The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
Stochastic Gradient Descent (SGD) is the method of choice for large scale problems, most notably in deep learning. Recent studies target improving convergence and speed of the SGD algorithm. In this paper, we equip the SGD algorithm and its advanced versions with an intriguing feature, namely handling constrained problems. Constraints such as orthogonality are pervasive in learning theory. Nevertheless...
This paper studies the average consensus problem for a discrete-time multi-agent system with first-order dynamics, we provide the sufficient and necessary stable condition that the step-size needs to satisfy for the multi-agent systems with an iterative method, and the cause of oscillation is investigated with use of matrix diagonal and eigenvalue analysis method. Some conclusions about the largest...
This paper deals with the design of a Proportional Multiple Integral (PMI) observer for Takagi-Sugeno systems with immeasurable premise variables subjected to unknown inputs. The introduction of an output error injection in the premise variable of the PMI observer allows to significantly reduce the conservatism of LMI-based stability conditions. A simulation example of a chaotic system is given to...
Iterative learning control is applicable to systems that make sweeps or passes through dynamics defined over a finite duration. Once each pass is complete all information generated as its dynamics evolve are available for use in designing the control action to be applied on the next sweep. The design problem is to construct a sequence of control inputs to enforce convergence to a specified reference...
Multiview canonical correlation analysis (MCCA) is an effective tool for analyzing the relationships among group- aligned multidimensional samples, which has been applied to the fields of pattern recognition and computer vision. In MCCA, its first-stage canonical variables are solved by a multivariate eigenvalue problem that can be computed by Horst method. However, how to use the algorithm for effectively...
In this paper we consider an algebraic Riccati equation arising from discrete linear quadratic optimal control problem. We discuss the solution of the discrete algebraic Riccati equation using modification of Newton's method. This modification consists of Exact Line search and Double Newton step. This method is applied to find the maximal symmetric solution of a discrete algebraic Riccati equation...
This paper presents a new model order selection technique for signal processing applications related to source localization or subspace orthogonal projection techniques in large dimensional regime (Random Matrix Theory) when the noise environment is Complex Elliptically Symmetric (CES) distributed, with unknown scatter matrix. The proposed method consists first in estimating the Toeplitz structure...
In this paper, the eigenvalue problem for self-adjoint linear Hamiltonian systems with nonlinear dependence on the spectral parameter and parameter dependent boundary conditions is considered. A Newton-type iterative solution method is presented. The method has a quadratic convergence and provides a two-sided estimate of the eigenvalue. With the help of this technique the boundary control problem...
We present the theory of sequences of random graphs and their convergence to limit objects. Sequences of random dense graphs are shown to converge to their limit objects in both their structural properties and their spectra. The limit objects are bounded symmetric functions on [0,1]2. The kernel functions define an equivalence class and thus identify collections of large random graphs who are spectrally...
This work develops a distributed optimization algorithm with guaranteed exact convergence for a broad class of left-stochastic combination policies. The resulting exact diffusion strategy is shown to have a wider stability range and superior convergence performance than the EXTRA consensus strategy. The exact diffusion solution is also applicable to non-symmetric left-stochastic combination matrices,...
In this paper, we prove the convergence and determine the best convergence factor of the gradient-based iterative algorithm for the equation Ax = b by analyzing the eigenvalues of the relevant matrices. A new property of eigenvalues related to the symmetric positive definite matrix is established. By using this property, we obtain a family of iterative algorithms for the linear matrix equation and...
Composite function minimization captures a wide spectrum of applications in both computer vision and machine learning. It includes bound constrained optimization and cardinality regularized optimization as special cases. This paper proposes and analyzes a new Matrix Splitting Method (MSM) for minimizing composite functions. It can be viewed as a generalization of the classical Gauss-Seidel method...
The objective of this paper is to extend the relaxed conditions on the existence of the backstepping controllers and observers obtained in a recent literature to the case of parabolic Partial Differential Equations (PDEs) with the same diffusion. Moreover, an observer-based output feedback controller is designed for the considered systems with the new conditions. Then, the corresponding exponential...
In this paper we revisit the well known and popular Normalized Subband Adaptive Filter (NSAF). Based on an analysis of the algorithm in the mean and using an analysis strategy presented in [1], we find that the NSAF can be seen as a Richardson iteration applied to a preconditioned augmented Wiener-Hopf equation. This equation is formulated in such a way that its convergence speed can be predicted...
This paper is concerned with the transient responses of a distributed multi-agent optimization protocol based on the alternating direction method of multipliers (ADMM). Utilizing a representation of an ADMM-based distributed optimization protocol as a feedback system with a linear dynamical system and a nonlinear relation implied by the subgradient, analysis of the transient responses is provided...
We consider generalized resistive systems, comprising linear Kirchhoff equations and non-linear element equations, depending on the flow through the element and on two adjacent nodal variables. The derivatives of the element equation should possess a special signature. For such systems we prove the global non-degeneracy of the Jacobi matrix and the applicability of globally convergent solution tracing...
The Cauchy problem for a linear second-order parabolic equation with 1-periodic measurable coefficients is studied in R + = {(x, t) : x ∊ Rd, t ≥ 0}, d ≥ 2. We focus at the behaviour of the fundamental solution as t → ∞. Its approximations are found with pointwise and integral error estimates of order O (t−(d+j+1)/2) and O(t−(j+1)/2), j = 0,1,…, as t → ∞, respectively. These results are applied to...
This paper presents a continuous fixed-time observer-based controller driving all states of an n-dimensional chain of integrators to the origin for a finite pre-established (fixed) time using a scalar input, when only the highest relative degree state can be measured. The uniform upper bound for the controller convergence time is calculated. Performance of the developed controller is demonstrated...
Recovery of graph signals from a limited number of sampled components is investigated. An iterative recovery algorithm is motivated and derived, including an analytic upper bound for the error. Simulation results are also presented.
The purpose of this work is to improve the tracking performance of the iterative learning control (ILC) by designing a new learning law that has the ability to update the input along both the time and iterative axes. First, the reference is generated by a high-order internal model (HOIM) along the iterative axis and can be approximated by an HOIM along the time axis. Then, the HOIM-based repetitive...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.