This paper presents a Multiple Kernel Learning (abbreviated as MKL) framework for the Support Vector Machine (SVM) with the (0, 1) loss function in the context of the binary classification task. Some KKT-like first-order optimality conditions are provided and then exploited to develop a fast ADMM algorithm to solve the nonsmooth nonconvex optimization problem. Convergence analysis of the ADMM is presented under certain technical assumptions. Numerical experiments on real data sets show that our MKL-L0/1-SVM can potentially outperform one of the leading approaches called SimpleMKL developed by Rakotomamonjy, Bach, Canu, and Grandvalet [Journal of Machine Learning Research, vol. 9, pp. 2491-2521, 2008] in terms of the accuracy of classification and the sparsity in the combination of kernels.
This paper presents a malware-detection pipeline for Windows API-call sequences that converts variable-length traces into fixed-length behavioral features. API tokens are embedded with Word2Vec and clustered with K-means to form a shared behavioral state space, and each trace is summarized by first-order Markov transition frequencies (S2 features). Experiments on 150,656 traces with a family-aware split (20% of families held out for testing) show that linear baselines trained on the full transition representation provide strong and scalable performance. Quantum models are evaluated under explicit computational limits by applying TruncatedSVD to low-dimensional subsets and comparing a fidelity-kernel QSVM and a variational quantum classifier against a matched classical SVC-RBF on identical stratified subsets. Beyond ROC/PR curves, security-oriented operating points are reported by selecting thresholds on validation to enforce high malware recall (Rmal >= 0.90), highlighting the recall trade-off and the scaling cost of quantum kernel evaluation as qubit count increases.
This paper studies the connected degree of G-V fuzzy matroids. The concept is generalized from the concept of matroid connectedness through plausible reasoning and has good expansibility and practicability. It is proved that a G-V fuzzy matroid M is disconnected if and only if its connected degree is the minimum (that is, Con(M) = 0), that is to say, it is connected if and only if Con(M) =/ 0. It is also proved that the connected degree Con(M(Ge)) = 1 of the fuzzy cycle matroid M(Ge) of a fuzzy graph Ge if and only if the fuzzy graph Ge is 2-connected. Finally, combined with the existing connected degree theory of fuzzifying matroids, and relying on the transformation relationship between closed G-V fuzzy matroids and fuzzifying matroids, the consistency of the definitions of the connected degrees of the two types of matroids is demonstrated, revealing the inherent unity of different fuzzy matroid models in connectedness analysis.
In this paper, a robust H infinity observer design method is proposed for a class of nonlinear descriptor systems subject to measurement noise and unknown disturbances. Specifically, a novel augmented system is constructed based on the output filtering method. Next, an observer resisting measurement noise is presented and the state estimation error system is derived. The observer parameters are parameterized by employing the generalized inverse technique for matrices. Finally, sufficient conditions ensuring the asymptotic stability of the state estimation error system with the H infinity performance index are derived on the basis of the linear matrix inequalities (LMIs). Compared to the conventional robust H infinity observers, simulation results demonstrate that the proposed method alleviates the adverse effects of measurement noise on the state estimation and results in smaller estimation errors.
This paper addresses the problem of refining settling-time upper-bound estimates for fixedtime stable systems under a three-term Lyapunov characterization. Without resorting to inequality-based techniques or special functions, a closed-form analytical upper bound for the settling-time function is derived using only elementary functions. Comparative analyses demonstrate that the proposed bound is tighter than existing results. The theoretical findings are further validated through the design of a fixed-time controller for a memristive neural network, with numerical simulations confirming the accuracy of the estimated settling time and the independence of the control scheme from initial conditions.
