Accelerated Matrix Recovery via Random Projection Based on Inexact Augmented Lagrange Multiplier Method

In this paper, a unified matrix recovery model was proposed for diverse corrupted matrices. Resulting from the separable structure of the proposed model, the convex optimization problem can be solved efficiently by adopting an inexact augmented Lagrange multiplier (IALM) method. Additionally, a random projection accelerated technique (IALM+RP) was adopted to improve the success rate. From the preliminary numerical comparisons, it was indicated that for the standard robust principal component analysis (PCA) problem, IALM+RP was at least two to six times faster than IALM with an insignificant reduction in accuracy; and for the outlier pursuit (OP) problem, IALM+RP was at least 6.9 times faster, even up to 8.3 times faster when the size of matrix was 2 000×2 000.

Quasi-cyclic Random Projection Code and Hardware Implementation

Random Projection Code (RPC) is a mechanism that combines channel coding and modulation together and realizes rate adaptation in the receiving end. Random projection code’s mapping matrix has significant influences on decoding performance as well as hardware implementation complexity. To reduce hardware implementation complexity, we design a quasi-cyclic mapping matrix for RPC codes. Compared with other construction approaches, our design gets rid of data filter component, thus reducing chip area 7284.95 um2, and power consumption 331.46 uW in 0.13 um fabrication. Our simulation results show that our method does not cause any performance loss and even gets 0.2 dB to 0.5 dB gain at BER 10-4.

Tensor Train Random Projection

This work proposes a Tensor Train Random Projection(TTRP)method for dimension reduction,where pairwise distances can be approximately preserved.Our TTRP is systematically constructed through a Tensor Train(TT)representation with TT-ranks equal to one.Based on the tensor train format,this random projection method can speed up the dimension reduction procedure for high-dimensional datasets and requires fewer storage costs with little loss in accuracy,comparedwith existingmethods.We provide a theoretical analysis of the bias and the variance of TTRP,which shows that this approach is an expected isometric projectionwith bounded variance,and we show that the scaling Rademacher variable is an optimal choice for generating the corresponding TT-cores.Detailed numerical experiments with synthetic datasets and theMNIST dataset are conducted to demonstrate the efficiency of TTRP.

Face Templates Encryption Technique Based on Random Projection and Deep Learning

Cancellable biometrics is the solution for the trade-off between two concepts:Biometrics for Security and Security for Biometrics.The cancelable template is stored in the authentication system’s database rather than the original biometric data.In case of the database is compromised,it is easy for the template to be canceled and regenerated from the same biometric data.Recoverability of the cancelable template comes from the diversity of the cancelable transformation parameters(cancelable key).Therefore,the cancelable key must be secret to be used in the system authentication process as a second authentication factor in con-junction with the biometric data.The main contribution of this paper is to tackle the risks of stolen/lost/shared cancelable keys by using biometric trait(in different feature domains)as the only authentication factor,in addition to achieving good performance with high security.The standard Generative Adversarial Network(GAN)is proposed as an encryption tool that needs the cancelable key during the training phase,and the testing phase depends only on the biometric trait.Additionally,random projection transformation is employed to increase the proposed system’s security and performance.The proposed transformation system is tested using the standard ORL face database,and the experiments are done by applying different features domains.Moreover,a security analysis for the proposed transformation system is presented.

Rate Adaptation for Decoding-and-Forward Relay Channel by Random Projections Codes

This paper investigates rate adaptation schemes for decoding-and-forward (DF) relay system based on random projections codes (RPC). We consider a classic three node relay system model, where relay node performs on half-duplex mode. Then, we give out receiving diversity relay scheme and coding diversity relay scheme, and present their jointly decoding methods. Furthermore, we discuss the performance of the two schemes with different power allocation coefficients. Simulations show that our relay schemes can achieve different gain with the help of relay node. And, we should allocate power to source node to just guarantee relay node can decode successfully, and allocate remain power to relay node as far as possible. In this way, this DF relay system not only achieves diversity gain, but also achieves higher and smooth spectrum efficiency.

基于MRP的可撤销模板设计及其分析

针对基于生物特征认证系统中的存储及传输安全问题,并且考虑到基于VQ(Vector Quantization)算法声纹认证系统训练数据少,存储空间和训练时间也比较小的优点,在基于VQ算法声纹认证系统的基础上,采用MRP(Multispace Random Projection)提出一种可撤销声纹模板.通过特征点与特征点,特征点与码字之间的距离变换前后保持不变,说明平均量化误差不变,从而证明该方法满足可撤销模板的性能保持性.通过随机矩阵和不定方程的分析证明该方法满足不可逆性,即是安全的.初步实验结果的认证率达96%,说明该方法的有效性.

基于VSRP与β-GWO-SVM的变压器故障辨识方法

针对变压器故障诊断准确率低的问题,提出一种基于VSRP与β-GWO-SVM的变压器故障诊断方法。首先,采用非常稀疏随机投影(very sparse random projection,VSRP)对原始数据进行维数约减,消除变量信息之间的冗余特征;其次,用β-混沌序列优化的灰狼算法(β-chaotic map enabled grey wolf optimizer,β-GWO)动态寻优支持向量机(support vector machine,SVM)的核参数与惩罚因子,获取VSRP与β-GWO-SVM相结合的综合故障诊断模型;最后,将实际变压器故障数据输入诊断模型,并与传统灰狼优化支持向量机模型(GWO-SVM)、粒子群优化支持向量机模型(particle swarm optimization-SVM,PSO-SVM)进行对比验证,诊断精度分别为91.87%、82.13%、75.10%,结果表明该文所提方法能够有效地提升变压器故障诊断性能。

基于随机部署RP的位置指纹定位算法

随着人们对位置信息需求的不断增加,室内定位技术迎来了前所未有的研究热潮.在分析影响基于位置指纹定位算法性能主要因素的基础上,提出了基于随机部署参考点(RP)的位置指纹定位算法.该方法首先将目标环境划分成多个子区域,在每个子区域中采用随机方式部署RP,从而减少定位成本、提高灵活性;为了在保证定位性能的前提下缩短定位时间,给出了两种数据剔除方法.仿真结果表明,该方法能够获得良好的定位性能,具有实际的应用价值.

Reliability analysis of structure with random parameters based on multivariate power polynomial expansion

A new method for calculating the failure probabilityof structures with random parameters is proposed based onmultivariate power polynomial expansion, in which te uncertain quantities include material properties, structuralgeometric characteristics and static loads. The structuralresponse is first expressed as a multivariable power polynomialexpansion, of which the coefficients ae then determined by utilizing the higher-order perturbation technique and Galerkinprojection scheme. Then, the final performance function ofthe structure is determined. Due to the explicitness of theperformance function, a multifold integral of the structuralfailure probability can be calculated directly by the Monte Carlo simulation, which only requires a smal amount ofcomputation time. Two numerical examples ae presented toillustate te accuracy ad efficiency of te proposed metiod. It is shown that compaed with the widely used first-orderreliability method （ FORM） and second-order reliabilitymethod （ SORM）, te results of the proposed method are closer to that of the direct Monte Carlo metiod,and it requires much less computational time.

Early filiform needle acupuncture for poststroke depression: a meta-analysis of 17 randomized controlled clinical trials

OBJECTIVE： To evaluate the effectiveness and safety of filiform needle acupuncture for poststroke depression, and to compare acupuncture with the therapeutic efficacy of antidepressant drugs. DATA RETRIEVAL： We retrieved data from the Chinese National Knowledge Infrastructure （1979-2012）, Wanfang （1980-2012）, VIP （1989-2012）, Chinese Biomedical Literature （1975- 2012）, PubMed （1966-2012）, Ovid Lww （-2012）, and Cochrane Library （-2012） Database using the internet. SELECTION CRITERIA： Randomized controlled trials on filiform needle acupuncture versus antidepressant drugs for treatment of poststroke depression were included. Moreover, the in- cluded articles scored at least 4 points on the Jadad scale. Exclusion criteria： other acupuncture therapies as treatment group, not stroke-induced depression patients, score 〈 4 points, non-ran- domized controlled trials, or animal trials. MAIN OUTCOME MEASURES： These were the Hamilton Depression Scale scores, clinical ef- fective rate, Self-Rating Depression Scale scores, Side Effect Rating Scale scores, and incidence of adverse reaction and events. RESULTS： A total of 17 randomized controlled clinical trials were included. Meta-analysis results displayed that after 4 weeks of treatment, clinical effective rate was better in patients treated with fill- form needle acupuncture than those treated with simple antidepressant drugs [relative risk = 1.11, 95% confidence interval （C/）： 1.03-1.21, P = 0.01]. At 6 weeks, clinical effective rate was similar between filiform needle acupuncture and antidepressant drug groups. At 2 weeks after filiform needle acupuncture, Hamilton Depression Scale （17 items） scores were lower than in the antide- pressant drug group （mean difference = -2.34, 95%CI： -3.46 to -1.22, P 〈 0.000,1）. At 4 weeks, Hamilton Depression Scale （24 items） scores were similar between filiform needle acupuncture and antidepressant drug groups. Self-Rating Depression Scale scores were lower in filiform needle acupuncture group than in the antidepressant drug group. Side Effect Rating Scale was used in only two articles, and no meta-analysis was conducted. Safety evaluation of the 17 arti- cles showed that gastrointestinal tract reactions such as nausea and vomiting were very common in the antidepressant drug group. Incidence of adverse reaction and events was very low in the filiform needle acupuncture group. CONCLUSION： Early filiform needle acupuncture for poststroke depression can perfectly con- trol depression. Filiform needle acupuncture is safe and reliable. Therapeutic effects of filiform needle acupuncture were better than those of antidepressant drugs.

Statistics of Projected Motion in One Dimension of a D-Dimensional Random Walker

We are studying the motion of a random walker in generalised d-dimensional continuum with unit step length (up to 10 dimensions) and its projected one dimensional motion numerically. The motion of a random walker in lattice or continuum is well studied in statistical physics but what will be the statistics of projected one dimensional motion of higher dimensional random walker is yet to be explored. Here in this paper, by addressing this particular type of problem, it shows that the projected motion is diffusive irrespective of any dimension;however, the diffusion rate is changing inversely with dimensions. As a consequence, it can be predicted that for the one dimensional projected motion of infinite dimensional random walk, the diffusion rate will be zero. This is an interesting result, at least pedagogically, which implies that though in infinite dimensions there is diffusion, its one dimensional projection is motionless. At the end of the discussion we are able to make a good comparison between projected one dimensional motion of generalised d-dimensional random walk with unit step length and pure one dimensional random walk with random step length varying uniformly between -h to h where h is a “step length renormalizing factor”.

考虑随机返工与紧急插单的可抢占设计项目群调度

鉴于多项目并行以及因随机返工和紧急插单而导致的任务抢占特征,使设计项目群调度问题难以求解,而传统随机动态规划方法因维数灾难只能获得小规模问题的最优策略,为构造面向大规模问题的高效近似方法,提出将原随机调度问题近似分解为各个状态下的确定性调度子问题,从而获得原问题的次优策略。建立了确定性调度子问题的混合整数规划模型,并提出基于元启发式算法和优先规则的求解方法。基于标准测试集PSPLIB设计了多组计算实验,在不同调度环境下验证了所构建模型和算法的有效性。实验结果表明,元启发式算法不但在单项目平均拖期比目标下较最优规则提升了12%以上,而且具有较高的计算效率,能够满足实际决策需求。

基于随机森林优化RL的电力工程造价预测

针对现阶段电力工程造价数据预测准确度较低的问题,提出了一种基于随机森林优化RL的相关预测算法。通过分析影响电力工程造价的因素,将其按照影响程度的大小进行分级,利用随机森林算法加以筛选,并采用筛选后的影响因素特征向量作为预测模型的训练和测试数据。同时,将电力工程造价预测问题转化为输电线路规划问题,再使用强化学习优化蚁群算法的参数来构建电力工程造价预测模型。经过实验对照,综合两种栅格尺寸结果,所提方案比两种对照组算法的电子工程造价分别降低了2.97%、3.78%。

公共政策评估中的随机控制实验:实施困境与突破策略

采用随机控制实验方法来进行政策评估,已成为公共政策或项目评估的“黄金标准”。但是,随机控制实验方法在中国政策评估过程中还运用较少,评估者缺乏对其实施困境与突破策略的深入认识。本研究在深入分析随机控制实验方法及其适用范围的基础上,对随机控制实验的实施困境与突破策略进行了系统的考察与总结。研究发现,采用随机控制实验开展政策评估主要面临着实验方法本身的复杂性、实验方法引发的公共价值分歧以及随机控制实验融入中国政策场域的挑战三个层面的困境。

多源多特征集成的南水北调工程丹江库区湿地时空格局演变

丹江口水库是中国南水北调工程的关键水源区。随着城镇化发展以及大坝二次建设,库区湿地生态系统变化显著,亟需湿地生态科学监测。以丹江库区为例,依托Google Earth Engine(GEE)平台,首先采用已有土地覆盖数据集生成湿地样本集,其次整合Landsat影像、DEM等数据构建多源特征集合,并基于Jeffries-Matusita(JM)距离进行特征优选,使用随机森林(RF)算法实现了1985~2023年丹江库区湿地制图。结果表明:①本文提出的样本采集流程可有效提高样本质量,为长时序分类样本采集提供参考;②湿地分类特征优选后特征数由37个减为27个,分类总体精度变化不大,优选后的特征应用于丹江库区湿地分类的平均总体精度(OA)以及平均数量和分配分歧指数(QADI)分别为89.53%和0.0802,说明特征优选有效减少信息冗余,提高影像分类效率;③1985~2023年,丹江库区湿地面积呈波动增加趋势,从1985年的17839.85 ha扩大到2023年的28872.48 ha,面积增长38.12%。总体来说,丹江库区湿地生态系统呈现出逐步恢复和优化的良好态势。

基于随机活动工期的装配式项目鲁棒性调度优化

将装配式项目的生产阶段、运输阶段和装配阶段作为一个整体进行研究,考虑了各阶段的调度性质及相互牵制特性,提出了分种类分批次的调度策略。针对随机活动工期的装配式项目调度问题,采用鲁棒性调度方法,建立了以项目鲁棒值和成本值为优化目标的生产-运输-装配一体化建造鲁棒性调度模型。依据调度性质设计了集成精英策略的非支配性遗传算法(Non-dominated Sorting Genetic Algorithms-II,NSGA-II)并对其进行了求解。最后选取装配式项目实际案例,验证了算法的有效性并对实验结果进行了分析。研究结果表明,通过改变活动时差的不同分布状态,可在提高项目鲁棒性的同时保证项目成本最低。

核酸检测的重排算法及其数值模拟

2022年春季上海遭受到奥密克戎变异株的侵袭,核酸检测是筛查新冠阳性病例的重要手段之一。目前社区常用核酸检测方式是10人/20人混合检测,重点人群单人单管复测。考虑到检测数量的庞大,二次复核易对非阳性人员造成恐慌,仿照CT成像原理,设计一种核酸检测重排方法,每人检测后将检测液分别随机置于多组不同试管中,根据不同组试管中的检测结果(投影数据),利用反投影方法,确定进入二次检测人员。计算机仿真结果表明,上述所设计的重排算法能够实现减少总检测数量和提高进入二次检测人员精度的目的。因此以上重排方法有望进一步应用到智能化的核酸采样机器人上,达到高效检测的目的。

基于“扶放有度”模式的“随机过程”项目式教学方法探索

该文利用“扶放有度”教学模式,结合项目式教学方法,对“随机过程”课程的教学方法进行了探索。以“随机过程”课程中的“平稳时间序列的预报”为例,将教学分为教师示证、教师辅导、同伴协作和独立表现四个阶段,在每个阶段对教学任务进行了精心设计,并且结合项目式教学方法在每个阶段增加了教学任务项目化环节,不但培养了学生的自主学习能力,还补充了学生的实践训练,实现了以学生为中心的教育理念,对于提高学生独立开展科学研究的能力具有重要作用。最后,该文还列出了几种提高研究生教学效果的方法。

基于可见光谱特征波长提取和分类算法的柑橘黄龙病快检研究

柑橘黄龙病(HLB)是由亚洲韧皮杆菌引起的一种严重病害,目前无法根治。其防控具有重要意义和经济价值。当前利用健康和患病叶片的光谱差异对其进行诊断显示了良好的应用前景。因患病叶片在叶绿素反射区及O—H伸缩振动区的可见光谱与健康叶片存在显著差异,而可见光谱检测在采集和数据处理方面具有成本低、简便的优势,研究可见光谱的黄龙病快速检测方法具有可行性和重要意义。为了减少光谱数据冗余和计算量,实现精准的黄龙病的早期鉴别以及降低黄龙病相似病症的误诊率,采集了黄龙病患病地区共160个叶片样本。经qPCR测定分别将其分类标定为健康、轻度疾病、重度疾病和缺镁症四类。根据叶片样本在可见光波段450~800 nm的反射光谱特征,通过S-G平滑以及降采样等预处理光谱数据后,为了优选出尽可能囊括光谱特征信息的特征波长,分别使用遗传算法(GA)、连续投影算法(SPA)以及竞争自适应重加权采样法(CARS)对采集到的可见光谱数据进行特征波长提取和降维优选出特征波长,进一步降低模型复杂度,提高预测精度。综合泛化能力及检测速度的考量,在定性判别分析模型的选择中采用训练速度快,分析准确率高的最小二乘支持向量机(LS-SVM)以及随机森林(RF)对两种变量筛选算法降维后的数据进行分类判别。通过对不同的模型的验证优选,筛选出最佳的快检方案。对比发现,在建立的模型中,SPA-RF模型与其他模型比较,对于训练集和测试集的判别准确率分别达到了100%和97.5%。结果表明,连续投影算法以及随机森林的组合分类模型可以很好地实现黄龙病早期的病理鉴别,同时也能够很好地识别出黄龙病病叶与其他相似病症的差异,为柑橘黄龙病快速检测及防治提供了一种方法依据。

自适应秩约束逆矩阵近似分解及其在语音增强中的应用

针对低秩约束和稀疏矩阵分解(constrained low-rank and sparse matrix decomposition, CLSMD)方法中硬阈值可能导致降噪后的语音信号分量丢失或出现孤立噪声问题,提出了一种自适应秩约束逆矩阵近似(adaptive rank constrained inverse matrix approximation, ARCIMA)分解方法。该方法首先采用能量阈值法初步估计低秩矩阵秩值,然后从语音信号子空间矩阵的结构特性出发,采用修正双边随机投影(modified bilateral random projections, MBRP)方法求解代表纯净语音信号的低秩矩阵,降低使用SVD方法的计算量,并通过Tikhonov正则化优化方法改善迭代求解过程中解的病态性。实验结果表明,该方法相比经典方法在多种噪声环境下取得了更好的PESQ得分,并且增强语音的时域波形也更接近原始信号的波形。该方法去噪性能在低信噪比噪声条件下具有优势。