Search

Real-world traffic sign recognition is an important step toward building autonomous vehicles, most of which highly dependent on deep neural networks (DNNs). Recent studies demonstrated that DNNs are surprisingly susceptible to adversarial examples. Many attack methods have been proposed to understand and generate adversarial examples, such as gradient-based attack, score-based attack, decision-based attack, and transfer-based attacks. However, most of these algorithms are ineffective in real-world road sign attack, because 1) iteratively learning perturbations for each frame is not realistic for a fast moving car and 2) most optimization algorithms traverse all pixels equally without considering their diverse contribution. To alleviate these problems, this article proposes the targeted attention attack (TAA) method for real-world road sign attack. Specifically, we have made the following contributions: 1) we leverage the soft attention map to highlight those important pixels and skip those zero-contributed areas—this also helps to generate natural perturbations; 2) we design an efficient universal attack that optimizes a single perturbation/noise based on a set of training images under the guidance of the pretrained attention map; 3) we design a simple objective function that can be easily optimized; and 4) we evaluate the effectiveness of TAA on real-world data sets. Experimental results validate that the TAA method improves the attack successful rate (nearly 10%) and reduces the perturbation loss (about a quarter) compared with the popular RP2 method. Additionally, our TAA also provides good properties, e.g., transferability and generalization capability. We provide code and data to ensure the reproducibility: https://github.com/AdvAttack/RoadSignAttack .

In this work, an efficient model for simulating bubble dispersion and coalescence due to turbulence is developed in the Euler-Lagrange framework. The primary liquid phase is solved on the Euler grid with the RANS turbulence model. Bubble motion is computed with the force balance equations. One-way coupling between two phases is assumed and the framework is designed for the computation of disperse bubbly flows at low Eötvös number. The turbulent dispersion of the dispersed phase is reconstructed with the continuous random walk (CRW) model. Bubble-bubble collisions and coalescence are accounted for deterministically. To accelerate the time-consuming search for potential collision partners in dense bubbly flows, the sweep and prune algorithm is employed, which can be utilized in arbitrary mesh types and sizes. Validation against experiments of turbulent pipe flows demonstrates that the one-way coupled EL-CRW dispersion model can well reproduce the bubble distribution in a typical dense bubbly pipe flow. Good agreement of the bubble size distribution at the pipe outlet between the simulation and the experiment is obtained.

Convolutional neural network (CNN) has been proven an effective way to extract high-level features from remote sensing (RS) images automatically. Many variants of the CNN model have been proposed, including principal component analysis network (PCANet), canonical correlation analysis network (CCANet), multiple-scale CCANet (MS-CCANet) and multiview CCANet (MCCANet). The PCANet is specialized for single view feature abstraction, while in many real-world practices, the RS data are frequently observed from many more views. Although CCANet, MS-CCANet and MCCANet can be applied to two or more view data, they consider only the pair-wise correlation by calculating a series of two-order covariance matrices. Moreover, it remains undiscovered how to handle high-order consistencies which can only be explored by simultaneously examining all views. In this paper, we propose the tensor canonical correlation analysis network (TCCANet) to tackle this problem. Particularly, TCCANet learns filter banks by simultaneously maximizing arbitrary number of views with high-order-correlation and solves the optimization problem by decomposing a covariance tensor. After the convolutional stage, we utilize binarization and block-wise histogram strategies to generate the final feature. Numerical experiment results on RSSCN7 and SAT-6 datasets demonstrate the advantages of TCCANet and MS-TCCANet for RS scene recognition.

Canonical correlation analysis (CCA) is a main technique of linear subspace approach for two-view dimension reduction by finding basis vectors with maximum correlation between the pair of variables. The shortcoming of the traditional CCA lies that it only handles data represented by two-view features and cannot reveal the nonlinear correlation relationship. In recent years, many variant algorithms have been developed to extend the capability of CCA such as discriminative CCA, sparse CCA, kernel CCA, locality preserving CCA and multiset canonical correlation analysis (MCCA). One representative work is Laplacian multiset canonical correlations (LapMCC) that employs graph Laplacian to exploit the nonlinear correlation information for multiview high-dimensional data. However, it possibly leads to poor extrapolating power because Laplacian regularization biases the solution towards a constant function. In this paper, we present Hessian multiset canonical correlations (HesMCC) for multiview dimension reduction. Hessian can properly exploit the intrinsic local geometry of the data manifold in contrast to Laplacian. HesMCC takes the advantage of Hessian and provides superior extrapolating capability and finally leverage the performance. Extensive experiments on several popular datasets for handwritten digits classification, face classification and object classification validate the effectiveness of the proposed HesMCC algorithm by comparing it with baseline algorithms including TCCA, KMUDA, MCCA and LapMCC.

Dispersed multiphase flow can be simulated by the Euler-Lagrange approach where the continuous phase is described on the Euler grid and the dispersed phase is treated as Lagrange points. The point-mass Euler-Lagrange method requires the bubble or particle size to be much smaller than the cell size when the interphase coupling is considered. Otherwise, numerical instabilities may arise. However, the point-mass prerequisite is hard to achieve, for example, if very fine grids are used to resolve turbulence in LES. Therefore, interpolation methods between the Euler and the Lagrange phase for finite-size bubbles which are bigger than or of the same size as numerical cells are discussed in this work. The spatially distributed coupling method is used for the Lagrange-to-Euler coupling. For the Euler-to-Lagrange coupling, a new approach is proposed and validated in different cases. Finally, a combination of coupling approaches is employed in a square bubble column reactor.

In the past decade, deep learning (DL) algorithms have been widely used for remote sensing (RS) image recognition tasks. As the most typical DL model, convolutional neural networks (CNNs) achieves outstand performance for big RS data classification. Recently, a variant of CNN, dubbed canonical correlation analysis network (CCANet), was proposed to abstract the two-view image features. Extensive experiments conducted on several benchmark databases validate the effectiveness of CCANet. However, the CCANet structure is powerless when the observations arrive from more than two sources. To serve the multiview purpose, in this letter, we propose multiview CCANets (MCCANets). Particularly, the MCCANet model learns the stacked multiperspective filter banks by the MCCA method and builds a deep convolutional structure. In the output stage, the binarization and the blockwise histogram are employed as nonlinear processing and feature pooling, respectively. To access the effectiveness of the MCCANet, we conduct a host of experiments on the RSSCN7 RS database. Extensive experimental results demonstrate that the MCCANet outperforms the two-view CCANet.

In recent years, deep learning has attracted an increasing amount of attention in machine learning and artificial intelligence areas. Currently, many deep learning network-related architectures such as deep neural networks (DNNs), convolutional neural network (CNN), wavelet scattering network (ScatNet) and principal component analysis network (PCANet) have been proposed. The most effective network is PCANet, which has achieved promising performance in image classification, such as for face, object and handwritten digit recognition. PCANet can only handle data that are represented by single-view features. In this paper, we present a canonical correlation analysis network (CCANet) to address image classification, in which images are represented by two-view features. The CCANet learns two-view multistage filter banks by a canonical correlation analysis (CCA) method and constructs a cascaded convolutional deep network. Then, we incorporate filters with binaryzation and block-wise histogram processes to form the final depth structure. In addition, we introduce a variation of CCANetdubbed RandNet-2in which the filter banks are randomly generated. Extensive experiments are conducted using the ETH-80, Yale-B, and USPS databases for object classification, face classification and handwritten digits classification, respectively. The experimental results demonstrate that the CCANet algorithm is more effective than PCANet, RandNet-1 and RandNet-2.

Deep networks are vulnerable to adversarial attacks from malicious adversaries. Currently, many adversarial learning algorithms are designed to exploit such vulnerabilities in deep networks. These methods focus on attacking and retraining deep networks with adversarial examples to do either feature manipulation or label manipulation or both. In this paper, we propose a new adversarial learning algorithm for finding adversarial manipulations to deep networks. We formulate adversaries who optimize game-theoretic payoff functions on deep networks doing multi-label classifications. We model the interactions between a classifier and an adversary from a game-theoretic perspective and formulate their strategies into a Stackelberg game associated with a two-player problem. Then we design algorithms to solve for the Nash equilibrium, which is a pair of strategies from which there is no incentive for either the classifier or the adversary to deviate. In designing attack scenarios, the adversary’s objective is to deliberately make small changes to test data such that attacked samples are undetected. Our results illustrate that game-theoretic modelling is significantly effective in securing deep learning models against performance vulnerabilities attached by intelligent adversaries.

In recent years, the advances in data collection and statistical analysis promotes canonical correlation analysis (CCA) available for more advanced research. CCA is the main technique for two-set data dimensionality reduction such that the correlation between the pairwise variables in the common subspace is mutually maximized. Over 80-years of developments, a number of CCA models have been proposed according to different machine learning mechanisms. However, the field lacks an insightful review for the state-of-art developments. This survey targets to provide a well-organized overview for CCA and its extensions. Specifically, we first review the CCA theory from the perspective of both model formation and model optimization. The association between two popular solution methods, i.e., eigen value decomposition (EVD) and singular value decomposition (SVD), are discussed. Following that, we present a taxonomy of current progresses and classify them into seven groups: 1) multi-view CCA, 2) probabilistic CCA, 3) deep CCA, 4) kernel CCA, 5) discriminative CCA, 6) sparse CCA and 7) locality preserving CCA. For each group, we demonstrate two or three representative mathematical models, identifying their strengths and limitations. We summarize the representative applications and numerical results of these seven groups in real-world practices, collecting the data sets and open-sources for implementation. In the end, we provide several promising future research directions that can improve the current state of the art.

The objective of the present study was to investigate the migration characteristics and model of chloride ions in salt storage asphalt mixtures under various working condition through natural leaching test. Three doses of NaCl-based salt storage fillers, two leaching temperatures, and seven freely leaching times were designed to simulate different work environment in this study. Auto ions analyser was utilized to characterize residual chloride ions content in salt storage asphalt mixtures before and after water leaching process, which indirectly stated the movement behaviour of sodium chloride from sample into water. And Fick's second law was used to establish migration model of residual chloride ions content in salt storage asphalt mixtures. The results show that residual chloride ions content and attenuation rate of salt storage asphalt mixtures all decreased with increasing leaching time. Also, high leaching temperature and content of salt storage fillers resulted in accelerating the motion of chloride ions in early leaching stage. The deviation analysis indicated that the established migration model of chloride ions had good accuracy except low usage of salt storage fillers or short water leaching time. The migration model could predict the residual chloride ions content in the salt storage asphalt mixtures under certain work condition.

This study investigates the potential drug-excipient interactions of polystyrene sulfonic acid (PSSA) and two weakly basic anticancer drugs, lapatinib (LB) and gefitinib (GB), in amorphous solid dispersions. Based on the strong acidity of the sulfonic acid functional group, PSSA was hypothesized to exhibit specific intermolecular acid-base interactions with both model basic drugs. Ultraviolet (UV) spectroscopy identified red shifts, which correlated well with the color change observed in lapatinib-PSSA solutions. Fourier transform infrared (FTIR) spectra suggest the protonation of the quinazoline nitrogen atom in both model compounds, which agrees well with data from the crystalline ditosylate salt of lapatinib. X-ray photoelectron spectroscopy (XPS) detected increases in binding energy of the basic nitrogen atoms in both lapatinib and gefitinib, strongly indicating protonation of these nitrogen atoms. (15)N solid-state NMR spectroscopy provided direct spectroscopic evidence for protonation of the quinazoline nitrogen atoms in both LB and GB, as well as the secondary amine nitrogen atom in LB and the tertiary amine nitrogen atom in GB. The observed chemical shifts in the LB-PSSA (15)N spectrum also agree very well with the lapatinib ditosylate salt where proton transfer is known. Additionally, the dissolution and physical stability behaviors of both amorphous solid dispersions were examined. PSSA was found to significantly improve the dissolution of LB and GB and effectively inhibit the crystallization of LB and GB under accelerated storage conditions due to the beneficial strong intermolecular acid-base interaction between the sulfonic acid groups and basic nitrogen centers.

During the past decade, deep learning (DL) has been proven an effective way for image recognition. Various kinds of DL models such as, graph neural network (GNN), gated graph neural networks (GG-NNs) and graph convolutional neural networks (GCN) have been proposed. The latest GCN algorithm has achieved prominent performance for knowledge graph classification and molecular recognition. However, it only calculates the simple zero-one adjacency matrix, which is powerless to distinguish the self-node and other nodes. In this paper, we propose a novel negative one-zero-one graph to tackle this problem. Particularly, the negative one-zero-one graph employs zero to identify the node itself. To further explore the distance structure of inner-class samples, we develop the negative one-zero-one graph to a Euclidean distance version. Based on the ordinary graph, we also propose three kinds of hypergraph construction methods, such as hypergraph label method, hypergraph Euclidean distance method and the improvement method of hypergraph Euclidean distance. Additionally, we extend GCN to image recognition tasks in this paper. Extensive experiments are conducted on PASCAL VOC 2007 dataset to validate the proposed algorithms. The experiments results prove that the proposed methods is superior to the some methods for image recognition.

With the rapid development of representation learning, deep learning has been proved to be an effective technique to extract high level features. Many variants have been reported including convolutional neural network (CNN), principle component analysis networks (PCANet) and canonical correlation analysis networks (CCANet). The representative CCANet utilizes CCA to learn two-view multi-stage filter banks and achieves significant superiority to PCANet for object recognition. However, CCANet tends to only use the output feature of the last convolutional stage, which ignores the previous different scale features. To surmount this problem, in this paper, we present a novel method dubbed multiple scale canonical correlation analysis networks (MS-CCANet). Specifically, the MS-CCANet learns more discriminative information by stacking multi-scale features of all the convolutional stages together. Extensive experiments are conducted on ETH-80 dataset to verify the effectiveness of MS-CCANet. The results demonstrate that the proposed MS-CCANet outperforms the state-of-art methods including PCANet and CCANet.