims at improving the optimization efficiency toward real-time operation. Extensive experiments on three benchmark data sets demonstrate the superiority of SSDMH over state-of-the-art cross-media hashing approaches. [ABSTRACT FROM AUTHOR]

We study the dual problem of image super-resolution (SR), which we term image compact-resolution (CR). Opposite to image SR that hallucinates a visually plausible high-resolution image given a low-resolution input, image CR provides a low-resolution version of a high-resolution image, such that the low-resolution version is both visually pleasing and as informative as possible compared to the high-resolution image. We propose a convolutional neural network (CNN) for image CR, namely, CNN-CR, inspired by the great success of CNN for image SR. Specifically, we translate the requirements of image CR into operable optimization targets for training CNN-CR: the visual quality of the compact resolved image is ensured by constraining its difference from a naively downsampled version and the information loss of image CR is measured by upsampling/super-resolving the compact-resolved image and comparing that to the original image. Accordingly, CNN-CR can be trained either separately or jointly with a CNN for image SR. We explore different training strategies as well as different network structures for CNN-CR. Our experimental results show that the proposed CNN-CR clearly outperforms simple bicubic downsampling and achieves on average 2.25 dB improvement in terms of the reconstruction quality on a large collection of natural images. We further investigate two applications of image CR, i.e., low-bit-rate image compression and image retargeting. Experimental results show that the proposed CNN-CR helps achieve significant bits saving than High Efficiency Video Coding when applied to image compression and produce visually pleasing results when applied to image retargeting. [ABSTRACT FROM AUTHOR]

Advanced algorithms can promote the development of energy management strategies (EMSs) as a key technology in hybrid electric vehicles (HEVs). Reinforcement learning (RL) with distributed structure can significantly improve training efficiency in complex environments, and multi-threaded parallel computing provides a reliable algorithm basis for promoting adaptability. Dedicated to trying more efficient deep reinforcement learning (DRL) algorithms, this paper proposed a deep q-network (DQN)-based energy and emission management strategy (E&EMS) at first. Then, two distributed DRL algorithms, namely asynchronous advantage actor-critic (A3C) and distributed proximal policy optimization (DPPO), were adopted to propose EMSs, respectively. Finally, emission optimization was taken into account and then distributed DRL-based E&EMSs were proposed. Regarding dynamic programming (DP) as the optimal benchmark, simulation results show that three DRL-based control strategies can achieve near-optimal fuel economy and outstanding computational efficiency, and compared with DQN, two distributed DRL algorithms have improved the learning efficiency by four times. [ABSTRACT FROM AUTHOR]

In this paper, we consider wireless-powered secure communication with an energy harvesting receiver, which is allowed to harvest energy from the transmitted signals but not to decode information, and there is thus a requirement to keep the information secret from this potential eavesdropper. Considering co-channel interference among signal links, we find the optimal transmit power to maximize the sum rate of the signal links, while ensuring the requirements of information secrecy and energy harvesting. Due to the non-convexity of the optimization problem formulated here, we first derive suboptimal solutions using an iterative algorithm based on a dual method. In order to address the limitations caused by the use of the iterative algorithm, i.e., long computation time and suboptimality, we design an efficient deep neural network (DNN) framework and a novel training strategy as a means of combining supervised and unsupervised training. Specifically, the DNN is first pre-trained using labeled training data with the suboptimal solutions obtained from the iterative algorithm in a supervised manner; further training is then applied to the DNN using a well designed loss function in an unsupervised manner to enhance the training performance. Simulation results reveal that the proposed scheme achieves a near-optimal performance with a lower computation time than existing schemes. We also verify that the pre-training and the new loss function are effective in improving the speed of training of the DNN. [ABSTRACT FROM AUTHOR]

Many classical Computer Vision problems, such as essential matrix computation and pose estimation from 3D to 2D correspondences, can be tackled by solving a linear least-square problem, which can be done by finding the eigenvector corresponding to the smallest, or zero, eigenvalue of a matrix representing a linear system. Incorporating this in deep learning frameworks would allow us to explicitly encode known notions of geometry, instead of having the network implicitly learn them from data. However, performing eigendecomposition within a network requires the ability to differentiate this operation. While theoretically doable, this introduces numerical instability in the optimization process in practice. In this paper, we introduce an eigendecomposition-free approach to training a deep network whose loss depends on the eigenvector corresponding to a zero eigenvalue of a matrix predicted by the network. We demonstrate that our approach is much more robust than explicit differentiation of the eigendecomposition using two general tasks, outlier rejection and denoising, with several practical examples including wide-baseline stereo, the perspective-n-point problem, and ellipse fitting. Empirically, our method has better convergence properties and yields state-of-the-art results. [ABSTRACT FROM AUTHOR]

In this paper, we introduce the algorithms of Orthogonal Deep Neural Networks (OrthDNNs) to connect with recent interest of spectrally regularized deep learning methods. OrthDNNs are theoretically motivated by generalization analysis of modern DNNs, with the aim to find solution properties of network weights that guarantee better generalization. To this end, we first prove that DNNs are of local isometry on data distributions of practical interest; by using a new covering of the sample space and introducing the local isometry property of DNNs into generalization analysis, we establish a new generalization error bound that is both scale- and range-sensitive to singular value spectrum of each of networks’ weight matrices. We prove that the optimal bound w.r.t. the degree of isometry is attained when each weight matrix has a spectrum of equal singular values, among which orthogonal weight matrix or a non-square one with orthonormal rows or columns is the most straightforward choice, suggesting the algorithms of OrthDNNs. We present both algorithms of strict and approximate OrthDNNs, and for the later ones we propose a simple yet effective algorithm called Singular Value Bounding (SVB), which performs as well as strict OrthDNNs, but at a much lower computational cost. We also propose Bounded Batch Normalization (BBN) to make compatible use of batch normalization with OrthDNNs. We conduct extensive comparative studies by using modern architectures on benchmark image classification. Experiments show the efficacy of OrthDNNs. [ABSTRACT FROM AUTHOR]

In this paper we consider maximum-likelihood (ML) MIMO detection under one-bit quantized observations and binary symbol constellations. This problem is motivated by the recent interest in adopting coarse quantization in massive MIMO systems—as an effective way to scale down the hardware complexity and energy consumption. Classical MIMO detection techniques consider unquantized observations, and many of them are not applicable to the one-bit MIMO case. We develop a new non-convex optimization algorithm for the one-bit ML MIMO detection problem, using a strategy called homotopy optimization. The idea is to transform the ML problem into a sequence of approximate problems, from easy (convex) to hard (close to ML), and with each problem being a gradual modification of its previous. Then, our attempt is to iteratively trace the solution path of these approximate problems. This homotopy algorithm is well suited to the application of deep unfolding, a recently popular approach for turning certain model-based algorithms into data-driven, and performance enhanced, ones. While our initial focus is on one-bit MIMO detection, the proposed technique also applies naturally to the classical unquantized MIMO detection. We performed extensive simulations and show that the proposed homotopy algorithms, both non-deep and deep, have satisfactory bit-error probability performance compared to many state-of-the-art algorithms. Also, the deep homotopy algorithm has attractively low computational complexity. [ABSTRACT FROM AUTHOR]

Minimally invasive procedures rely on image guidance for navigation at the operation site to avoid large surgical incisions. X-ray images are often used for guidance, but important structures may be not well visible. These structures can be overlaid from pre-operative 3-D images and accurate alignment can be established using 2-D/3-D registration. Registration based on the point-to-plane correspondence model was recently proposed and shown to achieve state-of-the-art performance. However, registration may still fail in challenging cases due to a large portion of outliers. In this paper, we describe a learning-based correspondence weighting scheme to improve the registration performance. By learning an attention model, inlier correspondences get higher attention in the motion estimation while the outlier correspondences are suppressed. Instead of using per-correspondence labels, our objective function allows to train the model directly by minimizing the registration error. We demonstrate a highly increased robustness, e.g. increasing the success rate from 84.9% to 97.0% for spine registration. In contrast to previously proposed learning-based methods, we also achieve a high accuracy of around 0.5mm mean re-projection distance. In addition, our method requires a relatively small amount of training data, is able to learn from simulated data, and generalizes to images with additional structures which are not present during training. Furthermore, a single model can be trained for both, different views and different anatomical structures. [ABSTRACT FROM AUTHOR]

A deep neural network of multiple nonlinear layers forms a large function space, which can easily lead to overfitting when it encounters small-sample data. To mitigate overfitting in small-sample classification, learning more discriminative features from small-sample data is becoming a new trend. To this end, this paper aims to find a subspace of neural networks that can facilitate a large decision margin. Specifically, we propose the Orthogonal Softmax Layer (OSL), which makes the weight vectors in the classification layer remain orthogonal during both the training and test processes. The Rademacher complexity of a network using the OSL is only $\frac {1}{K}$ , where $K$ is the number of classes, of that of a network using the fully connected classification layer, leading to a tighter generalization error bound. Experimental results demonstrate that the proposed OSL has better performance than the methods used for comparison on four small-sample benchmark datasets, as well as its applicability to large-sample datasets. Codes are available at: https://github.com/dongliangchang/OSLNet. [ABSTRACT FROM AUTHOR]

Deep learning has been successfully introduced for 2D-image denoising, but it is still unsatisfactory for hyperspectral image (HSI) denoising due to the unacceptable computational complexity of the end-to-end training process and the difficulty of building a universal 3D-image training dataset. In this paper, instead of developing an end-to-end deep learning denoising network, we propose an HSI denoising framework for the removal of mixed Gaussian impulse noise, in which the denoising problem is modeled as a convolutional neural network (CNN) constrained non-negative matrix factorization problem. Using the proximal alternating linearized minimization, the optimization can be divided into three steps: the update of the spectral matrix, the update of the abundance matrix, and the estimation of the sparse noise. Then, we design the CNN architecture and proposed two training schemes, which can allow the CNN to be trained with a 2D-image dataset. Compared with the state-of-the-art denoising methods, the proposed method has a relatively good performance on the removal of the Gaussian and mixed Gaussian impulse noises. More importantly, the proposed model can be only trained once by a 2D-image dataset but can be used to denoise HSIs with different numbers of channel bands. [ABSTRACT FROM AUTHOR]

This paper considers the design of optimal resource allocation policies in wireless communication systems, which are generically modeled as a functional optimization problem with stochastic constraints. These optimization problems have the structure of a learning problem in which the statistical loss appears as a constraint, motivating the development of learning methodologies to attempt their solution. To handle stochastic constraints, training is undertaken in the dual domain. It is shown that this can be done with small loss of optimality when using near-universal learning parameterizations. In particular, since deep neural networks (DNNs) are near universal, their use is advocated and explored. DNNs are trained here with a model-free primal-dual method that simultaneously learns a DNN parameterization of the resource allocation policy and optimizes the primal and dual variables. Numerical simulations demonstrate the strong performance of the proposed approach on a number of common wireless resource allocation problems. [ABSTRACT FROM AUTHOR]

Neural networks with rectified linear unit (ReLU) activation functions (a.k.a. ReLU networks) have achieved great empirical success in various domains. Nonetheless, existing results for learning ReLU networks either pose assumptions on the underlying data distribution being, e.g., Gaussian, or require the network size and/or training size to be sufficiently large. In this context, the problem of learning a two-layer ReLU network is approached in a binary classification setting, where the data are linearly separable and a hinge loss criterion is adopted. Leveraging the power of random noise perturbation, this paper presents a novel stochastic gradient descent (SGD) algorithm, which can provably train any single-hidden-layer ReLU network to attain global optimality, despite the presence of infinitely many bad local minima, maxima, and saddle points in general. This result is the first of its kind, requiring no assumptions on the data distribution, training/network size, or initialization. Convergence of the resultant iterative algorithm to a global minimum is analyzed by establishing both an upper bound and a lower bound on the number of non-zero updates to be performed. Moreover, generalization guarantees are developed for ReLU networks trained with the novel SGD leveraging classic compression bounds. These guarantees highlight a key difference (at least in the worst case) between reliably learning a ReLU network as well as a leaky ReLU network in terms of sample complexity. Numerical tests using both synthetic data and real images validate the effectiveness of the algorithm and the practical merits of the theory. [ABSTRACT FROM AUTHOR]

Hashing has been widely used for large-scale search due to its low storage cost and fast query speed. By using supervised information, supervised hashing can significantly outperform unsupervised hashing. Recently, discrete supervised hashing and feature learning based deep hashing are two representative progresses in supervised hashing. On one hand, hashing is essentially a discrete optimization problem. Hence, utilizing supervised information to directly guide discrete (binary) coding procedure can avoid sub-optimal solution and improve the accuracy. On the other hand, feature learning based deep hashing, which integrates deep feature learning and hash-code learning into an end-to-end architecture, can enhance the feedback between feature learning and hash-code learning. The key in discrete supervised hashing is to adopt supervised information to directly guide the discrete coding procedure in hashing. The key in deep hashing is to adopt the supervised information to directly guide the deep feature learning procedure. However, most deep supervised hashing methods cannot use the supervised information to directly guide both discrete (binary) coding procedure and deep feature learning procedure in the same framework. In this paper, we propose a novel deep hashing method, called deep discrete supervised hashing (DDSH). DDSH is the first deep hashing method which can utilize pairwise supervised information to directly guide both discrete coding procedure and deep feature learning procedure and thus enhance the feedback between these two important procedures. Experiments on four real datasets show that DDSH can outperform other state-of-the-art baselines, including both discrete hashing and deep hashing baselines, for image retrieval. [ABSTRACT FROM AUTHOR]

Many distributed deep learning systems have been published over the past few years, often accompanied by impressive performance claims. In practice these figures are often achieved in high performance computing (HPC) environments with fast InfiniBand network connections. For average deep learning practitioners this is usually an unrealistic scenario, since they cannot afford access to these facilities. Simple re-implementations of algorithms such as EASGD  for standard Ethernet environments often fail to replicate the scalability and performance of the original works. In this paper, we explore this particular problem domain and present MPCA SGD, a method for distributed training of deep neural networks that is specifically designed to run in low-budget environments. MPCA SGD tries to make the best possible use of available resources, and can operate well if network bandwidth is constrained. Furthermore, MPCA SGD runs on top of the popular Apache Spark  framework. Thus, it can easily be deployed in existing data centers and office environments where Spark is already used. When training large deep learning models in a gigabit Ethernet cluster, MPCA SGD achieves significantly faster convergence rates than many popular alternatives. For example, MPCA SGD can train ResNet-152  up to 5.3x faster than state-of-the-art systems like MXNet  , up to 5.3x faster than bulk-synchronous systems like SparkNet  and up to 5.3x faster than decentral asynchronous systems like EASGD . [ABSTRACT FROM AUTHOR]

How do we retrieve images accurately? Also, how do we rank a group of images precisely and efficiently for specific queries? These problems are critical for researchers and engineers to generate a novel image searching engine. First, it is important to obtain an appropriate description that effectively represent the images. In this paper, multimodal features are considered for describing images. The images unique properties are reflected by visual features, which are correlated to each other. However, semantic gaps always exist between images visual features and semantics. Therefore, we utilize click feature to reduce the semantic gap. The second key issue is learning an appropriate distance metric to combine these multimodal features. This paper develops a novel deep multimodal distance metric learning (Deep-MDML) method. A structured ranking model is adopted to utilize both visual and click features in distance metric learning (DML). Specifically, images and their related ranking results are first collected to form the training set. Multimodal features, including click and visual features, are collected with these images. Next, a group of autoencoders is applied to obtain initially a distance metric in different visual spaces, and an MDML method is used to assign optimal weights for different modalities. Next, we conduct alternating optimization to train the ranking model, which is used for the ranking of new queries with click features. Compared with existing image ranking methods, the proposed method adopts a new ranking model to use multimodal features, including click features and visual features in DML. We operated experiments to analyze the proposed Deep-MDML in two benchmark data sets, and the results validate the effects of the method. [ABSTRACT FROM PUBLISHER]

The generalization error of deep neural networks via their classification margin is studied in this paper. Our approach is based on the Jacobian matrix of a deep neural network and can be applied to networks with arbitrary nonlinearities and pooling layers, and to networks with different architectures such as feed forward networks and residual networks. Our analysis leads to the conclusion that a bounded spectral norm of the network's Jacobian matrix in the neighbourhood of the training samples is crucial for a deep neural network of arbitrary depth and width to generalize well. This is a significant improvement over the current bounds in the literature, which imply that the generalization error grows with either the width or the depth of the network. Moreover, it shows that the recently proposed batch normalization and weight normalization reparametrizations enjoy good generalization properties, and leads to a novel network regularizer based on the network's Jacobian matrix. The analysis is supported with experimental results on the MNIST, CIFAR-10, LaRED, and ImageNet datasets. [ABSTRACT FROM PUBLISHER]

The task of labeling samples is demanding and expensive. Active learning aims to generate the smallest possible training data set that results in a classifier with high performance in the test phase. It usually consists of two steps of selecting a set of queries and requesting their labels. Among the suggested objectives to score the query sets, information theoretic measures have become very popular. Yet among them, those based on Fisher information (FI) have the advantage of considering the diversity among the queries and tractable computations. In this work, we provide a practical algorithm based on Fisher information ratio to obtain query distribution for a general framework where, in contrast to the previous FI-based querying methods, we make no assumptions over the test distribution. The empirical results on synthetic and real-world data sets indicate that this algorithm gives competitive results. [ABSTRACT FROM AUTHOR]

Domain generalization aims to learn knowledge invariant across different distributions while semantically meaningful for downstream tasks from multiple source domains, to improve the model’s generalization ability on unseen target domains. The fundamental objective is to understand the underlying ”invariance” behind these observational distributions and such invariance has been shown to have a close connection to causality. While many existing approaches make use of the property that causal features are invariant across domains, we consider the invariance of the average causal effect of the features to the labels. This invariance regularizes our training approach in which interventions are performed on features to enforce stability of the causal prediction by the classifier across domains. Our work thus sheds some light on the domain generalization problem by introducing invariance of the mechanisms into the learning process. Experiments on several benchmark datasets demonstrate the performance of the proposed method against SOTAs. The codes are available at: https://github.com/lithostark/Contrastive-ACE. [ABSTRACT FROM AUTHOR]

This work proposes to develop and evaluate a deep learning framework that jointly optimizes k-t sampling patterns and reconstruction for head and neck dynamic contrast-enhanced (DCE) MRI aiming to reduce bias and uncertainty of pharmacokinetic (PK) parameter estimation. 2D Cartesian phase encoding k-space subsampling patterns for a 3D spoiled gradient recalled echo (SPGR) sequence along a time course of DCE MRI were jointly optimized in a deep learning-based dynamic MRI reconstruction network by a loss function concerning both reconstruction image quality and PK parameter estimation accuracy. During training, temporal k-space data sharing scheme was optimized as well. The proposed method was trained and tested by multi-coil complex digital reference objects of DCE images (mcDROs). The PK parameters estimated by the proposed method were compared with two published iterative DCE MRI reconstruction schemes using normalized root mean squared errors (NRMSEs) and Bland-Altman analysis at temporal resolutions of $\Delta \text{t}$ = 2s, 3s, 4s, and 5s, which correspond to undersampling rates of R = 50, 34, 25, and 20. The proposed method achieved low PK parameter NRMSEs at all four temporal resolutions compared with the benchmark methods on testing mcDROs. The Bland-Altman plots demonstrated that the proposed method reduced PK parameter estimation bias and uncertainty in tumor regions at temporal resolution of 2s. The proposed method also showed robustness to contrast arrival timing variations across patients. This work provides a potential way to increase PK parameter estimation accuracy and precision, and thus facilitate the clinical translation of DCE MRI. [ABSTRACT FROM AUTHOR]

Processing-in-memory (PIM) is a promising technique to accelerate deep learning (DL) workloads. Emerging DL workloads (e.g., ResNet with 152 layers) consist of millions of parameters, which increase the area and fabrication cost of monolithic PIM accelerators. The fabrication cost challenge can be addressed by 2.5-D systems integrating multiple PIM chiplets connected through a network-on-package (NoP). However, server-scale scenarios simultaneously execute multiple compute-heavy DL workloads, leading to significant interchiplet data volume. State-of-the-art NoP architectures proposed in the literature do not consider the nature of DL workloads. In this article, we propose a novel server scale 2.5-D manycore architecture called SWAP that accounts for the traffic characteristics of DL applications. Comprehensive experimental evaluations with different system sizes as well as diverse emerging DL workloads demonstrate that SWAP achieves significant performance and energy consumption improvements with much lower fabrication cost than state-of-the-art NoP topologies. [ABSTRACT FROM AUTHOR]

The field of meta-learning, or learning-to-learn, has seen a dramatic rise in interest in recent years. Contrary to conventional approaches to AI where tasks are solved from scratch using a fixed learning algorithm, meta-learning aims to improve the learning algorithm itself, given the experience of multiple learning episodes. This paradigm provides an opportunity to tackle many conventional challenges of deep learning, including data and computation bottlenecks, as well as generalization. This survey describes the contemporary meta-learning landscape. We first discuss definitions of meta-learning and position it with respect to related fields, such as transfer learning and hyperparameter optimization. We then propose a new taxonomy that provides a more comprehensive breakdown of the space of meta-learning methods today. We survey promising applications and successes of meta-learning such as few-shot learning and reinforcement learning. Finally, we discuss outstanding challenges and promising areas for future research. [ABSTRACT FROM AUTHOR]

This work proposes a novel deep learning (DL) framework to solve the electromagnetic inverse scattering (EMIS) problems. The proposed framework integrates the complex-valued deep convolutional neural network (DConvNet) into the supervised descent method (SDM) to realize both off-line training and on-line “imaging” prediction for EMIS. The offline training consists of two parts: 1) DConvNet training: the training dataset is created, and the proposed DConvNet is trained to realize the EM forward process and 2) SDM training: the trained DConvNet is integrated into the SDM framework, and the average descent directions between the initial prediction and the true label of SDM iterative schemes are learned based on the same dataset in part 1). In the online step, the contrasts (permittivities) reconstruction of scatterers is realized by the SDM iteration process based on learned descent directions, while its forward process is achieved by the trained complex-valued DConvNet. Ultimately, this framework provides a new perspective to integrate the prior information into the EMIS solving process with the maintained accuracy. Unlike the conventional SDM, the novel proposed framework can significantly shorten the computation and realize the real-time imaging. Various numerical examples and discussions are provided to demonstrate the efficiency and accuracy of the proposed novel framework. [ABSTRACT FROM AUTHOR]

The popularity of deep learning techniques renewed the interest in neural architectures able to process complex structures that can be represented using graphs, inspired by Graph Neural Networks (GNNs). We focus our attention on the originally proposed GNN model of Scarselli et al. 2009, which encodes the state of the nodes of the graph by means of an iterative diffusion procedure that, during the learning stage, must be computed at every epoch, until the fixed point of a learnable state transition function is reached, propagating the information among the neighbouring nodes. We propose a novel approach to learning in GNNs, based on constrained optimization in the Lagrangian framework. Learning both the transition function and the node states is the outcome of a joint process, in which the state convergence procedure is implicitly expressed by a constraint satisfaction mechanism, avoiding iterative epoch-wise procedures and the network unfolding. Our computational structure searches for saddle points of the Lagrangian in the adjoint space composed of weights, nodes state variables and Lagrange multipliers. This process is further enhanced by multiple layers of constraints that accelerate the diffusion process. An experimental analysis shows that the proposed approach compares favourably with popular models on several benchmarks. [ABSTRACT FROM AUTHOR]