LPNN-based approach for LASSO problem via a sequence of regularized minimizations

In compressive sampling theory, the least absolute shrinkage and selection operator (LASSO) is a representative problem. Nevertheless, the non-differentiable constraint impedes the use of Lagrange programming neural networks (LPNNs). We present in this article the -LPNN model, a novel algorithm that tackles the LASSO minimization together with the underlying theory support. First, we design a sequence of smooth constrained optimization problems, by introducing a convenient differentiable approximation to the non-differentiable -norm constraint. Next, we prove that the optimal solutions of the regularized intermediate problems converge to the optimal sparse signal for the LASSO. Then, for every regularized problem from the sequence, the -LPNN dynamic model is derived, and the asymptotic stability of its equilibrium state is established as well. Finally, numerical simulations are carried out to compare the performance of the proposed -LPNN algorithm with both the LASSO-LPNN model and a standard digital method.     

高速旋转机械用消除频率混叠现象的小波变换算法研究

高速旋转机械由于其工作转速高、控制精度高，其故障信号又属于微弱突变信号，造成常规监测方法准确率偏低，成为制约高速旋转机械故障监测进一步发展的最大障碍。本文设计了一种消除小波分解和重构过程中产生的频率混叠的算法。该算法利用傅里叶变换和傅里叶逆变换构成Mallat小波变换在分解和重构过程中所需的严格正交镜像滤波器，从而达到Mallat小波变换过程中必须具有的理想截止特性，最终去除掉多余的频率成分。利用改进前后的小波算法对实际信号进行分析，结果表明，本文设计的Mallat小波变换改进算法可消除信号分析中出现的频率混叠现象。     

基于分段加权的反向稀疏跟踪算法研究

为提高稀疏表示跟踪模型性能，提出一种分段加权的反向稀疏跟踪算法，将跟踪问题转化为在贝叶斯框架下寻找概率最高的候选对象问题，构造不同的分段权重函数来分别度量候选目标与正负模板的判别特征系数。通过池化来降低跟踪结果的不确定性干扰，选择正负模板加权系数差值最大的候选表示作为跟踪结果。实验表明，在光照变化、遮挡、快速运动、运动模糊情况下，所提出的算法可以确保跟踪结果的准确性和鲁棒性。     

飞秒激光光斑轮廓的傅里叶描述子重建及特征提取

在利用飞秒激光对单晶硅材料进行微尺度烧蚀时,衍生的等离子体发光光斑中蕴含着大量的加工信息。如何有效提取光斑图像的几何特征信息,对研究飞秒激光烧蚀加工工艺过程具有重要意义。轮廓特征是描述激光光斑最重要的特征之一,本文着重分析了等离子体光斑轮廓特征与光斑运动规律间的关系及光斑运动情况分类方法。首先,考虑到光斑微弱的特性,对光斑图像进行增强处理。其次,采用链码方法提取光斑轮廓特征,并利用傅里叶描述子低频特性,选取少量的描述子重建光斑图像轮廓特征,减少了复杂的计算量,有效提高了光斑图像分类的速度与质量。最后,运用Hu不变矩特征证明傅里叶描述子用于提取激光光斑对轮廓特征的有效性。     

Eliciting and measuring hesitance in decision-making

Hesitance is an innate human psychological phenomenon pervasive in our daily life. Although it is very complex and still not fully understood, there have been many studies related to this interesting topic. In this study, we proposed two fundamental problems about hesitance: the first is the generating problem, which is related to how to elicit hesitance information from human thinking; the second is the measuring problem, which is related to how to devise effective and reasonable methods to measure hesitance degrees from given hesitance information. With these two fundamental questions, we first discussed and analyzed several examples in real life involving hesitance, showing that it can indeed be recorded and represented. Then we roughly classified the hesitance information into two major classes: static hesitance and dynamic hesitance. Some simple and interesting methods were proposed to elicit hesitance, and more reasonable methods were proposed to measure the hesitance degree from the two classes of hesitance information. All methods that have been discussed in this study try to avoid complexity in every aspect while reserving strictness and reasonability; therefore, the present study provides suitable information for practitioners with different backgrounds.     

食醋总酸近红外快速检测方法的研究

利用傅立叶变换近红外光谱技术结合偏最小二乘法（PLS）对食醋的总酸进行快速检测分析。随机采集106个不同酸度食醋的实验数据、扫描近红外光谱图，选择最优的光谱预处理方法优化、建立模型。利用模型对15个未参与建模的食醋样品总酸的含量进行预测，外部验证模型的准确性。结果表明，食醋总酸实测值与对应采集的近红外漫反射光谱相关联建立的快速检测模型，其近红外食醋总酸模型的交叉验证决定系数（R2）为0.972 3，交叉验证均方差（RMSECV）为0.062 1。经外部验证后，该模型食醋总酸预测值和实测值的绝对偏差平均值为0.035，最大相对误差为1.383%，两者间相关性系数为0.995。该方法可以快速、高效、简便地用于食醋总酸的快速检测。     

多共振分量融合CNN的行星齿轮箱故障诊断

针对行星齿轮箱中各部件所激起的振动成分混叠、早期故障特征经常被较强的各级齿轮谐波成分以及环境噪声所湮没的问题，提出一种多共振分量融合卷积神经网络（multi-resonance component fusion based convolutional neural network，简称MRCF-CNN）的行星齿轮箱故障诊断方法。首先，对振动信号进行共振稀疏分解，得到包含齿轮谐波成分的高共振分量和可能包含轴承故障冲击成分的低共振分量；其次，构建多共振分量融合卷积神经网络，将得到的高、低共振分量和原始振动信号进行自适应的特征级融合，通过有监督的方式训练模型并进行行星齿轮箱故障诊断。对行星齿轮箱实验数据的分析结果表明，该方法能够有效分类行星齿轮箱中滚动轴承和齿轮的故障，成功对行星齿轮箱故障进行诊断，同时能够进一步增强卷积神经网络对振动信号所蕴含的故障信息的辨识能力。     

Affine hull based target representation for visual tracking

Handling appearance variations is a challenging issue in visual tracking. Existing appearance models are usually built upon a linear combination of templates. With such kind of representation, accurate visual tracking is not desirable when heavy appearance variations are in presence. Under the framework of particle filtering, we propose a novel target representation for tracking. Namely, the target candidates are represented by affine combinations of a template set, which leads to better capability in describing unseen target appearances. Additionally, in order to adapt this representation to dynamic contexts across a video sequence, a novel template update scheme is presented. Different from conventional approaches, the scheme considers both the importance of one template to a target candidate in the current frame and the recentness of the template that is kept in the template set. Comprehensive experiments show that the proposed algorithm achieves superior performances in comparison with state-of-the-art works.     

基于双向位图的CSR大规模图存储优化

大数据时代,Graph500是评测超级计算机处理数据密集型应用能力的重要工具,E级验证系统的图遍历处理能力主要受限于内存空间和访存带宽,尤其是内存空间利用率直接决定了图的测试规模和测试性能.针对天河E级验证系统小内存特征,提出了基于双向位图的大规模图数据压缩存储方法(bidirectional-bitmap based CSR,Bi-CSR),Bi-CSR在CSR矩阵压缩的基础上引入行方向位图和列方向位图协同完成稀疏矩阵压缩存储,行方向位图主要负责行方向位图的压缩存储与索引,列方向位图除了进一步压缩图存储空间,还负责为顶点遍历向量并行优化提供加速空间.Bi-CSR大幅度减少了稀疏矩阵存储空间.面向天河E级验证系统,当图输入规模为237时,Graph500的图存储空间节约效率接近70%,全系统稳定测试性能为2.131E+12TEPS,性能最大加速比超过100倍.