Abstract: With the emergence of new applications and the increasing cost of new semiconductor manufacturing technology, high energy-efficiency and flexibility are both critical for processors. Dynamically reconfigurable computing architecture, which has the both characteristics, is one of the most promising architectures for future processors. It has shown notable advantages in many important application fields, compared to conventional architectures. However, the fundamental reasons for the characteristics have never been deeply discussed in previous papers. This paper analyzes the reasons from the perspective of design methodology. The technologies of dynamically reconfigurable computing are continually evolving, and some cheering results of application have been achieved. This paper summarizes the latest progress in key technologies and provides an introduction of the application achievements.
Abstract: Aiming at the difficulty of obtaining sufficient labeled Hyperspectral image (HSI) data and the inconsistent feature distribution of different HSIs, a novel Unsupervised heterogeneous domain adaptation CycleGan (UHDAC) is proposed by using CycleGan to capture the transferable features in the absence of similar data. On the one hand, the two-way mapping is used to find the internal relationship between the source and target domain data, while the two-way adversary is used to constrain the source and target domain features, realizing the alignment of feature distributions. On the other hand, the CORAL loss function is introduced to minimize the distance between the second-order statistical difference between the source and target domain features, so as to solve the insufficient constraint of mapping relationship caused by the low consistency of HSI data structure in different domains. Experiments on three real HSI datasets show that UHDAC can effectively realize the unsupervised classification of target domain HSI with high classification accuracy by using the labeled HSI data in the source domain.
Abstract: Personalized recommendation systems predict potential demand by analyzing user preferences. Generally, user feedback information is inferred from implicit feedback or explicit feedback. Nevertheless, feedback can be contaminated by user's mis-operations or malicious operations, and may thus lead to incorrect results. We propose a novel Multi-feedback pairwise ranking method via Adversarial training (AT-MPR) for recommender to enhance the robustness and overall performance in the event of rating pollution. The MPR method extends Bayesian personalized ranking (BPR) to cover three types of feedback: positive, negative, and unobserved. It obtains user preferences in a probabilistic way through multiple feedbacks at different levels. To reduce the impact of feedback noise, we train an MPR objective function using minimax adversarial training. Experiments on two datasets show that the AT-MPR model achieves satisfactory performance and outperforms the state-of-the-art implicit feedback collaborative ranking models in two evaluation metrics.
Abstract: We designed a spatiotemporal generative adversarial network which given some initial data and random noise, generates a consecutive sequence of spatiotemporal samples that have a logical relationship. We build spatial discriminators and temporal discriminators to distinguish whether the samples generated by the generator meet the requirements for time and space coherence. The model is trained on the skeletal dataset and the Caltrans Performance Measurement System District 7 dataset. In contrast to traditional Generative adversarial networks (GANs), the proposed spatiotemporal GAN can generate logically coherent samples with the corresponding spatial and temporal features while avoiding mode collapse. In addition, we show that our model can generate different styles of spatiotemporal samples given different random noise inputs. This model will extend the potential range of applications of GANs to areas such as traffic information simulations and multiagent adversarial simulations.
Abstract: In the application of deep learning, the depth and width of the neural network structure have a great influence on the learning performance of the neural network. This paper focuses on structural optimization of depth and width, leveraging the information entropy model and decision tree strategy as feature selection and structural adjustment to optimize neural network candidates. Therefore, a decision tree-based heuristic optimization algorithm for neural network structural adjustment is proposed. Furthermore, the proposed approach is applied to fully-connected neural networks trained on the Iris dataset, and the proposed approach is verified effective via experimental simulation.
Abstract: By advances in cloud computing, users are allowed to remotely store their data in the cloud, manage the stored data without limitation of time and place, and give rights to visitors that want to access to their data. As may no longer possess data physically, the data owner has to ensure the integrity of the data with the public key given by Public key infrastructure (PKI). However, there are many security risks of the traditional PKI and the certificate management is complex. We utilize elliptic curve group to propose a certificateless signature to solve the above problem. To check the integrity of files stored in the cloud, we design a certificateless public verification mechanism based on the signature and further extend it to support batch auditing tasks. Meanwhile, an efficient key updating method is proposed to provide visitors a friendly data success environment. The security analysis proves that the proposed scheme is secure under the discrete logarithm assumption. Extensive theoretical analyses and experimental results show the effectiveness of the proposed scheme.
Abstract: Along with the extension of optical lithography to 28nm node and beyond, several effects that can be neglected in previous technology nodes become more and more prominent. One of the most striking effects is the 3D mask effect where the mask transmittance and phase are influenced by the mask topography. This study started with Kirchhoff mask model based experiment of 28nm node SRAM cell features based on Attenuated Phase-shift mask (Att. PSM), which were calibrated into Optical proximity correction (OPC) process models. A 28nm test model was constructed with Kirchhoff mask modeling approach. Once the standard Kirchhoff effects were working on the test wafer, the 3D mask effects were included for the same data. Then the interactive relations among these mask data were investigated and found to generate as much as 15nm edge shift differences. The mask model was then refit including the 3D mask effect and compared with the Kirchhoff mask model in the simulator to better understand the impact of 3D mask effect on patterns.
Abstract: Message authentication is a process that allows sender Alice to transmit a source state to receiver Bob such that the latter is assured of the authenticity. We study this problem in the physical-layer where the channel is degraded by a Rayleigh fading and a Gaussian noise. We propose an efficient construction for our problem. Our idea is to amplify a basic codeword of the source state with a small but dynamic scaling factor. This scaling factor is asymptotically 1. It thus essentially incurs no power expansion. Our authentication is achieved through a statistical test. Our scheme is provably secure and performs better than the related existing schemes. Further, our error analysis allows it to be realized with very practical parameters.
Abstract: Differential-linear cryptanalysis has attracted much attention since proposed to attack DES in 1994, and then some generalized theories are developed to complement and unify the method. However, the links between differential-linear cryptanalysis and other important cryptanalysis methods have been still missing. The motivation is to fix the gap. By establishing some boolean equations, we propose the mathematical links among differential, linear and differential-linear attacks. We then generalise the definition of capacity and present some properties of the capacity of differential function. The links and properties are employed to explore the relationships between multidimensional differential-linear hulls and integral distinguishers. We show that a multidimensional differential-linear hull of certain correlation always implies the existence of an integral distinguisher and a zero-correlation linear hull, while a special integral distinguisher indicates the existence of a multidimensional differential-linear hull.
Abstract: Three of the most essential criteria for cryptographically strong Boolean functions are resiliency, high nonlinearity and high algebraic degree. We give a technique for constructing 1-resilient Boolean functions with high nonlinearity via modifying PS- class bent functions. The main technique is to extend the support of bent functions in PS- class by additionally defining two different plateaued functions on two suitably chosen subspaces. A large class of highly nonlinear 1-resilient functions which were not known earlier are obtained.
Abstract: Cyclic codes are an important class of linear codes. Cyclic codes are applied in data storage, communication systems and consumer electronics due to their efficient encoding and decoding algorithms. We utilize the cyclotomy of order 2 to study the cyclic codes with length n=2pe and dimension k=n/2=pe. The cyclic codes from our construction are either optimal or almost optimal among all cyclic codes with the same length and dimension. We get the enumeration of these cyclic codes and study the hull of cyclic codes of length n over Fq. We obtain the range of ℓ=dim(Hull(C)). Finally, we construct and enumerate cyclic codes of length n having hull of given dimension.
Abstract: Compressed sensing (CS) exploits the sparsity of images or image patches in a transform domain or synthesis dictionary to reconstruct images from undegraded images. Because the synthesis dictionary learning methods involves NP-hard sparse coding and expensive learning steps, sparsifying transform based blind compressed sending (BCS) has been shown to be effective and efficient in applications, while also enjoying good convergence guarantees. By minimizing the rank of an overlapped patch group matrix to efficiently exploit the nonlocal self-similarity features of the image, while the sparsifying transform model imposes the local features of the image. We propose a combined low-rank and adaptive sparsifying transform (LRAST) BCS method to better represent natural images. We utilized the patch coordinate (PCD) descent algorithm to optimize the method, and this enforced the intrinsic local sparsity and nonlocal self-similarity of the images simultaneously in a unified framework. The experimental results indicated a promising performance, even in comparison to state-of-theart methods.
Abstract: We studied a novel image thresholding method using symmetric co-occurrence matrix probability information, in which, the relative homogeneity characteristics of the segmented object and background is explored, square distance co-occurrence matrix thresholding criterion is reformed, and the relative criterion of square distance symmetric co-occurrence matrix is proposed. The new thresholding method applies the spatial information of image and takes the relative information between classes into account. The experimental results show that, comparing with the existing correlation methods, the proposed method can extract the object more integrity, and reserve the edge more clearly.
Abstract: Early and accurately detecting faults is crucial for the modern manufacturing system. We proposed a novel Deep fault diagnosis (DFD) method for rotating machinery with scarce labeled samples. A spectrogram of the raw vibration signal is calculated by applying a Short-time Fourier transform (STFT). Several candidate Support vector machine (SVM) models are trained with different combinations of features in the feature pool with scarce labeled samples. By evaluating the pretrained SVM models on the validation set, the most discriminative features and best-performed SVM models can be selected, which are used to make predictions on the unlabeled samples. The predicted labels reserve the expert knowledge originally carried by the SVM model. They are combined together with the scarce fine labeled samples to form an Augmented training set (ATS). Finally, a novel 2D deep Convolutional neural network (CNN) model is trained on the ATS to learn more discriminative features and a better classifier. Experimental results on two fault diagnosis datasets demonstrate the effectiveness of the proposed DFD.
Abstract: Ultrasound computed tomography (UCT) is considered to have great potential for the early diagnosis of knee joint injuries. Combining prior knowledge and sound speeds of tissues, UCT can diagnose a variety of symptoms and the extent of tissue lesion. However, no existing studies can reconstruct the sound speed distributions of knee joints in UCT. Therefore, in this study, sound speed distributions of knee joints are reconstructed to provide an imaging basis for the formulation of a treatment plan. In addition, Full waveform inversion (FWI) methods are utilized to estimate highresolution images. However, for media with large variations in sound speeds, traditional FWI methods may lead to a cycle-skipping phenomenon and incorrect convergence direction. Hence, this study proposes a multi-centerfrequency source based FWI method to overcome the above limitations. A penalized least-squares optimization problem is constructed to obtain a numerical solution of the sound speed distribution. Computer simulations are conducted to prove the effectiveness and robustness of the proposed method. Furthermore, the reasons for selecting the center frequencies of sources are analyzed. Two numerical knee joint phantoms are used to evaluate the performance. The results suggest that the biases of the reconstructed images are less than 1% under a 5% noise condition.
Abstract: In this paper, we investigate a class of p-ary cyclic codes whose duals have two zeros for some special cases and calculate their weight distributions explicitly. The results show that the codes have at most five nonzero weights. Moreover, they contain some optimal threeweight codes meeting the Griesmer bound.
Abstract: We present a robust frequency synchronization algorithm for the Coherent optical Offset quadrature-amplitude modulation based Filter-bank multicarrier (CO-FBMC-OQAM) systems. It is a two-stage method operating in the frequency domain, which expands the estimated range of the normalized fractional frequency offset to (-1, +1). We designed and introduced a cascading cross-correlation function of real-valued pilots to roughly estimate the frequency offset in the first stage. The residual frequency offset was accurately tracked by a non-cascading cross-correlation function afterwards. This approach enlarges the maximum estimated range, and also resists interference and noise. A CO-FBMC-OQAM system with the sampling rate of 50 Gsample/s was numerically investigated to evaluate the frequency-offset estimation and compensation capability. All obtained results prove that the proposed frequency synchronization method can correct the laser-frequency offset and improve the transmission performance significantly.
Abstract: An application layer privacy data protection scheme combining dynamic and static analysis is proposed. Android component life cycle and system calls are first studied, and the taint propagation path under the cross-component scenario in static analysis is optimized. Based on the static analysis, a privacy preserving container is designed and implemented on both the Framework layer and the Native layer of Android. The scheme generates a privacy protection policy file by constructing leakage paths for privacy data propagation in Android applications, and monitors privacy leakage in the running environment of the target application according to the policy file. Experiments show that the proposed scheme can effectively protect user privacy while running third-party applications.
Abstract: Cyberspace mimic domain name system (CMDNS) adopts dynamic heterogeneous redundant architecture with strategic decision mechanism to control the effectiveness of uncertain disturbance. There is lack of methods to evaluate the availability and awareness security of CMDNS. To further describe and analyze the characteristics of CMDNS accurately, the Generalized stochastic Petri net (GSPN) is used to model the attack disturbance and defense of mimic Domain name system (DNS), and the availability and awareness security of Dissimilar redundancy system (DRS) and mimic DNS under different disturbance intensities are compared. We compared the different effect of local service query and real network query on the average response delay. The results show that the introduction of mimic architecture will inevitably pay the corresponding delay cost which increase 9.3% compared with traditional local DNS, but it has little effect on the service which only increases by 1.9% compared with the DNS transmission delay at the network communication level.
Abstract: This paper studies surrounding dynamically multi-targets with second-order integrator/nonlinear systems and continuous time-varying topology. The continuous time-varying topology refers to the continuous change of the topology rather than switching in several fixed topologies. Surrounding multi-targets is an extension of containment control. The distributed control algorithm is designed based on the containment control theory. By transforming the system of containment control into error stability, the condition that the error between multi-agent systems and multi-targets to be constant can be obtained by using the algebraic graph theory, matrix theory and Lyapunov function stability analysis. The multiagent systems can dynamically surround multi-targets by designing the deviation vector. Finally, simulation experiments verify the effectiveness of the proposed algorithm.
Abstract: A fast method of transmit beamforming for large planar Multiple-input multiple-output (MIMO) radar is proposed. For MIMO radar with large elements, it is difficult to design a large number of orthogonal waveforms. We design a weight matrix to connect orthogonal waveforms and array elements. We optimize the weight matrix such that the transmit beampattern approximates the desired transmit beampattern. By introducing an auxiliary variable, a cyclic optimization algorithm based on Fast fourier transform (FFT) is proposed to obtain the optimal weight matrix, which is very efficient and can satisfy the real-time requirements. Numerical examples are provided to verify the effectiveness of the proposed algorithm.
Abstract: Due to unique features, Storage class memory (SCM) technologies such as Phase change memory (PCM) open up new opportunities for architect engineers. In such a scenario, we present a true PCM storage system with an FPGA storage controller to explore ISP benefits. Our contributions are summarized as follows. 1) Propose a heterogeneous ISP architecture which uses FPGA working as storage controller and accelerator; 2) Offer a new research direction for designing storage devices which inherently support ISP; 3) Present a novel storage system which eliminate data transfer; 4) The first evaluation of ISP on a real SCM device; 5) Demonstrate significant performance gains by using efficient data flow and consuming extremely small amounts of host resources. Besides, the proposed system can be extended to handle other kinds of data applications by adding corresponding accelerator and data conversion module in FPGA. Multinode system can be realized for more aggressive results.
Abstract: A 2998MHz, 60MW klystron has been developed for future linear accelerators of synchrotron radiation sources. In order to save development and construction time, the 2998MHz, 60MW klystron is a modified version of an existing 2856MHz klystron used in linear accelerator of Beijing Electron-Positron Collider Ⅱ. Based on the simulations, it was shown that the scaled version klystron was able to produce more than 60MW at a 350kV beam voltage. A prototype was built and tested. It successfully produced 50MW output power with an RF pulse width of 4.0μs, which exceeded the operating requirement of China's High Energy Photon Source. It also produced 62MW with an RF pulse width of 1.6μs, which proved the validity of new RF circuit design. This paper describes the details of the progresses of this klystron and advantages of this tube development.
Abstract: Guided transport vehicle (GTV) uses satellite positioning technology, lidar, and highperformance video sensors to achieve real-time positioning and tracing, these devices cost a lot. In addition, GTV shifts left and right during operation due to there is no fixed orbits constraint. It is very important to detect the lateral offset in real time to ensure the accurate tracing of the vehicle along the dotted line. Aiming at the above issues, this paper proposes a low-cost high precision tracing method for GTV based on integrated positioning, high resolution estimation of the vehicles positioning information is implemented by Cubature Kalman filters (CKF) with a loosely coupled mode based on GPS/SINS in this method; Considering that GTV has a fixed driving route, a two-stage map matching algorithm based on Hidden Markov model (HMM) is proposed, which further improves the accuracy; The lateral offset distance is detected based on the positioning information, which provides theoretical support for the subsequent control of the precise operation of the vehicle. The algorithm's feasibility has been verified by real vehicle experiments, the results show that the proposed algorithm achieves a centimeter-level positioning accuracy, and when the lateral offset distance is not less than 65cm, the detection accuracy of the lateral offset distance is more than 92%.
Abstract: Railway point machines (RPMS) are one of the key equipments in the railway system to switch different routes for the trains. Condition monitoring for RPMs is a vital measure to keep train operation safe and reliable. Taking convenience and low cost into consideration, a novel intelligent condition monitoring method for RPMs based on sound analysis is proposed. Time-domain and frequency-domain features are obtained, and normalized using z-score standardization method to eliminate the influences of different dimensions. Binary particle swarm optimization (BPSO) is utilized to select the most significant discrimination feature subset. The effects of the selected optimal features are verified using Support vector machine (SVM), 1-Nearest neighbor (1NN), Random forest (RF), and Naive Bayes (NB). Experiment results indicate SVM performs best on identification accuracy and computing cost compared with the other three classifiers. The identification accuracies on normal switching and reverse switching processes reach 100% and 99.67%, respectively, indicating the feasibility of the proposed method.