基于改进sigmoid激活函数的深度神经网络训练算法研究

针对深度神经网络训练过程中残差随着其传播深度越来越小而使底层网络无法得到有效训练的问题,通过分析传统sigmoid激活函数应用于深度神经网络的局限性,提出双参数sigmoid激活函数。一个参数保证激活函数的输入集中坐标原点两侧,避免了激活函数进入饱和区,一个参数抑制残差衰减的速度,双参数结合有效的增强了深度神经网络的训练。结合DBN对MNIST数据集进行数字分类实验,实验表明双参数 sigmoid激活函数能够直接应用于无预训练深度神经网络,而且提高了sigmoid激活函数在有预训练深度神经网络中的训练效果。     

随机并行梯度下降光束净化实验研究

利用自适应光学技术进行光束净化是高能激光系统中一项重要的研究内容.为实现光束净化系统的小型化和低成本,基于系统性能评价函数无模型最优化的波前畸变校正方法是适合的技术方案.就随机并行梯度下降(SPGD)最优化算法在光束净化系统中的应用展开研究.针对高能激光束常见的像差分布进行了SPGD波前校正的数值模拟,在此基础上构建了37单元自适应光学光束净化实验平台,讨论了双边扰动梯度估计和迭代增益系数自适应变化对算法收敛特性的影响.数值模拟与实验结果验证了SPGD算法对不同程度波前畸变的校正能力,表明了SPGD光束净化方案的可行性.     

自适应光学系统随机并行梯度下降控制算法仿真与分析

随机并行梯度下降算法能不依赖波前传感器直接对系统性能进行优化。以32单元变形镜为校正器,采用随机并行梯度下降算法建立了自适应光学系统仿真模型。通过分析该系统对静态波前畸变的校正能力,验证了随机并行梯度下降算法的收敛性;讨论了算法增益系数、随机扰动幅度与收敛速度的关系,并指出通过算法增益系数的自适应调整可以改进算法的收敛速度。     

大功率调制气流声源阵列的相干合成实验研究

完成了调制气流声源阵列的相干合成实验. 提出了利用主动相位控制方法实现调制气流声源阵列相干合成的思路, 介绍了基于随机并行梯度下降算法的声源阵列相干合成的原理. 对利用该算法实现声源阵列的相干合成进行了数值模拟, 完成了双调制气流声源阵列在远场的相干合成实验, 并给出算法参数的合理设置方案. 实验结果显示, 基频成分的相干合成效果明显, 算法收敛时测点处的声压级相比单源发射增加了4 dB, 接近于各单源功率谱中基频成分相干合成、其他频率成分非相干合成的结果; 结果表明实验中算法能够有效控制各调制气流声源辐射声波的相位, 取得了明显的相干合成效果. 关键词： 调制气流声源 相干合成 随机并行梯度下降 高阶谐波     

基于粒子群优化算法的测光红移回归预测

星系的红移在天文研究中极其重要,星系测光红移的预测对研究宇宙大尺度结构及演变有着重要的研究意义。利用斯隆巡天项目发布的SDSS DR13的150 000个星系的测光及光谱数据进行分析,首先根据颜色特征并基于聚类的方法对星系进行分类,由分类结果可知早型星系的占比较大。对比了三种不同的机器学习算法对早型星系进行测光红移回归预测实验,并找出最优的方法。实验中将星系样本中u, g, r, i, z五个波段的测光值以及两两做差得到的10个颜色特征作为输入数据,首先构建BP网络,使用BP算法对星系的测光红移进行回归预测;然后利用遗传算法（GA）优化BP网络各层参数,将优化后的GA-BP算法应用于早型星系的回归预测试验中。考虑到GA算法的复杂操作会影响预测效率,并且粒子群算法（PSO）不仅稳定性高且操作简单,因此将粒子群算法应用到星系样本中早型星系的测光红移回归预测实验中,进而采用粒子群算法优化BP网络（PSO-BP）。实验中将光谱红移作为期望值,采用均方差（MSE）作为误差分析指标来评判三种算法的精度,将PSO-BP回归预测结果与BP网络模型、GA-BP网络模型进行比较。由实验结果可知,BP网络的MSE值为0.001 92,GA-BP网络的MSE值0.001 728,PSO-BP网络的MSE值为0.001 708。实验结果表明,所用到的PSO-BP优化模型在精度上优于BP神经网络模型和GA-BP神经网络模型,分别提高了11.1%和1.2%;在效率上优于传统的K近邻（KNN）测光红移估计算法, 克服了KNN算法中遍历所有数据样本进行训练的缺点并且其泛化性能优于其它BP网络优化模型。     

基于改进蚁群算法的无线传感器网络路由的优化

为了延长无线传感器网络(Wireless Sensor Network ,WSN)的生命周期,均衡各个节点间能量消耗,针对现有的WSN路由优化算法存在的问题,提出了一种基于改进蚁群算法的路由优化算法。首先通过对蚁群算法和遗传算法的优劣性比较,在蚁群算法的基础上,结合遗传算法的选择、交叉和变异的操作,从而提高蚁群算法的搜索速度和寻优能力。最优路径评价函数综合考虑节点能耗及节点的剩余能量,使剩余能量多的节点优先参与数据转发,均衡节点间的能量消耗。通过与经典蚁群算法及遗传算法的对比实验表明,随着数据转发轮数增加,改进的蚁群算法能耗小,剩余能量多,网络生命周期明显延长;随着整个网络运行时间的增长,改进的蚁群算法,节点均衡能耗性好,最优路径搜索的成功率也明显优于其他两种算法。     

Numerical and experimental study on coherent beam combining of fibre amplifiers using simulated annealing algorithm

We present the numerical and experimental study on the coherent beam combining of fibre amplifiers by means of simulated annealing (SA) algorithm. The feasibility is validated by the Monte Carlo simulation of correcting static phase distortion using SA algorithm. The performance of SA algorithm under time-varying phase noise is numerically studied by dynamic simulation. It is revealed that the influence of phase noise on the performance of SA algorithm gets stronger with an increase in amplitude or frequency of phase noise; and the laser array that contains more lasers will be more affected from phase noise. The performance of SA algorithm for coherent beam combining is also compared with a widely used stochastic optimization algorithm, i.e., the stochastic parallel gradient descent (SPGD) algorithm. In a proof-of-concept experiment we demonstrate the coherent beam combining of two 1083~nm fibre amplifiers with a total output power of 12~W and 93% combining efficiency. The contrast of the far-field coherently combined beam profiles is calculated to be as high as 95%.     

基于SPGD算法的非保偏-保偏光自适应偏振转换

报道了一种基于随机并行梯度下降(SPGD)算法的高消光比非保偏-保偏光自适应偏振转换系统。该系统利用偏振控制器对非保偏光的偏振分量进行直接控制,通过SPGD算法对输出的偏振消光比进行优化,最终实现了自适应的非保偏-保偏光的偏振转换。理论上,结合SPGD算法和偏振控制器的原理,对系统进行分析,建立了非保偏-保偏光自适应偏振转换的数学模型。实验上,利用该系统实现了非保偏到保偏光的转换,获得了14.1 dB的线偏振光输出;并利用该系统将任意方向（0~360）偏振态的线偏振光转换为期望偏振态的高消光比线偏光,其输出线偏光的平均消光比约为12 dB。     

基于GA-BP的移动通信设备故障诊断

针对通信设备故障发生随机性强,影响因素多,对应的故障诊断有高度非线性和不确定性的特点,采用BP神经网络算法,优化的GA-BP神经网络算法和POS-BP神经网络算法分别搭建基站设备故障诊断模型,提取设备故障历史数据进行MATLAB仿真,准确预测设备故障类型,帮助提高代维公司调度管理的智能化水平,提高基站设备运维的执行效率。仿真结果表明：本文的BP,GA-BP和POS-BP神经网络算法都能够实现设备故障类别的预测,且GA-BP神经网络算法相比BP和POS-BP神经网络算法对通信设备故障诊断有更好的适应性。     

基于混沌人工蜂群算法的无线传感器网络覆盖优化

针对无线传感器网络随机播撒的节点严重冗余并且导致网络寿命短、覆盖效率不高等缺陷,提出了一种混沌人工蜂群算法的无线传感器网络覆盖优化算法;将节点的利用率和覆盖率作为优化目标函数,建立与之对应的数学模型,之后用混沌人工蜂群算法改善人工蜂群算法陷入局部最优、收敛慢等问题,提高算法收敛速度和精度,对节点覆盖模型进行求解,得出网络最优覆盖方案;通过实验仿真,提出的算法提高了无线传感器网络的覆盖率,覆盖率可达93.48%以上,减少了网络节点冗余,提高了网络寿命,降低了网络成本。     

自适应光学随机并行梯度下降算法波前整形规律仿真

本文首先介绍了基于Zernike模式的SPGD算法对大气湍流畸变波前的整形原理,通过推导得到了关于性能指标的简明表达式,使SPGD算法收敛速率得到明显提升。然后建立了自适应光学随机并行梯度下降算法波前整形系统模型,主要对SPGD算法收敛速率、整形能力和整形效果随波前畸变量和变形镜模型的变化规律作了较为详细的仿真研究,整体定性结果表明：三者的变化规律有一定的相似性,同时利用最小二乘法得到了关于整形能力和整形效果变化规律的定量表达式,若从自适应光学波前整形系统的实时性和简单性考虑,在保证一定整形效果的情况下,选择37单元变形镜对畸变波前的3~27（25）阶Zernike像差进行整形即可。     

基于遗传算法优化BP神经网络的高炉喷煤优化

高炉炼铁是一个复杂的多变量系统,而现行的操作制度是基于炉长经验的参数设置模式,导致能源尤其是煤粉的消耗常常处于“盲目”状态。本文综合炼铁工艺理论和高炉专家经验,针对白云鄂博矿石冶炼的特殊性,采用筛选出的优化数据,利用遗传算法所固有的全局搜索性能优化BP神经网络模型的权值和阈值,分别建立了基于遗传算法优化BP神经网络的高炉喷煤量优化预测模型以及工艺指标(铁水[Si]含量及入炉焦比)预测模型。优化数据的利用使得上述模型可以根据高炉当前炉况输出喷煤量的最佳优化设定值,并预测出相对应的工艺指标变化趋势。实际应用表明,本方法能够给现场操作人员提供操作指导,实现高炉稳定顺行、提高经济效益的目的。     

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     

Training Multilayer Perceptron with Genetic Algorithms and Particle Swarm Optimization for Modeling Stock Price Index Prediction

Predicting stock market (SM) trends is an issue of great interest among researchers, investors and traders since the successful prediction of SMs’ direction may promise various benefits. Because of the fairly nonlinear nature of the historical data, accurate estimation of the SM direction is a rather challenging issue. The aim of this study is to present a novel machine learning (ML) model to forecast the movement of the Borsa Istanbul (BIST) 100 index. Modeling was performed by multilayer perceptron–genetic algorithms (MLP–GA) and multilayer perceptron–particle swarm optimization (MLP–PSO) in two scenarios considering Tanh (x) and the default Gaussian function as the output function. The historical financial time series data utilized in this research is from 1996 to 2020, consisting of nine technical indicators. Results are assessed using Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and correlation coefficient values to compare the accuracy and performance of the developed models. Based on the results, the involvement of the Tanh (x) as the output function, improved the accuracy of models compared with the default Gaussian function, significantly. MLP–PSO with population size 125, followed by MLP–GA with population size 50, provided higher accuracy for testing, reporting RMSE of 0.732583 and 0.733063, MAPE of 28.16%, 29.09% and correlation coefficient of 0.694 and 0.695, respectively. According to the results, using the hybrid ML method could successfully improve the prediction accuracy.     

Stochastic asymptotical synchronization of chaotic Markovian jumping fuzzy cellular neural networks with mixed delays and the Wiener process based on sampled-data control

We investigate the stochastic asymptotical synchronization of chaotic Markovian jumping fuzzy cellular neural networks (MJFCNNs) with discrete, unbounded distributed delays, and the Wiener process based on sampled-data control using the linear matrix inequality (LMI) approach. The Lyapunov-Krasovskii functional combined with the input delay approach as well as the free-weighting matrix approach is employed to derive several sufficient criteria in terms of LMIs to ensure that the delayed MJFCNNs with the Wiener process is stochastic asymptotical synchronous. Restrictions (e.g., time derivative is smaller than one) are removed to obtain a proposed sampled-data controller. Finally, a numerical example is provided to demonstrate the reliability of the derived results.     

结构材料辐照损伤的分子动力学程序GPU并行化及优化

基于NIVIDIA公司的CUDA架构对结构材料辐照损伤的分子动力学程序在单个GPU上进行并行化,并对影响程序运行效率的相关因素进行分析和测试.经过一系列优化,当粒子数为两百万时,对比单CPU的执行时间,优化后的GPU程序其双精度加速比可达112倍,单精度加速比达到了三百倍,为后续扩展多GPU结构材料辐照损伤的分子动力学程序奠定基础.     

Evolution of Brains and Computers: The Roads Not Taken

When computers started to become a dominant part of technology around the 1950s, fundamental questions about reliable designs and robustness were of great relevance. Their development gave rise to the exploration of new questions, such as what made brains reliable (since neurons can die) and how computers could get inspiration from neural systems. In parallel, the first artificial neural networks came to life. Since then, the comparative view between brains and computers has been developed in new, sometimes unexpected directions. With the rise of deep learning and the development of connectomics, an evolutionary look at how both hardware and neural complexity have evolved or designed is required. In this paper, we argue that important similarities have resulted both from convergent evolution (the inevitable outcome of architectural constraints) and inspiration of hardware and software principles guided by toy pictures of neurobiology. Moreover, dissimilarities and gaps originate from the lack of major innovations that have paved the way to biological computing (including brains) that are completely absent within the artificial domain. As it occurs within synthetic biocomputation, we can also ask whether alternative minds can emerge from A.I. designs. Here, we take an evolutionary view of the problem and discuss the remarkable convergences between living and artificial designs and what are the pre-conditions to achieve artificial intelligence.     

基于自适应变异果蝇优化算法和广义回归神经网络的布里渊散射谱特征提取

针对布里渊光时域反射光纤传感系统散射谱的高精度特征提取的要求,提出了一种基于自适应变异果蝇优化算法和广义回归神经网络的布里渊散射谱特征提取算法。不仅利用了广义回归神经网络在逼近能力、学习速度、模型的泛化等方面具有的优势,而且采用搜索能力较强的自适应变异果蝇优化算法进一步增强了神经网络的学习能力,从而提高了布里渊散射谱的拟合度和频移提取的准确度。在布里渊散射谱中心频率为11.213 GHz,线宽为40～50,30～60和20～70 MHz的散射谱白噪声实验模型中,将新算法分别与基于有限元分析的Levenberg-Marquardt拟合法、粒子群优化和拉凡格式混合拟合法、最小二乘法进行预测比较,新算法获得的最大拟合频移误差为0.4 MHz,平均拟合度为0.991 2,均方根误差为0.024 1。仿真结果表明所提出的算法拟合度较好,绝对误差小。因此,将此算法用于基于布里渊光时域反射的分布式光纤传感系统,可有效提高布里渊散射谱的拟合度和频移提取的准确度。     

A Transfer Learning Approach on the Optimization of Edge Detectors for Medical Images Using Particle Swarm Optimization

Edge detection is a fundamental image analysis task, as it provides insight on the content of an image. There are weaknesses in some of the edge detectors developed until now, such as disconnected edges, the impossibility to detect branching edges, or the need for a ground truth that is not always accessible. Therefore, a specialized detector that is optimized for the image particularities can help improve edge detection performance. In this paper, we apply transfer learning to optimize cellular automata (CA) rules for edge detection using particle swarm optimization (PSO). Cellular automata provide fast computation, while rule optimization provides adaptability to the properties of the target images. We use transfer learning from synthetic to medical images because expert-annotated medical data is typically difficult to obtain. We show that our method is tunable for medical images with different properties, and we show that, for more difficult edge detection tasks, batch optimization can be used to boost the quality of the edges. Our method is suitable for the identification of structures, such as cardiac cavities on medical images, and could be used as a component of an automatic radiology decision support tool.     

Implications of Noise on Neural Correlates of Consciousness: A Computational Analysis of Stochastic Systems of Mutually Connected Processes

Random fluctuations in neuronal processes may contribute to variability in perception and increase the information capacity of neuronal networks. Various sources of random processes have been characterized in the nervous system on different levels. However, in the context of neural correlates of consciousness, the robustness of mechanisms of conscious perception against inherent noise in neural dynamical systems is poorly understood. In this paper, a stochastic model is developed to study the implications of noise on dynamical systems that mimic neural correlates of consciousness. We computed power spectral densities and spectral entropy values for dynamical systems that contain a number of mutually connected processes. Interestingly, we found that spectral entropy decreases linearly as the number of processes within the system doubles. Further, power spectral density frequencies shift to higher values as system size increases, revealing an increasing impact of negative feedback loops and regulations on the dynamics of larger systems. Overall, our stochastic modeling and analysis results reveal that large dynamical systems of mutually connected and negatively regulated processes are more robust against inherent noise than small systems.