A recurrent perceptron learning algorithm for cellular neural networks

A supervised learning algorithm for obtaining the template coefficients in completely stable Cellular Neural Networks (CNNs) is analysed in the paper. The considered algorithm resembles the well-known perceptron learning algorithm and hence called as Recurrent Perceptron Learning Algorithm (RPLA) when applied to a dynamical network. The RPLA learns pointwise defined algebraic mappings from initial-state and input spaces into steady-state output space; despite learning whole trajectories through desired equilibrium points. The RPLA has been used for training CNNs to perform some image processing tasks and found to be successful in binary image processing. The edge detection templates found by RPLA have performances comparable to those of Canny's edge detector for binary images.     

Stable reformulation of transfer matrix method for wave propagation in layered anisotropic media.

The numerical instability problem in the standard transfer matrix method has been resolved by introducing the layer stiffness matrix and using an efficient recursive algorithm to calculate the global stiffness matrix for an arbitrary anisotropic layered structure. For general anisotropy the computational algorithm is formulated in matrix form. In the plane of symmetry of an orthotropic layer the layer stiffness matrix is represented analytically. It is shown that the elements of the stiffness matrix are as simple as those of the transfer matrix and only six of them are independent. Reflection and transmission coefficients for layered media bounded by liquid or solid semi-spaces are formulated as functions of the total stiffness matrix elements. It has been demonstrated that this algorithm is unconditionally stable and more efficient than the standard transfer matrix method. The stiffness matrix formulation is convenient in satisfying boundary conditions for different layered media cases and in obtaining modal solutions. Based on this method characteristic equations for Lamb and surface waves in multilayered orthotropic media have been obtained. Due to the stability of the stiffness matrix method, the solutions of the characteristic equations are numerically stable and efficient. Numerical examples are given.     

高效选取线性独立的原子特征构建精确机器学习势函数

机器学习势由于具有与第一性原理计算相当的准确性,且低得多的计算成本,在原子模拟中极具前景. 然而原子机器学习势的可靠性、速度和可迁移性在很大程度上取决于原子构型的表示. 适当地选取用作机器学习程序输入的描述符是一个成功的机器学习表示的关键. 本文发展了一种简单有效的方法,可以基于训练数据固有的相关性,从大量待选的描述符中自动选取一组最佳的线性独立原子特征. 通过对几个具有较少冗余线性独立嵌入密度描述符的基准分子构建嵌入原子神经网络势的应用,证明了这种新方法的有效性和准确性. 该算法可以大大简化原子特征的初始选取,并极大地提高原子机器学习势的性能.     

MIGAN: Mutual-Interaction Graph Attention Network for Collaborative Filtering

Many web platforms now include recommender systems. Network representation learning has been a successful approach for building these efficient recommender systems. However, learning the mutual influence of nodes in the network is challenging. Indeed, it carries collaborative signals accounting for complex user-item interactions on user decisions. For this purpose, in this paper, we develop a Mutual Interaction Graph Attention Network “MIGAN”, a new algorithm based on self-supervised representation learning on a large-scale bipartite graph (BGNN). Experimental investigation with real-world data demonstrates that MIGAN compares favorably with the baselines in terms of prediction accuracy and recommendation efficiency.     

飞轮微振动与整星耦合特性参数补偿算法研究

为了提高飞轮与整星耦合特性影响成像像移理论预估的精度,提出了一种星上飞轮安装刚度和阻尼参数的数值补偿算法.首先,建立补偿算法数学模型,基于飞轮径向摇摆模态以及线性整星微振动传递函数模型,通过联合飞轮在专用测量平台上的扰振实测数据以及星上安装后所致成像像移实测数据构建优化函数,实现对星上飞轮安装刚度以及阻尼参数的补偿.其次,采用该算法对某型号卫星上的飞轮安装刚度以及阻尼进行了补偿和估计,最后对参数补偿前后的飞轮扰振所致像移分别进行理论预估与对比.对比结果表明:参数补偿后的预估像移数据较补偿前预估像移数据更接近实测像移数据且径向摇摆模态固有频率曲线更加明显,证明此参数补偿算法可行.本研究显著降低了整星微振动分析过程中飞轮在星上安装刚度与测试平台安装刚度不一致所带来的分析误差,为飞轮与整星耦合特性安装等效参数求解提供了一种思路.     

Symplectic analysis for periodical electro-magnetic waveguides

Symplectic analysis is introduced into electro-magnetic waveguide theory, by using Hamiltonian system theory in which the transverse electric and magnetic field vectors are the dual vectors. The method can accommodate arbitrary anisotropic material and includes the interface conditions between adjacent segments of the waveguide. An electro-magnetic stiffness matrix is introduced which relates to the two ends of each segment of the waveguide. Both the pass- and stop-band stiffness matrices for plane waveguides with constant cross-section are given analytically and also a transformation matrix is given to permit abrupt changes of cross-section to occur. The variational principle is applied to obtain the segment combination algorithm needed to generate the electro-magnetic stiffness matrix related to the two ends of the fundamental periodical segment. Then the Wittrick-Williams algorithm is used to extract the eigenvalues. Thereafter, an energy band analysis is performed for a periodical waveguide, e.g., a grating, by using the symplectic eigensolutions.     

Learning algorithm for a neural network with binary synapses

A learning mechanism for neural networks with binary synapses is defined and investigated. The algorithm is based on minimizing the energy of an Ising model. A replica symmetric calculation gives a parameter range where perfect learning is possible. A simple descent algorithm is studied by numerical simulation; and storage capacities, learning times and basins of attraction are determined.     

Inverse vibration problem for un-damped 3-dimensional multi-story shear building models

Various researchers have contributed to the identification of the mass and stiffness matrices of two dimensional (2-D) shear building structural models for a given set of vibratory frequencies. The suggested methods are based on the specific characteristics of the Jacobi matrices, i.e., symmetric, tri-diagonal and semi-positive definite matrices. However, in case of three dimensional (3-D) structural models, those methods are no longer applicable, since their stiffness matrices are not tri-diagonal. In this paper the inverse problem for a special class of vibratory structural systems, i.e., 3-D shear building models, is investigated. A practical algorithm is proposed for solving the inverse eigenvalue problem for un-damped, 3-D shear buildings. The problem is addressed in two steps. First, using the target frequencies, a so-called normalized eigenvector matrix, which is a banded matrix containing the information related to the frequencies and mode shapes of the target structural system, is determined. In this regard, similar to the solution of inverse problem for 2-D shear building structural models in which an auxiliary structure is constructed by adding constraints (or springs) to the original system, three auxiliary structures are proposed to solve the problem for 3-D cases. In the second step, the normalized eigenvector matrix is utilized to obtain the normalized stiffness matrix; in turn, this matrix is decomposed into the stiffness and mass matrices of the system. Finally, a numerical example is presented to demonstrate the efficiency of the proposed algorithm in determining the mass and stiffness matrices of a 3-D structural model for a given set of target vibrational frequencies.     

大口径光学元件气囊抛光工具刚度可控性

为保证气囊抛光过程中抛光运动的高稳定性和均匀材料去除率,对气囊抛光非球面过程中气囊工具刚度的可控性进行了研究。通过分析气囊抛光大口径光学元件时工具的受力情况,计算了气囊工具的刚度,并分析了气囊抛光工具刚度对抛光时材料去除的影响及气囊工具刚度的影响因素。设计了气囊工具刚度控制算法并进行模拟试验,仿真结果表明,在刚度标准值根据加工要求设定以后,即可通过调节工件对气囊工具的反作用力,使得气囊抛光大口径光学元件过程中气囊工具刚度可控。     

基于分支设置的质量导引相位展开算法

二维相位展开方法是近年来较为活跃、引起关注的一个研究课题,它在许多测量应用中有着重要的作用。尽管掩膜阻断算法在多个领域都有成功应用实例,该算法存在着固有的缺陷。为了克服掩膜阻断算法的缺陷,综合分支阻断方法和质量导引方法的优点,提出一种基于分支设置的质量导引相位展开新算法。它先以一个初始质量图来引导分支的设置,然后把分支对应的相位质量设置为最低,从而产生一个新的质量图,最后按新质量图来引导相位展开,并使用几个包裹相位图来验证此方法的有效性。计算机模拟相位图和实际相位图的相位展开结果表明,在存在复杂轮廓不连续和高噪声的情况下,该算法优于模板阻断算法,能得到较好的相位展开结果。     

Parameter estimation for chaotic systems using the cuckoo search algorithm with an orthogonal learning method

We study the parameter estimation of a nonlinear chaotic system,which can be essentially formulated as a multidimensional optimization problem.In this paper,an orthogonal learning cuckoo search algorithm is used to estimate the parameters of chaotic systems.This algorithm can combine the stochastic exploration of the cuckoo search and the exploitation capability of the orthogonal learning strategy.Experiments are conducted on the Lorenz system and the Chen system.The proposed algorithm is used to estimate the parameters for these two systems.Simulation results and comparisons demonstrate that the proposed algorithm is better or at least comparable to the particle swarm optimization and the genetic algorithm when considering the quality of the solutions obtained.     

Generalized Kohonen's competitive learning algorithms for ophthalmological MR image segmentation

Kohonen's self-organizing map is a two-layer feedforward competitive learning network. It has been used as a competitive learning clustering algorithm. In this paper, we generalize Kohonen's competitive learning (KCL) algorithm with fuzzy and fuzzy-soft types called fuzzy KCL (FKCL) and fuzzy-soft KCL (FSKCL). These generalized KCL algorithms fuse the competitive learning with soft competition and fuzzy c-means (FCM) membership functions. We then apply these generalized KCLs to MRI and MRA ophthalmological segmentations. These KCL-based MRI segmentation techniques are useful in reducing medical image noise effects using a learning mechanism. They may be particularly helpful in clinical diagnosis. Two real cases with MR image data recommended by an ophthalmologist are examined. First case is a patient with Retinoblastoma in her left eye, an inborn malignant neoplasm of the retina frequently metastasis beyond the lacrimal cribrosa. The second case is a patient with complete left side oculomotor palsy immediately after a motor vehicle accident. Her brain MRI with MRA, skull routine, orbital CT, and cerebral angiography did not reveal brainstem lesions, skull fractures, or vascular anomalies. These generalized KCL algorithms were used in segmenting the ophthalmological MRIs. KCL, FKCL and FSKCL comparisons are made. Overall, the FSKCL algorithm is recommended for use in MR image segmentation as an aid to small lesion diagnosis.     

基于改进的径向基函数神经网络的铁磁谐振系统混沌控制

面向中性点直接接地电力系统发生的铁磁谐振过电压所显现的混沌特性，在径向基函数神经网络的基础上，提出引进一种极大熵学习算法对该混沌系统进行控制.该方法通过最优化一个目标函数导出中心向量的学习规则，充分利用网络隐层的聚类功能，极大改善网络的回归和学习能力.对具体的铁磁谐振系统的数值实验证实了该方法在针对铁磁谐振过电压混沌控制中的有效性和可行性. 关键词： 中性点直接接地系统 混沌控制 径向基函数 极大熵原理     

一种新型广义RBF神经网络在混沌时间序列预测中的研究

提出了一种新颖的广义径向基函数神经网络模型,其径向基函数(RBF)的形式由生成函数确定.然后,给出了易实现的梯度学习算法,同时为了进一步提高网络的收敛速度和网络性能,又给出了基于卡尔曼滤波的动态学习算法.为了验证网络的学习性能，采用基于卡尔曼滤波算法的新型广义RBF网络预测模型对Mackey-Glass混沌时间序列和Henon映射进行了仿真.结果表明,所提出的新型广义RBF神经网络模型能快速、精确地预测混沌时间序列,是研究复杂非线性动力系统辨识和控制的一种有效方法. 关键词： 广义径向基函数神经网络 卡尔曼滤波 梯度下降学习算法 混沌时间序列 预测     

Guided filter and adaptive learning rate based non-uniformity correction algorithm for infrared focal plane array

Imaging non-uniformity of infrared focal plane array (IRFPA) behaves as fixed-pattern noise superimposed on the image, which affects the imaging quality of infrared system seriously. In scene-based non-uniformity correction methods, the drawbacks of ghosting artifacts and image blurring affect the sensitivity of the IRFPA imaging system seriously and decrease the image quality visibly. This paper proposes an improved neural network non-uniformity correction method with adaptive learning rate. On the one hand, using guided filter, the proposed algorithm decreases the effect of ghosting artifacts. On the other hand, due to the inappropriate learning rate is the main reason of image blurring, the proposed algorithm utilizes an adaptive learning rate with a temporal domain factor to eliminate the effect of image blurring. In short, the proposed algorithm combines the merits of the guided filter and the adaptive learning rate. Several real and simulated infrared image sequences are utilized to verify the performance of the proposed algorithm. The experiment results indicate that the proposed algorithm can not only reduce the non-uniformity with less ghosting artifacts but also overcome the problems of image blurring in static areas.     

Who Will Score? A Machine Learning Approach to Supporting Football Team Building and Transfers

Background: the machine learning (ML) techniques have been implemented in numerous applications, including health-care, security, entertainment, and sports. In this article, we present how the ML can be used for building a professional football team and planning player transfers. Methods: in this research, we defined numerous parameters for player assessment, and three definitions of a successful transfer. We used the Random Forest, Naive Bayes, and AdaBoost algorithms in order to predict the player transfer success. We used realistic, publicly available data in order to train and test the classifiers. Results: in the article, we present numerous experiments; they differ in the weights of parameters, the successful transfer definitions, and other factors. We report promising results (accuracy = 0.82, precision = 0.84, recall = 0.82, and F1-score = 0.83). Conclusion: the presented research proves that machine learning can be helpful in professional football team building. The proposed algorithm will be developed in the future and it may be implemented as a professional tool for football talent scouts.     

A Multi-Strategy Adaptive Comprehensive Learning PSO Algorithm and Its Application

In this paper, a multi-strategy adaptive comprehensive learning particle swarm optimization algorithm is proposed by introducing the comprehensive learning, multi-population parallel, and parameter adaptation. In the proposed algorithm, a multi-population parallel strategy is designed to improve population diversity and accelerate convergence. The population particle exchange and mutation are realized to ensure information sharing among the particles. Then, the global optimal value is added to velocity update to design a new velocity update strategy for improving the local search ability. The comprehensive learning strategy is employed to construct learning samples, so as to effectively promote the information exchange and avoid falling into local extrema. By linearly changing the learning factors, a new factor adjustment strategy is developed to enhance the global search ability, and a new adaptive inertia weight-adjustment strategy based on an S-shaped decreasing function is developed to balance the search ability. Finally, some benchmark functions and the parameter optimization of photovoltaics are selected. The proposed algorithm obtains the best performance on 6 out of 10 functions. The results show that the proposed algorithm has greatly improved diversity, solution accuracy, and search ability compared with some variants of particle swarm optimization and other algorithms. It provides a more effective parameter combination for the complex engineering problem of photovoltaics, so as to improve the energy conversion efficiency.     

AutoMH: Automatically Create Evolutionary Metaheuristic Algorithms Using Reinforcement Learning

Machine learning research has been able to solve problems in multiple domains. Machine learning represents an open area of research for solving optimisation problems. The optimisation problems can be solved using a metaheuristic algorithm, which can find a solution in a reasonable amount of time. However, the time required to find an appropriate metaheuristic algorithm, that would have the convenient configurations to solve a set of optimisation problems properly presents a problem. The proposal described in this article contemplates an approach that automatically creates metaheuristic algorithms given a set of optimisation problems. These metaheuristic algorithms are created by modifying their logical structure via the execution of an evolutionary process. This process employs an extension of the reinforcement learning approach that considers multi-agents in their environment, and a learning agent composed of an analysis process and a process of modification of the algorithms. The approach succeeded in creating a metaheuristic algorithm that managed to solve different continuous domain optimisation problems from the experiments performed. The implications of this work are immediate because they describe a basis for the generation of metaheuristic algorithms in an online-evolution.     

Partial eigenstructure assignment for undamped vibration systems using acceleration and displacement feedback

A new method for partial eigenstructure assignment using acceleration and displacement feedback for undamped vibration systems is presented in this paper. Firstly, a necessary and sufficient condition is proposed for the incremental mass and stiffness matrices that modify some eigenpairs while keeping other eigenpairs unchanged. Secondly, based on this condition, an algorithm for determining the required control gain matrices of acceleration and displacement feedback, which assign the desired eigenstructure, is developed. This algorithm is easy to implement, and works directly on the second-order system model. More importantly, the algorithm allows the control matrix to be specified beforehand and also leads naturally to a small norm solution of the feedback gain matrices. Finally, some numerical examples are given to demonstrate the effectiveness and accuracy of the proposed algorithm.     

用于压缩感知磁共振成像的分割字典学习算法

字典学习算法可以根据数据本身的特点构建稀疏域中的基,从而使数据的表示更加稀疏.该文在传统的字典学习算法基础上提出了分割字典学习算法,由于部分磁共振图像组织结构简单、可以进行图像分割,因此可根据此特点来优化字典中基函数的构建,使磁共振图像的表达更为稀疏,从而获得更高的重建图像质量.该文利用模拟数据和真实数据进行了重建实验,结果表明与传统的字典学习算法相比,分割字典学习算法能进一步改善重建图像质量.