Research on exponential regularization approach for nonnegative Tucker3 decomposition

Methods of nonnegative tensor factorization (NTF), such as NTF1, NTF2, etc., are extension of nonnegative matrix factorization (NMF) for multi-way data analysis. As an existing NTF method, nonnegative Tucker3 decomposition (NTD) is researched for three-way decomposition in this paper. Firstly, an approach utilizing matrix exponentials built on Tikhonov-type regularization to enforce sparseness is proposed to extract image features instead of exclusively using Tucker tensor decomposition. Meanwhile, updating algorithms, derived from updating rules of NMF, are allowed to efficiently implement updating of mode matrices and core tensors alternatively for accuracy. Then, experimental cases of alternating least squares (ALS) and conjugate nonnegative constraints, called nonnegative alternating least squares (NALS), are studied to remedy data overfitting in computing procedures. Lastly, the proposed method exhibits more advantageous results than other algorithms of Tucker3 for feature extraction, thanks to computer simulations performed in the context of data analysis.     

An Overview of Infrared Spectroscopy Based on Continuous Wavelet Transform Combined with Machine Learning Algorithms: Application to Chinese Medicines,Plant Classification,and Cancer Diagnosis

Abstract

Infrared spectroscopy has been a workhorse technique for materials analysis and can result in positively identifying many different types of material. In recent years there have been reports using wavelet analysis and machine learning algorithms to extract features of Fourier transform infrared spectrometry (FTIR). The machine learning algorithms contain back-propagation neural network (BPNN), radial basis function neural network (RBFNN), and support vector machine (SVM). This article reviews the important advances in FTIR analysis employing a continuous wavelet transform (CWT) and machine learning algorithms, especially in the applications of the method for Chinese medicine identification, plant classification, and cancer diagnosis.     

EAST装置上W波段多道极向相关反射仪的研制

在EAST装置上安装了X模极化W波段多道相关反射仪,用于测量等离子体芯部密度涨落。该诊断利用低损耗(<3dB)多工器将4个不同频率(79.2GHz,85.2GHz,91.8GHz和96GHz)的微波耦合在一起,通过同一个天线发射。反射波由两个极向分离(～5cm)的天线接收,通过下变频技术实现外差测量。通过对两个极向天线接收的信号进行相关分析,获得芯部湍流垂直速度。对2018年低约束模式(L模)放电进行分析发现,在电子回旋共振加热(ECRH)等离子体中,芯部湍流垂直速度在电子逆磁漂移方向。而在注入同向中性束(co-NBI)后,芯部湍流垂直速度变为离子逆磁漂移方向。     

NO分子宏观气体热力学性质的理论研究

采用量子统计系综理论,研究了基态NO分子宏观气体摩尔熵、摩尔内能、摩尔热容等热力学性质.首先应用课题组前期建立的变分代数法(variational algebraic method, VAM)计算获得了基态NO分子的完全振动能级,得到的VAM振动能级作为振动部分,结合欧拉-麦克劳林渐进展开公式的转动贡献,应用于经典的热力学与统计物理公式中,从而计算得到了1000-5000 K温度范围内NO宏观气体的摩尔内能、摩尔熵和摩尔热容.将不同方法计算得到的摩尔热容结果分别与实验值进行比较,结果表明基于VAM完全振动能级获得的结果优于其他方法获得的理论结果.振动部分采用谐振子模型对无限能级求和计算热力学性质的方法有一定的局限性,应当使用有限的完全振动能级进行统计求和.     

排名聚合算法在少量长列表聚合中的性能比较分析

排名聚合将多个排名列表聚合成一个综合排名列表,可应用于推荐系统、链路预测、元搜索、提案评选等.当前已有工作从不同角度对不同排名聚合算法进行了综述、比较,但存在算法种类较少、数据统计特性不清晰、评价指标不够合理等局限性.不同排名聚合算法在提出时均声称优于已有算法,但是用于比较的方法不同,测试的数据不同,应用的场景不同,因此何种算法最能适应某一任务在很多情况下仍不甚清楚.本文基于Mallows模型,提出一套生成统计特性可控的不同类型的排名列表的算法,使用一个可应用于不同类型排名列表的通用评价指标,介绍9种排名聚合算法以及它们在聚合少量长列表时的表现.结果发现启发式方法虽然简单,但是在排名列表相似度较高、列表相对简单的情况下,能够接近甚至超过一些优化类方法的结果;列表中平局数量的增长会降低聚合排名的一致性并增加波动;列表数量的增加对聚合效果的影响呈现非单调性.整体而言,基于距离优化的分支定界方法 (FAST)优于其他各类算法,在不同类型的排名列表中表现非常稳定,能够很好地完成少量长列表的排名聚合.     

交错跃迁Hofstadter梯子的量子流相

为玻色Hofstadter梯子模型引入交错跃迁,来扩展模型支持的量子流相.基于精确对角化和密度矩阵重整化群计算发现,无相互作用时,系统中包含横流相、涡旋相和纵流相;横流相来自均匀跃迁时Hofstadter梯子模型的Meissner相,纵流相是交错跃迁时才可见的流相.强相互作用极限下系统的超流区也包含横流相、纵流相和涡旋相,但存在更多的相变级数;超流区的横流相、纵流相之间存在相变但Mott区的不存在,把Mott区的"横、纵流相"称为Mott-均匀相,在Mott区只存在均匀相和涡旋相.跃迁的交错会压缩涡旋相存在的区域,使Mott区最终只剩下均匀相;跃迁的交错不仅能驱动Mott-超流相变,还使磁通的改变也能够驱动系统的Mott-超流相变.对这一系统的研究丰富了磁通系统中的量子流相,同时为研究拓扑流特性提供了模型支持.     

基于Retinex理论与概率非局部均值的红外图像增强方法

针对传统红外图像增强算法中细节模糊及过度增强的问题,提出了一种基于Retinex理论与概率非局部均值相结合的红外图像增强方法.首先通过单尺度Retinex方法调整图像中过暗与过亮部分的灰度级;然后利用概率非局部均值对图像进行分解处理得到基本层与细节层,对基本层采用直方图均衡化拉伸对比度,对细节层采用非线性函数进行增强;最后,将不同层次的结果融合得到对比度与细节增强的红外图像.用该方法对多组不同场景的红外图像进行仿真实验,并将其与多种增强方法进行主、客观对比分析,结果表明所提方法在红外图像的细节及对比度增强方面都获得了更好的效果.     

GaInP/GaAs/Ge三结太阳电池不同能量质子辐照损伤模拟

以GaInP/GaAs/Ge三结太阳电池为研究对象,开展了能量为0.7, 1, 3, 5, 10 MeV的质子辐照损伤模拟研究,建立了三结太阳电池结构模型和不同能量质子辐照模型,获得了不同质子辐照条件下的I-V曲线,光谱响应曲线,结合已有实验结果验证了本文模拟结果,分析了三结太阳电池短路电流、开路电压、最大功率、光谱响应随质子能量的变化规律,利用不同辐照条件下三结太阳电池最大输出功率退化结果,拟合得到了三结太阳电池最大输出功率随位移损伤剂量的退化曲线.研究结果表明,质子辐照会在三结太阳电池中引入位移损伤缺陷,使得少数载流子扩散长度退化幅度随质子能量的减小而增大,从而导致三结太阳电池相关电学参数的退化随质子能量的减小而增大.相同辐照条件下,中电池光谱响应退化幅度远大于顶电池光谱响应退化幅度,中电池抗辐照性能较差,同时中电池长波范围内光谱响应的退化幅度比短波范围更大,表明中电池相关电学参数的退化主要来源于基区损伤.     

Surface passivation in n-type silicon and its application in silicon drift detector

Based on the surface passivation of n-type silicon in a silicon drift detector(SDD), we propose a new passivation structure of SiO2/Al2O3/SiO2 passivation stacks. Since the SiO2 formed by the nitric-acid-oxidation-of-silicon(NAOS)method has good compactness and simple process, the first layer film is formed by the NAOS method. The Al2O3 film is also introduced into the passivation stacks owing to exceptional advantages such as good interface characteristic and simple process. In addition, for requirements of thickness and deposition temperature, the third layer of the SiO2 film is deposited by plasma enhanced chemical vapor deposition(PECVD). The deposition of the SiO2 film by PECVD is a low-temperature process and has a high deposition rate, which causes little damage to the device and makes the SiO2 film very suitable for serving as the third passivation layer. The passivation approach of stacks can saturate dangling bonds at the interface between stacks and the silicon substrate, and provide positive charge to optimize the field passivation of the n-type substrate.The passivation method ultimately achieves a good combination of chemical and field passivations. Experimental results show that with the passivation structure of SiO2/Al2O3/SiO2, the final minority carrier lifetime reaches 5223 μs at injection of 5×10¹⁵ cm^-3. When it is applied to the passivation of SDD, the leakage current is reduced to the order of nA.     

Multicaloric and coupled-caloric effects

The multicaloric effect refers to the thermal response of a solid material driven by simultaneous or sequential application of more than one type of external field.For practical applications,the multicaloric effect is a potentially interesting strategy to improve the efficiency of refrigeration devices.Here,the state of the art in multi-field driven multicaloric effect is reviewed.The phenomenology and fundamental thermodynamics of the multicaloric effect are well established.A number of theoretical and experimental research approaches are covered.At present,the theoretical understanding of the multicaloric effect is thorough.However,due to the limitation of the current experimental technology,the experimental approach is still in progress.All these researches indicated that the thermal response and effective reversibility of multiferroic materials can be improved through multicaloric cycles to overcome the inherent limitations of the physical mechanisms behind single-field-induced caloric effects.Finally,the viewpoint of further developments is presented.