The development of intrusion detection systems (IDS) that are adapted to allow routers and network defence systems to detect malicious network traffic disguised as network protocols or normal access is a critical challenge. This paper proposes a novel approach called SCDNN, which combines spectral clustering (SC) and deep neural network (DNN) algorithms. First, the dataset is divided into k subsets based on sample similarity using cluster centres, as in SC. Next, the distance between data points in a testing set and the training set is measured based on similarity features and is fed into the deep neural network algorithm for intrusion detection. Six KDD-Cup99 and NSL-KDD datasets and a sensor network dataset were employed to test the performance of the model. These experimental results indicate that the SCDNN classifier not only performs better than backpropagation neural network (BPNN), support vector machine (SVM), random forest (RF) and Bayes tree models in detection accuracy and the types of abnormal attacks found. It also provides an effective tool of study and analysis of intrusion detection in large networks.

Magnetoelectric multiferroics in which ferroelectricity and magnetism coexist have attracted extensive attention because they provide great opportunities for the mutual control of electric polarization by magnetic fields and magnetization by electric fields. From a practical point view, the main challenge in this field is to find proper multiferroic materials with a high operating temperature and great magnetoelectric sensitivity. Here we report on the magnetically tunable ferroelectricity and the giant magnetoelectric sensitivity up to 250 K in a Y-type hexaferrite, BaSrCoZnFe 11 AlO 22 . Not only the magnitude but also the sign of electric polarization can be effectively controlled by applying low magnetic fields (a few hundreds of Oe) that modifies the spiral magnetic structures. The magnetically induced ferroelectricity is stabilized even in zero magnetic field. Decayless reproducible flipping of electric polarization by oscillating low magnetic fields is shown. The maximum linear magnetoelectric coefficient reaches a high value of ~ 3.0×10 3 ps/m at 200 K.In the past several years, spiral magnetic order induced multiferroics and magnetoelectric (ME) effects have been observed in a number of transition metal oxides such as TbMnO 3 , RMn 2 O 5 , CoCr 2 O 4 , and others [1][2][3] . In these spiral magnets, the magnetic order and ferroelectricity are inherently coupled and thus pronounced ME effects could be expected. 4,5 The microscopic mechanism has been well described with the spin current model 6 or the inverse Dzyaloshinskii-Moriya (DM) interaction model 7 . However, the ME effects in these spiral magnets are not useful for practical applications because they occur at low temperatures and require a large magnetic field of several tesla. Recently, the hexaferrites with helical spin order have been suggested as promising candidates for high temperature multiferroics. It was reported that some Y-type hexaferrites, such as (Ba,Sr) 2 Zn 2 Fe 12 O 22 and Ba 2 Mg 2 Fe 12 O 22 , can show magnetically induced ferroelectricity and pronounced ME effects due to modifications of spiral magnetic structures by applying magnetic fields [8][9][10][11] . Although the magnetic ordering temperatures of these Y-type hexaferrites are above room temperature, their ME effects are observable only below ~ 130 K. Subsequently, ME effects were also observed in Z-type 12,13 , M-type 14 , and U-type 15 hexaferrites.Especially, the low field ME effect in a Z-type hexaferrite, Sr 3 Co 2 Fe 24 O 41 , happens at room temperature, representing a big step towards practical applications 12. Nevertheless, there are still some critical problems to be overcome. For instance, although magnetic control of electric polarization at room temperature has been achieved, the reversal of electric polarization by magnetic fields has been realized only at low temperatures 10,11,14 . We prepared the Y-type hexaferrites by solid state reaction in oxygen. The as-sintered samples are not insulating enough at high temperatures, and thus a post-annealing...

Abstract Background The three-month difference in turnaround time between Xpert and conventional phenotypic drug susceptibility testing (pDST) causes patient treatment challenges when pDST rifampin (RIF) susceptibility results and earlier Xpert results disagree, resulting in unnecessary tuberculosis (TB) patient exposure to toxic second-line drugs. Here, the prevalence of discordant RIF susceptibility test results, specifically Xpert (resistant) versus pDST (susceptible) results, was determined. Methods Tuberculosis patients enrolled between January 2015 and June 2018 at Beijing Chest Hospital who consecutively tested positive for RIF resistance using Xpert then negative using pDST were studied. DNA sequences and minimal inhibitory concentration (MIC) results provided insights for understanding discordant results. Results Of 26,826 patients with suggestive TB symptoms undergoing Xpert MTB/RIF testing, 728 diagnosed as RIF-resistant were evaluated. Of these, 118 (16.2%) exhibiting Xpert RIF resistance and phenotypic RIF susceptibility yielded 104 successfully subcultured isolates; of these, 86 (82.7%) harbored rpoB gene RRDR mutations and 18 (17.3%) did not. Leu511Pro (25.0%) and Leu533Pro (17.3%) mutants were most frequently associated with discordant RIF susceptibility test results. Of the 86 isolates with rpoB mutations, 42 (48.8%) with MICs ≤1.0 mg/L were assigned to the RIF-susceptible group, with Leu511Pro the most common mutation observed. Isolates with very low bacterial load were most frequently misdiagnosed as RIF-resistant by Xpert. Conclusion Approximately one-sixth of RIF-resistant TB isolates identified via Xpert yielded discordant pDST results due to questionable interpretation of specific “disputed” mutations. Thus, a diagnostic flow chart should be used to correctly interpret Xpert RIF-resistance results to best guide patient treatment.

Graph sampling with noise is a fundamental problem in graph signal processing (GSP). Previous works assume an unbiased least square (LS) signal reconstruction scheme and select samples greedily via expensive extreme eigenvector computation. A popular biased scheme using graph Laplacian regularization (GLR) solves a system of linear equations for its reconstruction. Assuming this GLR-based scheme, we propose a reconstruction-cognizant sampling strategy to maximize the numerical stability of the linear system-i.e., minimize the condition number of the coefficient matrix. Specifically, we maximize the eigenvalue lower bounds of the matrix, represented by left-ends of Gershgorin discs of the coefficient matrix. To accomplish this efficiently, we propose an iterative algorithm to traverse the graph nodes via Breadth First Search (BFS) and align the leftends of all corresponding Gershgorin discs at lower-bound threshold T using two basic operations: disc shifting and scaling. We then perform binary search to maximize T given a sample budget K. Experiments on real graph data show that the proposed algorithm can effectively promote large eigenvalue lower bounds, and the reconstruction MSE is the same or smaller than existing sampling methods for different budget K at much lower complexity.

No abstract

Learning a suitable graph is an important precursor to many graph signal processing (GSP) pipelines, such as graph signal compression and denoising. Previous graph learning algorithms either i) make assumptions on graph connectivity (e.g., graph sparsity), or ii) make edge weight assumptions such as positive edges only. In this paper, given an empirical covariance matrixC computed from data as input, we consider an eigen-structural assumption on the graph Laplacian matrix L: the first K eigenvectors of L are pre-selected,

No abstract

No abstract