Search

Oil-paper insulation is the key insulation structure of the transformers, whose aging condition is closely related to the operations of the equipment. The degree of polymerization (DP) is the direct parameter characterizing the aging condition of oil-paper insulation. Recently, the near infrared spectroscopy (NIRS) measurement powered by quantitative analysis is used to evaluate DP of the insulating papers mainly. While the applications of NIRS for DP evaluation is constrained because the present spectral quantitative analysis method is not accurate and stable enough, especially for onsite tests. In this paper, we propose a Gaussian process regression (GPR) method to predict DP of the oil- paper insulation in laboratory as well as in field. Firstly, the basic principles of GPR algorithm are illustrated. A GPR model for DP prediction is established based on the spectra of differently aged insulating paper samples which are prepared in laboratory. The GPR model show a high prediction accuracy both for training set and testing set. The established GPR model is finally applied on the DP prediction of insulating papers originating from a de-tanked transformer. The accurate predicted DP and the reliable aging assessment results indicate that the established model can be implemented on site.

The present work presents a hybrid identification of recursive least squares with a metaheuristic called radial movement optimization (RMO) applied to the joints of a cylindrical robotic manipulator. The main contribution of the research is to consider the covariance matrix with RMO. At the end a comparative analysis will be made with some classic methods. The results of the identifications were tested with the RMSE. A covariance matrix has also been generated from manipulator identifications.

International audience; As part of the Historic Photographic Paper Classification Challenge, a multitude of approaches to quantifying paper texture similarity have been developed. These approaches have yielded encouraging results when applied to very controlled datasets containing photomicrographs of familiar specimens. In this paper, we report on the k-nearest neighbors classification performance of two multiscale analysis-based texture similarity approaches when applied to a much larger reference collection of silver gelatin photographic papers. The clusters for this data set were derived from a visual sorting experiment conducted by art conservators and paper experts later extended through crowd-sourcing. The results show that these texture similarity approaches, when combined with a simple k-nearest neighbors classification algorithm, yield workable performances with accuracy of up to 69%. We discuss this outcome in the context of available data and the cross-validation procedure used, then provide suggestions for improvement.

This paper investigates the reconfigurable intelligent surface (RlS)-aided multiple-input-single-output (MISO) systems with imperfect channel state information (CSI), where RIS-related channels are modeled by Rician fading. Considering the overhead and complexity in practical systems, we employ the low-complexity maximum ratio combining (MRC) at the base station (BS), and configure the phase shifts of the RIS based on long-term statistical CSI. Specifically, we first estimate the overall channel matrix based on the linear minimum mean square error (LMMSE) estimator, and evaluate the performance of mean square error (MSE) and normalized MSE (NMSE). Then, with the estimated channel, we derive the closed-form expressions of the ergodic rate. The derived expressions show that with Rician RIS-related channels, the rate can maintain at a non-zero value when the transmit power is scaled down proportionally to 1 /M or 1 /N2, where M and N are the number of antennas and reflecting elements, respectively. However, if all the RIS-related channels are purely Rayleigh, the transmit power of each user can only be scaled down proportionally to $1/\sqrt M $ or 1/N. Finally, numerical results verify the promising benefits from the RIS to traditional MISO systems.

Fault Tree Analysis is widely used in the reliability evaluation of critical systems, such as railway and automotive systems, power grids, and nuclear power plants. While there are efficient algorithms for calculating the probability of failure in static fault trees, Mean Time to First Failure (MTFF) evaluation remains challenging due to state space explosion. Recently, structural and symmetry reduction methods were proposed to counteract this phenomenon. However, systems with a large number of different and highly interconnected components preclude the use of reduction techniques. Their MTFF analysis requires the solution of a system of linear equations whose size is exponential in the number of components in the system. In this paper, we propose a solution leveraging Tensor Trains as a compressed vector and matrix representation. We build upon Binary Decision Diagram-based techniques to avoid explicit state space enumeration and use linear equation solvers developed specifically for Tensor Trains to efficiently solve the arising linear systems. As a result, our novel approach complements the existing reduction-based techniques and makes some previously intractable models possible to analyze. We evaluate our approach on an industrial case study adapted from a railway system, and other openly available benchmark models.

Optimal linear transmitter beamformers in multi-antenna multi-user systems are of the Minimum Mean Squared Error (MMSE) type (dual uplink MMSE receivers). MMSE designs make an optimal compromise between noise enhancement and interference suppression and reduce to matched filters at low SNR and zero-forcing at high SNR. We consider a realistic scenario of user channels of varying attenuation and constrain the beamformers to either zero-force or ignore each interference term. This leads to a reduced-order zero-forcing (RO-ZF) design in which the number of interference sources being zero-forced increases with SNR. We apply a simple large systems analysis (applicable to Massive MIMO) to determine the asymptotic performance of RO-ZF designs, determine the optimal ZF orders, and compare to optimal and ZF linear and Dirty Paper Coding (DPC) designs. RO-ZF designs lead to variable reductions of computational complexity and channel state information (CSI) requirements (esp. in future multi-cell extensions), both important considerations in Massive MIMO systems.

This paper proposes a novel three-dimensional (3D) cluster-based sixth-generation (6G) model for millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) channels. In the proposed model, a novel method based on visibility region (VR) is developed, where the smooth and consistent birth and death of clusters in both the time domain and space domain are successfully captured. By using the developed method, the space-time non-stationarity and time-space consistency of channels can be modeled for the first time. Based on the proposed model, important channel statistical properties are derived and thoroughly investigated. Simulation results demonstrate that the smooth birth and death of clusters, time-space consistency, and space-time non-stationarity are modeled sufficiently.

Brain-inspired hyperdimensional computing (HDC) is an emerging computational paradigm that has achieved success in various domains. HDC mimics brain cognition and lever-ages hyperdimensional vectors with fully distributed holographic representation and (pseudo)randomness. Compared to the traditional machine learning methods, HDC offers several critical advantages, including smaller model size, less computation cost, and one-shot learning capability, making it a promising candidate in low-power platforms. Despite the growing popularity of HDC, the robustness of HDC models has not been systematically explored. This paper presents a study on the robustness of HDC to errors in associative memory—the key component storing the class representations in HDC. We perform extensive error injection experiments to the associative memory in a number of HDC models (and datasets), sweeping the error rates and varying HDC configurations (i.e., dimension and data width). Empirically, we observe that HDC is considerably robust to errors in the associative memory, opening up opportunities for further optimizations. Further, results show that HDC robustness varies significantly with different HDC configurations such as data width. Moreover, we explore a low-cost error masking mechanism in the associative memory to enhance its robustness.

International audience; Reconfigurable intelligent surfaces (RISs) represent a new technology that can shape the radio wave propagation and thus offers a great variety of possible performance and implementation gains. Motivated by this, we investigate the achievable sum-rate optimization in a broadcast channel (BC) that is equipped with an RIS. We exploit the well-known duality between the Gaussian multiple-input multiple-output (MIMO) BC and multiple-access channel (MAC) to derive an alternating optimization (AO) algorithm which jointly optimizes the users’ covariance matrices and the RIS phase shifts in the dual MAC. The optimal users’ covariance matrices are obtained by a dual decomposition method in which each iteration is solved in closed-form. The optimal RIS phase shifts are also computed using a derived closed-form expression. Furthermore, we present a computational complexity analysis for the proposed AO algorithm. Simulation results show that the proposed AO algorithm can provide significant achievable sum-rate gains in a BC.

The popular technique of secure signaling over Gaussian MIMO wiretap channels, which makes use of artificial noise (AN) to increase secrecy rates, is considered. First, we briefly review the current state of affairs in this area and then provide new analytical results and insights on the usefulness of AN (jamming) to boost secrecy rates. The settings considered here go beyond the total transmit power constraint and include a number of additional constraints, such as interference (cognitive radio) and energy-harvesting constraints, for which the feasible set is not isotropic anymore so that the standard tools of the analysis cannot be used. By closely examining optimal precoding for the information-bearing and AN signals, we identify a number of cases where it is optimal to transmit no artificial noise at all (so that all the transmit power goes to the information-bearing signal). These cases include a fixed (no fading) MIMO WTC with single eavesdropper (for which we give a direct matrix-theoretic proof using novel matrix inequalities), multi-eavesdropper (com-pound) degraded and reversely-degraded channels, and multi-eavesdropper channels where there exists a dominant eavesdropper (for which we give a precise definition) or when the eavesdroppers collude. To improve secrecy rates by using AN, one has to look elsewhere.

Neural networks (NNs) are a driving force behind the ongoing information revolution, with a broad spectrum of applications affecting most aspects of science and technology. The interest in robust neural computation under adversarial noise has increased lately, due applications in sensitive tasks ranging from healthcare to finance and autonomous vehicles. This has ignited an influx of research on the topic, which for the most part focuses on obtaining robustness by altering the training process. In contrast, this paper surveys and develops a recently proposed novel approach to obtain robustness after training, by adding redundancy to the network and to the data in the form of error correcting codes.Since neural networks are essentially a concatenation of linear classifiers, we focus on obtaining robustness for a single linear classifier by coding the input and the classifier, and then apply the results on the network. We address two different types of adversaries, a worst-case one and an average-case one. For a worst-case adversary, that can choose the input to be attacked, we focus on binarized classifiers and show that the problem is related to construction of certain linear codes with restricted weight patterns. As a result, it is shown that the parity code can obtain robustness against any 1-erasure in any binarized NN, and no decoding is required. For an average-case adversary, that is given a uniformly random input to be attacked, it is shown that the optimal weights for any classifier and any code are given by the Fourier coefficients of that classifier. We demonstrate the latter experimentally, exposing improved accuracy-robustness tradeoff in neural classification of several popular datasets under state-of-the-art attacks.

The idea of this study is to validate a list of keywords derived from a scientific article by a domain expert from years of knowledge with prominent document similarity algorithms. For this study, a list of handcrafted keywords generated by Electric Double Layer Capacitor (EDLC) experts are chosen, and relevant documents to EDLC are considered for the comparison. Then, different similarity calculation algorithms were employed in different settings on the documents such as using the whole texts of the documents, selecting the positive sentences of the documents, and generating similarity score with automatically extracted keywords from the documents. The experiment’s outcome provides us with findings that the machine-generated keywords are mostly similar to the curated list by the domain experts. This study also suggests the preferable algorithms for similarity calculation and automated key-phrase extraction for the EDLC domain.

This paper considers an Adaptive Antenna Array (AAA), which weights are shared with those of a Feed-Forward (FF) part of the channel equalizer and the array output signal is combined with that of the Feed-Backward (FB) part of the equalizer. The array and this distributed equalizer operate as an indivisible adaptive filter, providing a good performance and an ability to receive a multipath data signal in the presence of external interference sources. The architecture of the array/equalizer and mathematical description of the Recursive Least Mean Squares (RLS), Least Mean Square (LMS) and Normalized LMS (NLMS) algorithms for weights calculation of the adaptive filter are presented. The simulation results demonstrate the efficiency of the proposed solution.

For the stacked paper placed parallel to each other, the characteristics of the side texture image are analyzed, and an algorithm for counting compensation combined with the overall information is proposed. Considering that the most algorithms only deal with one-dimensional information, it is easy to generate the blind area of counting, resulting in false detection and missing detection. On this basis, this paper counts the edges of the paper in combination with the overall image information. This method detects the binarized side image of the paper at the suspected position by using a rectangular template to eliminate the false detection position and recover the missed detection position. The experimental results show that the accuracy of the algorithm is significantly improved, and the counting error is controlled within the required range.

This brief presents a random number generator (RNG) based on irregular sampling of regular waveform method where the irregular signal is obtained by combining Fibonacci-Galois ring oscillators with an XOR gate. The RNG is implemented on a FPGA (field-programmable gate array). The regular waveform generated by the digital clock manager of the FPGA, is sampled at times corresponding to certain number of rising edges of the irregular signal, and the resulting bit stream is subjected to statistical tests of randomness. It is demonstrated that the resulting bit sequence from the proposed RNG satisfies NIST 800-22 test suit and Rabbit and SmallCrush batteries from TestU01 library without any need for post-processing such as Von Neumann or XOR. A comparison between the methods regular sampling of irregular waveform and irregular sampling of regular waveform is given in terms of robustness against external interference. The impact of selection of Fibonacci-Galois polynomials is discussed. Using digital design flow for TSMC 65nm process, an asic implementation of the proposed RNG is given having 1115 gates and 4.811 mW estimated power.

This paper studies the design of fabric drape prediction model based on BP neural network, and predicts the static and dynamic drape coefficients according to the number of warp and weft yarns, density of warp and weft yarns, weight per square meter and thickness of fabric. The effect of the number of intermediate layer neurons on the accuracy of neural network was studied. The influence of activation function on neural network prediction and initial value on local optimization. The simulation results show that the prediction results of fabric drape by BP neural network are good and meet the expected requirements.

The work is devoted to estimation of the amplitude frequency characteristics of the optical measuring system that implements method of phase triangulation. Method based on the analysis of images during spatial modulation of an optical source encoding a binary Gray code and on the spatial-frequency analysis of the image observed by the photo detector.

We deal with the waveform design problem for the radar-communications coexistence (RCC) that ensures low levels of waveform correlations and meanwhile implements efficient information embedding (IE) via the fast-time waveform transmission. To tackle the problem, we propose two unimodular waveform designs, in which the integrated sidelobe level (ISL) is evaluated to find the optimal waveform correlations for the radar transmission and the attributes of waveform phases with large degrees of freedom are exploited to embed communication symbols. The first design is an offline strategy simply using the phase initialization for waveform generation to embed communication symbols. It requires only a predefined database that maps the waveform initialization to the emitted version of it, which gives the lowest ISL of the waveform. The second design serves as an online strategy employing the phase difference between any two waveform elements for the implementation of IE, and its accuracy to approaching desirable phases is elaborately transformed into a quadratic form and jointly designed with the minimization of ISL. We formulate this design as a nonconvex optimization problem, whose compound objective involves the outer product of the waveform vector that is to be designed. Moreover, we apply the majorization-minimization technique to solve the formulated problem, based on which we also devise an efficient algorithm that iterates with a closed-form solution. Simulation results verify the effectiveness of the proposed designs for the RCC.

In this paper we propose a statistical approach for the optimization of the phased codes used in radar. By considering each code as a realization of a random variable we show that we can build good codes if the distribution of the random variable is suitably chosen. We then consider sequences generated by the skew tent map and we show that for some values of the bifurcation parameter the invariant probability density coincides with the desired one. To confirm this idea we compute the maximal peak of the ambiguity function outside zero relative to the bifurcation parameter of the skew tent map.

In this paper, we investigate the fast downlink beamforming for full-dimension multiple-input multiple-output systems under correlated Rician channels. Under the assumption that the base station (BS) has only statistical channel state information (CSI), we decouple each user’s beamforming vector and derive their optimal beamforming vector through the maximization of the average signal-to-leakage-plus-noise ratio (SLNR) lower bound. Then, to reduce the computation time, a model-driven deep learning (DL)-based beamforming algorithm is proposed, as well as a data-driven algoriothm for comparison. In the model-driven DL-based beamforming algorithm, the process of obtaining the beamforming vector is separated into two parallel neural networks which are constructed and trained independently. The proposed algorithms can achieve similar ergodic rate as the optimal beamforming algorithm with much less computation time, and the model-driven algorithm requires less computing resource than the data-driven algorithm.

We consider private information retrieval (PIR) for distributed storage systems with noncolluding nodes where data is stored using a non maximum distance separable (MDS) linear code. Recently, it was shown that when data is stored using certain non-MDS codes, the MDS-PIR capacity can be achieved, and is indeed the capacity of the system. In this paper, for storage codes not belonging to this class, we present a heuristic algorithm for their decomposition into punctured subcodes and a PIR protocol based on these punctured subcodes. The code decomposition is guided by the generalized Hamming weights of the storage code. We show that the proposed PIR protocol can achieve a larger PIR rate than that of all existing PIR protocols.

Quantum Approximate Optimization Algorithm (QAOA) provides approximate solution to combinatorial optimization problems. It encodes the cost function using a $p$ -level quantum circuit where each level consists a problem Hamiltonian followed by a mixing Hamiltonian. Despite the promises, few real-world applications (besides the pedagogical MaxCut problem) have exploited QAOA. The success of QAOA relies on the classical optimizer, variational parameter setting, and quantum circuit design and compilation. In this study, we implement QAOA and analyze its performance for a broader Quadratic Unconstrained Binary Optimization (QUBO) formulation to solve real-word applications such as, partially occluded object detection problem. Furthermore, we analyze the effects of above influential factors on QAOA performance. We propose a 3-level improvement of hybrid quantum-classical optimization for object detection. We achieve more than 13X execution speedup by choosing L-BFGS-B as classical optimizer at the first level and 5.50X additional speedup by exploiting parameter symmetry and more than 1.23X acceleration using parameter regression at the second level. We empirically show that the circuit will achieve better fidelity by optimally rescheduling gate operations (especially for deeper circuits) at the third level.

A Mean shift algorithm is a clustering & unsupervised algorithm. It is also called as “Mode seeking Algorithm”. A mean-shift algorithm used for image processing and also in computer vision. It doesn't require specified clusters like k-mean. Basically it has two types BMS & MS. By studying these two, it looks like BMS is slightly better than MS because it changes the dataset in each iteration with computed value. It's nonparametric algorithm and it can control no. of the cluster because the mean shift is much slower and much time complex and output depend on windows size. (Windows size is trivial). The BMS is an accelerated version which uses the original data only in the first step, then re-smoothes previous estimates. A BMS does not predefined shape on data cluster. To overcome such a situation, we can use the BMS and K-Medoid algorithm because k-medoid is partitioning technique of a cluster. It has better scalability for the larger datasets and much efficient so using these two we can build mean shift algorithm faster than before.

The conventional approach to pre-process data for compression is to apply transforms such as the Fourier, the Karhunen-Loeve, or wavelet transforms. One drawback from adopting such an approach is that it is independent of the use of the compressed data, which may induce significant optimality losses when measured in terms of final utility (instead of being measured in terms of distortion). We therefore revisit this paradigm by tayloring the data pre-processing operation to the utility function of the decision-making entity using the compressed (and therefore noisy) data. More specifically, the utility function consists of an Lp-norm, which is very relevant in the area of smart grids. Both a linear and a non-linear use-oriented transforms are designed and compared with conventional data pre-processing techniques, showing that the impact of compression noise can be significantlv reduced.

We propose a multistage differentiable method to select convolutional channels and construct light neural networks from a heavy network for inference on a subset of a big data set. The selection proceeds backward in layers and utilizes sparse penalty to diversify channel scores. The resulting light network gains sizable accuracy over the baseline heavy network.

Polar codes natively lack the flexibility that is desired for practical application. Namely, Arikan’s polar code definition can only achieve code lengths that are powers of two. Rate-matching techniques, known as puncturing and shortening, have been applied to polar codes to grant a flexible block length. By considering polarizing kernels of alternate dimensions, Multi-kernel polar codes improve natural block length flexibility. With the recent advent of the 3GPP 5th generation New Radio specification, there now exists an industry standard for length-flexible polar codes. This paper outlines various state-of-the-art flexible polar coding schemes, such as puncturing, shortening, and multi-kernel construction, and evaluates their efficacy with respect to the newly designed 3GPP standard. Simulations and an in-depth analysis are presented.

Cell tracking automatically in time-lapse image sequences is important for understanding the dynamic pattern of micro-cell. In this paper, we present a novel method for tracking cell with shape feature based on the delta generalized labeled multi-Bernoulli (delta-GLMB) filter which is of great research significance. The delta-GLMB filter with cell shape parameters can improve the tracking accuracy. This approach is evaluated and compared with raw detection using the generalized optimal sub-pattern assignment (GOSPA) metric on real N2DH-SIM cell sequences. Experiment results show that the delta-GLMB filter can provide the shape information as well as the better estimation than raw detection and KTH method.

Convolutional analysis operator learning (CAOL) methods train an autoencoding convolutional neural network (CNN) in an unsupervised learning manner, to more accurately solve inverse problems. Block Proximal Gradient method using a Majorizer (BPG-M) achieved fast and convergent CAOL, by using sharp majorizers and the momentum terms. This paper proposes a model-based image reconstruction (MBIR) method using autoencoding CNNs trained via CAOL, for sparse-view computational tomography (CT). We apply BPG-M to rapidly and stably solve the corresponding block multi-nonconvex optimization problem. Numerical experiments show that, for sparse-view CT, 1) the proposed MBIR method outperforms the standard MBIR method using edge-preserving regularization; 2) larger parameter dimensions of autoencoding CNNs improve reconstruction accuracy of the proposed MBIR method; and 3) when using BPG-M, sharper majorization is more critical for accelerating its convergence than giving more weights on extrapolation.

In this paper, we study a multiple access monostatic backscatter communication (BackCom) system, where one reader talks to multiple backscatter nodes (BNs). We propose to use nonorthogonal multiple access (NOMA) to improve the efficiency of BackCom system. Specifically, the reader uses the feedback signal during the training phase to separate the multiple BNs into two groups. The reader then randomly pairs the BNs from these two groups to implement NOMA. We propose to set the reflection coefficients for these two groups differently to further distinguish the reflected signal power for the paired BNs. To characterize the network performance, we derive an analytical expression for the success rate which shows the percentage of bits that can be successfully decoded in one time slot. We also present the design guideline of the reflection coefficients for the two groups to maximize the system performance. Our numerical results show that increasing the readers transmit power has little impact on the system performance when the channel condition on the transmission link is less severe. Instead, the proper choice of the reflection coefficients for the two groups of BNs can greatly enhance the BackCom system.

In recent years, ultra-broadband communication using densification of millimeter-wave (mmWave) smallcells has attracted attention owing to its ability to accommodate increasing mobile data traffic. To make full use of mmWave access, we need to deploy ultra-broadband backhauling lines such as 10Gbps Ethernet everywhere at the expense of very high installation costs. The mmWave Edge Cloud system proposed in this paper is a solution to this problem. Based on achievable user context information e.g. mobility and traffic demand, user data can be transmitted and stored in storage (prefetching) installed at the smallcell site in advance. By using this, mmWave access can bring out sufficient performance even with low-capacity backhaul. For the analysis, we have developed a new mobility model, traffic model, and prefetching algorithm. Numerical results showed that the proposed method with 1Gbps-backhaul and prefetching can achieve 95% of the ideal system rate while reduce the average access latency by a factor of 67% as compared to the case without prefetching.

In spatial modulation (SM) and quadrature SM (QSM), the maximum-likelihood (ML) decoder provides the optimum solution with high decoding complexity at the receiver side. This paper presents a novel low-complexity algorithm for decoding the SM and QSM symbols, referred to as the min-max algorithm. This is an intelligent searching algorithm, particularly designed for the tree-search of the SM and QSM decoders. The proposed algorithm expands the minimum Euclidean distance (ED) by adding a single node at each step, without considering the order of the branches. The expanding process stops if the minimum ED occurs at the end of a fully expanded branch. It is shown that the proposed algorithm achieves the optimum ML bit error rate performance with a significant reduction in the decoding complexity comparing with SM-ML and QSM-ML, as well as other existing sphere decoding algorithms. Simulations and mathematical analysis are provided to assess the decoding performance and complexity of the proposed algorithm.

In this paper, we present a practical Dirty Paper Coding (DPC) scheme using sum codes based on LDPC codes. While a typical sum-code-based DPC scheme uses a constrained decoder as part of the encoding operation, such an approach fails in the case of LDPC codes since the constrained decoder based on a standard LDPC decoder tends to get stuck at highly suboptimal points in typical scenarios. We get around this difficulty via a simple code construction method based on bit- mapping or permutation. Examples included in this paper show that compared to standard communication schemes the proposed DPC scheme yields substantial savings in terms of the

We present a CRC-aided LDPC coding scheme that can outperform the underlying LDPC code under ordered statistic decoding (OSD). In this scheme, the CRC is used jointly with the LDPC code to construct a candidate list, instead of conventionally being regarded as a detection code to prune the list generated by the LDPC code alone. As an example we consider a (128,64)5 G LDPC code with BP decoding, which we can outperform by 2 dB using a (128,72) LDPC code in combination with a 8-bit CRC under OSD of order t=3. The proposed decoding scheme for CRC-aided LDPC codes also achieves a better performance than the conventional scheme where the CRC is used to prune the list. A manageable complexity can be achieved with iterative reliability based OSD, which is demonstrated to perform well with a small OSD order.

We propose a binary message passing decoding algorithm for product codes based on generalized minimum distance decoding (GMDD) of the component codes, where the last stage of the GMDD makes a decision based on the Hamming distance metric. The proposed algorithm closes half of the gap between conventional iterative bounded distance decoding (iBDD) and turbo product decoding based on the Chase–Pyndiah algorithm, at the expense of some increase in complexity. Furthermore, the proposed algorithm entails only a limited increase in data flow compared to iBDD.

As the current soldier occupational skill appraisal platform is widely used in the troops, the paper quality problem needs to be solved as soon as possible. The article analyzed the paper difficulty distribution computing model, and then constructed models for the paper generating process, put forward an improved genetic algorithm. The new improved algorithm adopted sub-integer coding method and made some changes to the genetic operators. The result of simulation proved the validity of the algorithm, and the paper generated through the algorithm could satisfy the restrictions well.

This paper considers that write-once memory (WOM) codes can be seen as a type of dirty-paper code. The current state of the memory, which is known to the encoder, plays the role of the known interference of dirty-paper coding. Erez, Shamai and Zamir showed that lattice strategies can achieve the capacity of the known-interference channel. In this paper, lattices are used to design a WOM code. Encoding is performed modulo a shaping lattice with respect to a lattice fundamental region to obtain a codeword, to be added to the current state of the memory. The fundamental region is designed to accommodate the limitations of the flash memory system, particularly, that values can only increase. The criterion for evaluation is average number of writes. In order to improve the average number of writes, “coset select” bits are introduced, to maximize the average number of writes. For an eight-dimensional lattice, numerical results for practical parameter choices show a promising trend.

With today’s eye-tracking technologies, it is possible to analyze cognitive processes such as reading. There are many studies on that topic. We have studied some of these studies that analyze eye movements recording during reading by various machine learning algorithms.

With the popularization of informationization teaching, the item bank system becomes a significant instrument for assistant teaching and plays a key role in improving the quality of teaching. This research analyses the shortcomings of the exiting item bank system in examination paper on the background of Chinese education. A design of auto-generating examination paper algorithm based on hybrid genetic algorithm is proposed to reduce the one-sided and subjective phenomenon in testing paper. Firstly, this paper establishes a mathematical model based on the attributes to be studied of testing paper and adds corresponding constraints to the model. Then, the genetic algorithm incorporates some features for testing paper and optimizes the global non-convergence and precocious phenomena by combining simulated annealing algorithm. Finally, the feasibility and effectiveness of the new algorithm paper model is verified, which can provide reference for promoting the intelligent and efficient development of teaching.

The increase in the number of online published research papers can be attributed to the recent developments of the internet and web technologies. However, researchers and online users have a difficult time getting relevant and accurate information due to information explosion on the internet. In this paper, we seek to establish which algorithms and similarity metric combinations can be used to optimise the search and recommendation of articles in a research paper recommender systems. Our investigation utilised non-linear classification algorithms with text similarity measures. An offline evaluation approach is utilised to determine the models accuracy and performance, while various similarity metrics are assessed using available datasets. We will utilise the Recursive PARTitioning (rpart), Random Forest and Boosted machine learning algorithms on research paper similarity evaluation datasets. The rpart algorithm generally performed well when compared to the Boosted and the Random Forest algorithms by getting an average accuracy and time efficiency of 80.73 and 2.354628 seconds respectively. The cosine similarity performed best when compared with the other similarity metrics. New similarity metrics and measures are going to be proposed. It has been established in this paper that there are better combinations of metrics and algorithms when attempting to develop models that can be used for research paper similarity evaluation and recommendation. Further challenges and open issues are identified.

For the restoration problem of shredded paper broken by shredder machines with the same marginal feature, a new method based on ant colony algorithm with classification is proposed in this paper. Firstly, shredded paper feature vector can be extracted by image space information. Secondly, the similar matrix and the marginal distance matrix are defined by the feature vector and left-right part image information respectively for every pieces of paper. At last, the original order of shredded paper can be ensured by ant colony algorithm depending on the minimum marginal distance sum among adjacent pieces of paper. For the stitching and restoration of large-scale broken pieces of paper without obvious marginal feature, the classification processing in this paper can greatly reduce the solution scale and improve the algorithm efficiency. The actual experimental result shows the good stability, efficiency and strong robustness of the algorithm.

Traditional data mining algorithms are exceptional at seeing patterns in data that humans cannot, but are often confused by details that are obvious to the organic eye. Algorithms that include humans “in-the-loop” have proved beneficial for accuracy by allowing a user to provide direction in these situations, but the slowness of human interactions causes execution times to increase exponentially. Thus, we seek to formalize frameworks that include humans “over-the-loop”, giving the user an option to intervene when they deem it necessary while not having user feedback be an execution requirement. With this strategy, we hope to increase the accuracy of solutions with minimal losses in execution time. This paper describes our vision of this strategy and associated problems.

This paper considers a media access control (MAC) scheme, which employs Reed-Solomon codes for error-correction, protocol sequences for packet scheduling, and successive interference cancellation (SIC) for demodulation. The construction of the protocol sequences is based on the Chinese remainder theorem (CRT), and a bound on the pairwise Hamming cross-correlation is established. After proving some properties of Hamming cross-correlation of CRT sequences, we illustrate how to allocate the code rates for a system with SIC. A lower bound on system throughput is established.

In this paper, we propose a novel low-complexity signal detection technique based on compressed sensing for a full-duplex generalized spatial modulation (FD-GSM) system, where a communication node transmits data symbols via some antennas and receives data symbols via the remaining antennas at the same time. Thus, each antenna operates as either transmitter or receiver in each symbol time, which implies FD-GSM can be implemented with half-duplex antennas. In particular, parallel orthogonal matching pursuit (POMP) algorithm is exploited as a sparse signal recovery algorithm for detecting GSM signals. The self-interference (SI) due to full-duplex operation is assumed to be completely removed by help of the recently proposed SI cancellation techniques. The proposed signal detection technique significantly outperforms the conventional OMP algorithm in terms of symbol error rate (SER). Interestingly, there exists an optimal number of active antennas for maximizing effective throughput.

We propose a close-to-capacity dirty-paper coding framework which employs multi-level low density parity-check (LDPC) and trellis coded quantization. The proposed coding framework is robust in the sense that it performs close to capacity in the high as well as the low rate regimes. This is in contrast to existing practical DPC schemes which perform well at one of these regimes, but never both. In order to evaluate the performance of our scheme, we consider its application to a cognitive radio channel. At a block length of $2\times 10^5$, the designed dirty-paper coding scheme operates within 0.95, 0.58 and 0.6 dB of the theoretical limit at transmission rates of 0.5, 1.0 and 1.5 bits/sample, respectively. As far as the authors are aware, this is the best performance reported in the literature so far.

This paper studies linear beamforming based on the generalized singular value decomposition (GSVD) for the two-receiver multiple-input multiple-output (MIMO) Gaussian broadcast channel with confidential messages. The transmitter has two independent messages, each of which is intended for one of the receivers but needs to be kept as secret as possible from the other. Recently, it has been proved that, under an input power-covariance constraint, the secret dirty paper coding (S-DPC) scheme is optimal, but under the average power constraint, there is not a computable secrecy capacity expression for the general MIMO case. In fact, for this case, the secrecy capacity region must in general be found through an exhaustive search over the set of all possible matrix power constraints. Clearly, this exhaustive search, as well as the complexity of dirty-paper encoding and decoding, motivates us to consider low complexity linear beamforming technique whose performance is close to the optimal S-DPC scheme. In this paper, we propose a GSVD-based beamforming scheme for the general MIMO broadcast channel with confidential messages. Moreover, an optimal power allocation is obtained to maximize the sum-secrecy rates for the GSVD-based beamforming technique, under the average power constraint. Numerical results are presented to illustrate that the secrecy rate region of the linear precoding approach is nearly identical to that of the optimal S-DPC scheme.

By analyzing the method of present generating test paper, the article represents a method of generating test paper based on genetic algorithm. A mathematical model for generating test paper is built. By putting forward the intercross operator and mutation operator in consistence with the independent coding, an automatic generation system for test paper is realized.

This work presents a self-adaptation algorithm to automatically adjust the peaking settings of a continuous-time linear equalizer (CTLE) in a high-speed PAM4 receiver. A statistical approach is adopted to improve the robustness and flexibility of the adaptation algorithm. The PAM4 top level distribution around the peak value of several consecutive top levels guides the CTLE to attain the optimal digital settings. By performing the peak detection at the output of the equalization summer, the effect of digital equalization is inherently accounted for by the proposed algorithm. To demonstrate the effectiveness of the proposed algorithm, a PAM4 receiver employing two stages of CTLE and 1-tap feed forward equalizer (FFE) is designed to compensate up to 11.8-dB channel loss for a data rate of 56 Gb/s.

A novel approach is here presented to embed a low-priority data stream into a high-priority data stream for OFDM wireless communications. The main improvement with respect to conventional hierarchical modulations is that the low-priority data stream can be decoded from the received subcarrier symbols without any previous equalization in case of multipath fading channel. As a consequence, the expected performance is independent of the channel estimation method used to equalize the channel. The low-priority signal is superimposed on the high-priority symbols using a dirty paper coding technique. As a result of the analysis of the proposed method, several improvements to reduce the interference on the high-priority symbols are introduced. The performance of our scheme has been compared with that of a conventional hierarchical modulation using DVB-T system parameters, showing a considerable improvement for multipath fading channels.

Based on the analysis of the types and characteristics of Chinese traditional paper-cut patterns, we propose a new algorithm for extracting and vectoring the patterns. On the contour extraction, the algorithm uses a grayscale, binary, contour tracking method. On the vectorization, we improved the segment Displaced curve segments so that you can make the final contour paper cutting pattern vector into a control point and a curve. Using this algorithm, we can quickly get the vector contour of the pattern of manual cutting. Further, we can carry on the processing, such as filling, deformation, combination, animation control, so as to complete the core function of the paper cutting animation software. The algorithm is also suitable for the extraction and the vector of common bitmap image, although the algorithm is put forward for the particularity of the paper cutting patterns.

Costa's "writing on dirty paper" capacity result establishes that full state pre-cancellation can be attained in the Gel'fand-Pinsker channel with additive state and additive Gaussian noise. The "carbon copy onto dirty paper" channel is the extension of Costa's model to the compound setting: here a common message is to be decoded at M receivers, each observing the sum of the channel input, Gaussian noise and one of M possible state sequences. As in the Gel'fand-Pinsker channel, the state sequences are anti-causally known at the encoder, which can thus pre-code its transmissions so as to minimize the adverse effects of the channel states. In this paper we derive the capacity to within 2.25 bits-per-channel-use of the M-user "carbon copying onto dirty paper" channel in which the state sequences are Gaussian distributed with the same variance and the same pairwise correlation.