Search

The existence of thick clouds covers the comprehensive Earth observation of optical remote sensing images (RSIs). Cloud removal is an effective and economical preprocessing step to improve the subsequent applications of RSIs. Deep learning (DL)-based methods have attracted much attention and achieved state-of-the-art results. However, most of these methods suffer from the following issues: 1) ignore the physical characteristics of RSIs; 2) require paired images with/without cloud or extra auxiliary images; and 3) demand the cloud mask. These issues might have limited the flexibility of existing networks. In this article, we propose a novel low-rank regularized self-supervised network (LRRSSN) that couples model-driven and data-driven methods to remove the thick cloud from multitemporal RSIs (MRSIs). First, motivated by the equal importance of image and cloud components as well as their intrinsic characteristics, we decompose the observed image into low-rank image and structural sparse cloud components. In this way, we obtain a model-driven thick cloud removal method where the spectral–temporal low-rank correlation of the image component and the spectral structural sparsity of the cloud component are effectively exploited. Second, to capture the complex nonlinear features of different scenarios, the data-driven self-supervised network that does not require external training datasets is designed to explore the deep prior of the image component. Third, the coupled model-driven and data-driven LRRSSN is optimized by an efficient half quadratic splitting (HQS) algorithm. Finally, without knowing the exact cloud mask, we estimate the cloud mask to preserve information in cloud-free areas as much as possible. Experiments conducted in synthetic and real-world scenarios demonstrate the effectiveness of the proposed approach.

The great success of deep neural networks is built upon their over-parameterization, which smooths the optimization landscape without degrading the generalization ability. Despite the benefits of over-parameterization, a huge amount of parameters makes deep networks cumbersome in daily life applications. On the other hand, training neural networks without over-parameterization faces many practical problems, e.g., being trapped in the local optimal. Though techniques such as pruning and distillation are developed, they are expensive in fully training a dense network as backward selection methods; and there is still a void on systematically exploring forward selection methods for learning structural sparsity in deep networks. To fill in this gap, this paper proposes a new approach based on differential inclusions of inverse scale spaces. Specifically, our method can generate a family of models from simple to complex ones along the dynamics via coupling a pair of parameters, such that over-parameterized deep models and their structural sparsity can be explored simultaneously. This kind of differential inclusion scheme has a simple discretization, dubbed Deep structure splitting Linearized Bregman Iteration (DessiLBI), whose global convergence in learning deep networks could be established under the Kurdyka-Łojasiewicz framework. Particularly, we explore several applications of DessiLBI, including finding sparse structures of networks directly via the coupled structure parameter and growing networks from simple to complex ones progressively. Experimental evidence shows that our method achieves comparable and even better performance than the competitive optimizers in exploring the sparse structure of several widely used backbones on the benchmark datasets. Remarkably, with early stopping, our method unveils “winning tickets” in early epochs: the effective sparse network structures with comparable test accuracy to fully trained over-parameterized models, that are further transferable to similar alternative tasks. Furthermore, our method is able to grow networks efficiently with adaptive filter configurations, demonstrating the good performance with much less computational cost. Codes and models can be downloaded at https://github.com/DessiLBI2020/DessiLBI.

Rank aggregation with pairwise comparisons has shown promising results in elections, sports competitions, recommendations, and information retrieval. However, little attention has been paid to the security issue of such algorithms, in contrast to numerous research work on the computational and statistical characteristics. Driven by huge profit, the potential adversary has strong motivation and incentives to manipulate the ranking list. Meanwhile, the intrinsic vulnerability of the rank aggregation methods is not well studied in the literature. To fully understand the possible risks, we focus on the purposeful adversary who desires to designate the aggregated results by modifying the pairwise data in this paper. From the perspective of the dynamical system, the attack behavior with a target ranking list is a fixed point belonging to the composition of the adversary and the victim. To perform the targeted attack, we formulate the interaction between the adversary and the victim as a game-theoretic framework consisting of two continuous operators while Nash equilibrium is established. Then two procedures against HodgeRank and RankCentrality are constructed to produce the modification of the original data. Furthermore, we prove that the victims will produce the target ranking list once the adversary masters the complete information. It is noteworthy that the proposed methods allow the adversary only to hold incomplete information or imperfect feedback and perform the purposeful attack. The effectiveness of the suggested target attack strategies is demonstrated by a series of toy simulations and several real-world data experiments. These experimental results show that the proposed methods could achieve the attacker's goal in the sense that the leading candidate of the perturbed ranking list is the designated one by the adversary.

Low physical activity is a major risk factor for cardiovascular diseases (CVDs). This study aimed to estimate the global, regional, national, and sex-age-specific burden of CVDs attributed to low physical activity from 1990 to 2019.

Training neural networks is recently a hot topic in machine learning due to its great success in many applications. Since the neural networks’ training usually involves a highly nonconvex optimization problem, it is difficult to design optimization algorithms with perfect convergence guarantees to derive a neural network estimator of high quality. In this article, we borrow the well-known random sketching strategy from kernel methods to transform the training of shallow rectified linear unit (ReLU) nets into a linear least-squares problem. Using the localized approximation property of shallow ReLU nets and a recently developed dimensionality-leveraging scheme, we succeed in equipping shallow ReLU nets with a specific random sketching scheme. The efficiency of the suggested random sketching strategy is guaranteed by theoretical analysis and also verified via a series of numerical experiments. Theoretically, we show that the proposed random sketching is almost optimal in terms of both approximation capability and learning performance. This implies that random sketching does not degenerate the performance of shallow ReLU nets. Numerically, we show that random sketching can significantly reduce the computational burden of numerous backpropagation (BP) algorithms while maintaining their learning performance.

Abstract: In this paper, a novel method for synthetic aperture radar (SAR) imaging is proposed. The approach is based on L 1/2 regularization to reconstruct the scattering field, which optimizes a quadratic error term of the SAR observation process subject to the interested scene sparsity. Compared to the conventional SAR imaging technique, the new method implements SAR imaging effectively at much lower sampling rate than the Nyquist rate, and produces high-quality images with reduced sidelobes and increased resolution. Also, over the prevalent greedy pursuit and L 1 regularization based SAR imaging methods, there are remarkable performance improvements of the new method. On one hand, the new method significantly reduces the number of measurements needed for reconstruction, as supported by a phase transition diagram study. On the other hand, the new method is more robust to the observation noise. These fundamental properties of the new method are supported and demonstrated both by simulations and real SAR data experiments.

Abstract: The displaced phase center antenna (DPCA) technique is an effective strategy to achieve wide-swath synthetic aperture radar (SAR) imaging with high azimuth resolution. However, traditionally, it requires strict limitation of the pulse repetition frequency (PRF) to avoid non-uniform sampling. Otherwise, any deviation could bring serious ambiguity if the data are directly processed using a matched filter. To break this limitation, a recently proposed spectrum reconstruction method is capable of recovering the true spectrum from the nonuniform samples. However, the performance is sensitive to the selection of the PRF. Sparse regularization based imaging may provide a way to overcome this sensitivity. The existing time-domain method, however, requires a large-scale observation matrix to be built, which brings a high computational cost. In this paper, we propose a frequency domain method, called the iterative spectrum reconstruction method, through integration of the sparse regularization technique with spectrum analysis of the DPCA signal. By approximately expressing the observation in the frequency domain, which is realized via a series of decoupled linear operations, the method performs SAR imaging which is then not directly based on the observation matrix, which reduces the computational cost from O(N 2) to O(N log N) (where N is the number of range cells), and is therefore more efficient than the time domain method. The sparse regularization scheme, realized via a fast thresholding iteration, has been adopted in this method, which brings the robustness of the imaging process to the PRF selection. We provide a series of simulations and ground based experiments to demonstrate the high efficiency and robustness of the method. The simulations show that the new method is almost as fast as the traditional mono-channel algorithm, and works well almost independently of the PRF selection. Consequently, the suggested method can be accepted as a practical and efficient wide-swath SAR imaging technique.