A “Nonconvex+Nonconvex” approach for image restoration with impulse noise removal

In this paper, we propose a new method for image restoration problems, which are degraded by impulsive noise, with nonconvex data fitting term and nonconvex regularizer.The proposed method possesses the advantages of nonconvex data fitting and nonconvex regularizer simultaneously, namely, robustness for impulsive noise and efficiency for restoring neat edge images.Further, we propose an efficient algorithm to solve the “Nonconvex+Nonconvex” structure problem via using the alternating direction minimization, and prove that the algorithm is globally convergent when the regularization parameter is known. However, the regularization parameter is unavailable in general. Thereby, we combine the algorithm with the continuation technique and modified Morozov’s discrepancy principle to get an improved algorithm in which a suitable regularization parameter can be chosen automatically. The experiments reveal the superior performances of the proposed algorithm in comparison with some existing methods.     

A generating set search method using curvature information

Direct search methods have been an area of active research in recent years. On many real-world problems involving computationally expensive and often noisy functions, they are one of the few applicable alternatives. However, although these methods are usually easy to implement, robust and provably convergent in many cases, they suffer from a slow rate of convergence. Usually these methods do not take the local topography of the objective function into account. We present a new algorithm for unconstrained optimisation which is a modification to a basic generating set search method. The new algorithm tries to adapt its search directions to the local topography by accumulating curvature information about the objective function as the search progresses. The curvature information is accumulated over a region thus smoothing out noise and minor discontinuities. We present some theory regarding its properties, as well as numerical results. Preliminary numerical testing shows that the new algorithm outperforms the basic method most of the time, sometimes by significant relative margins, on noisy as well as smooth problems. This work was supported by the Norwegian Research Council (NFR).     

Robust optimization of noisy blackbox problems using the Mesh Adaptive Direct Search algorithm

Blackbox optimization problems are often contaminated with numerical noise, and direct search methods such as the Mesh Adaptive Direct Search (MADS) algorithm may get stuck at solutions artificially created by the noise. We propose a way to smooth out the objective function of an unconstrained problem using previously evaluated function evaluations, rather than resampling points. The new algorithm, called Robust-MADS is applied to a collection of noisy analytical problems from the literature and on an optimization problem to tune the parameters of a trust-region method.     

Image restoration with a high-order total variation minimization method

In this paper, we propose a fast and efficient way to restore blurred and noisy images with a high-order total variation minimization technique. The proposed method is based on an alternating technique for image deblurring and denoising. It starts by finding an approximate image using a Tikhonov regularization method. This corresponds to a deblurring process with possible artifacts and noise remaining. In the denoising step, a high-order total variation algorithm is used to remove noise in the deblurred image. We see that the edges in the restored image can be preserved quite well and the staircase effect is reduced effectively in the proposed algorithm. We also discuss the convergence of the proposed regularization method. Some numerical results show that the proposed method gives restored images of higher quality than some existing total variation restoration methods such as the fast TV method and the modified TV method with the lagged diffusivity fixed-point iteration.     

Wavelet frame based blind image inpainting

Image inpainting has been widely used in practice to repair damaged/missing pixels of given images. Most of the existing inpainting techniques require knowing beforehand where those damaged pixels are, either given as a priori or detected by some pre-processing. However, in certain applications, such information neither is available nor can be reliably pre-detected, e.g. removing random-valued impulse noise from images or removing certain scratches from archived photographs. This paper introduces a blind inpainting model to solve this type of problems, i.e., a model of simultaneously identifying and recovering damaged pixels of the given image. A tight frame based regularization approach is developed in this paper for such blind inpainting problems, and the resulted minimization problem is solved by the split Bregman algorithm first proposed by Goldstein and Osher (2009) [1]. The proposed blind inpainting method is applied to various challenging image restoration tasks, including recovering images that are blurry and damaged by scratches and removing image noise mixed with both Gaussian and random-valued impulse noise. The experiments show that our method is compared favorably against many available two-staged methods in these applications.     

Almost Surely Convergent Global Optimization Algorithm Using Noise-Corrupted Observations

A new recursive algorithm for searching the global minimizer of a function is proposed when the function is observed with noise. The algorithm is based on switches between the stochastic approximation and the random search. The combination of SA with RS is not a new idea in such combination, the difficulty consists in creating a good switching rule and in designing an efficient method to reduce the noise effect. The proposed switching rule is easily realizable, the noise reducing method is effective, and the whole recursive optimization algorithm is simply calculated. It is proved that the algorithm a.s. converges to the global minimizer and is asymptotically normal. In comparison with existing methods, the proposed algorithm not only requires much weaker conditions, but also is more efficient as shown by simulation.     

近似函数高阶导数的高精度积分逼近方法

主要研究稳定计算近似函数的高阶导数的积分逼近方法,方法因由Lanczos提出故也称为Lanczos算法.利用Legendre多项式的正交性,提出了一类逼近近似函数高阶导数的高精度积分方法,即构造出一系列积分算子Dn,h(m)去逼近噪声函数的高阶导数,且这些积分算子具有O(δ(2n+2)/(2n+m+2))的收敛速度,其中δ为近似函数的噪声水平.数值模拟结果表明提出的方法是稳定而有效的.     

Newton's method for stochastic semilinear wave equations driven by multiplicative time-space noise

Semilinear wave equations with additive or one-dimensional noise are treatable by various iterative and numerical methods. We study more difficult models of semilinear wave equations with infinite dimensional multiplicative spatially correlated noise. Our proof of probabilistic second-order convergence of some iterative methods is based on Da Prato and Zabczyk's maximal inequalities.     

A sparsity preserving stochastic gradient methods for sparse regression

We propose a new stochastic first-order algorithm for solving sparse regression problems. In each iteration, our algorithm utilizes a stochastic oracle of the subgradient of the objective function. Our algorithm is based on a stochastic version of the estimate sequence technique introduced by Nesterov (Introductory lectures on convex optimization: a basic course, Kluwer, Amsterdam, 2003). The convergence rate of our algorithm depends continuously on the noise level of the gradient. In particular, in the limiting case of noiseless gradient, the convergence rate of our algorithm is the same as that of optimal deterministic gradient algorithms. We also establish some large deviation properties of our algorithm. Unlike existing stochastic gradient methods with optimal convergence rates, our algorithm has the advantage of readily enforcing sparsity at all iterations, which is a critical property for applications of sparse regressions.     

Application of particle swarm optimization in chaos synchronization in noisy environment in presence of unknown parameter uncertainty

In this paper, particle swarm optimization (PSO) is applied to synchronize chaotic systems in presence of parameter uncertainties and measurement noise. Particle swarm optimization is an evolutionary algorithm which is introduced by Kennedy and Eberhart. This algorithm is inspired by birds flocking. Optimization algorithms can be applied to control by defining an appropriate cost function that guarantees stability of system. In presence of environment noise and parameter uncertainty, robustness plays a crucial role in succeed of controller. Since PSO needs only rudimentary information about the system, it can be a suitable algorithm for this case. Simulation results confirm that the proposed controller can handle the uncertainty and environment noise without any extra information about them. A comparison with some earlier works is performed during simulations.     

A lower bound on convergence rates of nonadaptive algorithms for univariate optimization with noise

This paper considers complexity bounds for the problem of approximating the global minimum of a univariate function when the function evaluations are corrupted by random noise. We take an average-case point of view, where the objective function is taken to be a sample function of a Wiener process and the noise is independent Gaussian. Previous papers have bounded the convergence rates of some nonadaptive algorithms. We establish a lower bound on the convergence rate of any nonadaptive algorithm.     

A New Version of the Price's Algorithm for Global Optimization

We present an algorithm for finding a global minimum of a multimodal,multivariate function whose evaluation is very expensive, affected by noise andwhose derivatives are not available. The proposed algorithm is a new version ofthe well known Price's algorithm and its distinguishing feature is that ittries to employ as much as possible the information about the objectivefunction obtained at previous iterates. The algorithm has been tested on alarge set of standard test problems and it has shown a satisfactorycomputational behaviour. The proposed algorithm has been used to solveefficiently some difficult optimization problems deriving from the study ofeclipsing binary star light curves.     

Surrogate test for noise-contaminated dynamics in the Duffing oscillator

To identify random signals from nonlinear system under stochastic background is very difficult, and standard dynamical methods are generally not applicable. The pseudo-periodic surrogate algorithm recently developed by Small is introduced to test the sample time series in the Duffing oscillator under the Gaussian white noise excitation. The correlation dimensions of the noisy periodic, noise-induced chaotic and random-dominant responses of the system are compared with their corresponding artificial data respectively. Meanwhile, the leading Lyapunov exponents by Rosenstein’s algorithm are also presented to validate the identification idea on the system’s sample time series.     

Nonlinear programming methods in the presence of noise

The problem of minimizing a nonlinear function with nonlinear constraints when the values of the objective, the constraints and their gradients have errors, is studied. This noise may be due to the stochastic nature of the problem or to numerical error.Various previously proposed methods are reviewed. Generally, the minimization algorithms involve methods of subgradient optimization, with the constraints introduced through penalty, Lagrange, or extended Lagrange functions. Probabilistic convergence theorems are obtained. Finally, an algorithm to solve the general convex (nondifferentiable) programming problem with noise is proposed.Originally written for presentation at the 1976 Budapest Symposium on Mathematical Programming.     

A stepsize control algorithm for SDEs with small noise based on stochastic Runge–Kutta Maruyama methods

A variable stepsize control algorithm for solution of stochastic differential equations (SDEs) with a small noise parameter ?? is presented. In order to determine the optimal stepsize for each stage of the algorithm, an estimate of the global error is introduced based on the local error of the Stochastic Runge?CKutta Maruyama (SRKM) methods. Based on the relation of the stepsize and the small noise parameter, the local mean-square stochastic convergence order can be different from stage to stage. Using this relation, a strategy for producing and controlling the stepsize in the numerical integration of SDEs is proposed. Numerical experiments on several standard SDEs with small noise are presented to illustrate the effectiveness of this approach.     

A least-squares approach for discretizable distance geometry problems with inexact distances

A branch-and-prune (BP) algorithm is presented for the discretizable distance geometry problem in $$\mathbb {R}^K$$ with inexact distances. The algorithm consists in a sequential buildup procedure where possible positions for each new point to be localized are computed by using distances to at least K previously placed reference points and solving a system of quadratic equations. Such a system is solved in a least-squares sense, by finding the best positive semidefinite rank K approximation for an induced Gram matrix. When only K references are available, a second candidate position is obtained by reflecting the least-squares solution through the hyperplane defined by the reference points. This leads to a search tree which is explored by BP, where infeasible branches are pruned on the basis of Schoenberg’s theorem. In order to study the influence of the noise level, numerical results on some instances with distances perturbed by a small additive noise are presented.     

A block encryption algorithm based on dynamic sequences of multiple chaotic systems

A block encryption algorithm using dynamic sequences generated by multiple chaotic systems is proposed in this paper. In this algorithm, several one-dimension chaotic maps generate pseudo-random sequences, which are independent and approximately uniform. After a series of transformations, the sequences constitute a new pseudo-random sequence uniformly distributing in the value space, which covers the plaintext by executing Exclusive-OR and shifting operations some rounds to form the cipher. This algorithm makes the pseudo-random sequence possess more concealment and noise like characteristic, and overcomes the periodic malpractice caused by the computer precision and single chaotic system. Simulation results show that the algorithm is efficient and useable for the security of communication system.     

基于数据挖掘技术的股票收益率方向研究

本文依据数据挖掘技术对股票收益率的变化方向进行探究。通过小波多尺度分解,将股票价格转化为不同频率域下的子序列数据、并对其中的高频序列进行降噪。构建极度梯度提升树(XGBoost)、以及其它主流机器学习算法,对沪深300和中证500指数中成分股的涨跌进行了拟合并预测。研究发现XGBoost的平均准确率分别达到了54.69%和55.13%,同时依据预测信号构建的投资策略可产生稳定收益,表明该方法具备较强的预测能力。在此基础上,对机器学习算法存在的“黑箱”问题进行了阐述和研究,对模型选股的逻辑进行了探析:提出一种因子权重的度量方法,研究发现市净率、市盈率、能量潮等指标在模型中是较为重要的判别指标,并通过偏相依关系度量了模型中各因子对于股价涨跌方向的边际影响,得到模型倾向于选择市盈率、市净率较小的股票等一些结论,使算法的逻辑更为清楚。     

An Adaptive Subdivision Algorithm for the Identification of the Diffusion Coefficient in Two-dimensional Elliptic Problems

In this work, we consider the identification problem of the diffusion coef-ficient in two-dimensional elliptic equations. For parameterization, we use the zonation method: the diffusion coefficient is assumed to be a piecewise constant space function and unknowns are both the diffusion coefficient values and the geometry of the zones. An algorithm based on geometric principles is developed in order to determine the boundaries between the zones. This algorithm uses the refinement indicators which are easily computed from the gradient of the objective function. The efficiency of the algorithm is proved by testing it in some simple cases with and without noise on the data.      

Reconstruction of convex bodies from surface tensors

We present two algorithms for reconstruction of the shape of convex bodies in the two-dimensional Euclidean space. The first reconstruction algorithm requires knowledge of the exact surface tensors of a convex body up to rank s for some natural number s. When only measurements subject to noise of surface tensors are available for reconstruction, we recommend to use certain values of the surface tensors, namely harmonic intrinsic volumes instead of the surface tensors evaluated at the standard basis. The second algorithm we present is based on harmonic intrinsic volumes and allows for noisy measurements. From a generalized version of Wirtinger's inequality, we derive stability results that are utilized to ensure consistency of both reconstruction procedures. Consistency of the reconstruction procedure based on measurements subject to noise is established under certain assumptions on the noise variables.