小样本轴承故障诊断研究综述

随着数据时代的来临,基于数据驱动的轴承故障诊断方法表现出了优越的性能,但是此类方法依赖大量标记数据,而在实际生产过程中很难收集到大量的数据,因此小样本的轴承故障诊断具有很高的研究价值。对小样本条件下的轴承故障诊断方法进行了回顾,并将其分为两类：基于数据的方法和基于模型的方法。其中基于数据的方法是从数据角度对原始样本进行扩充;基于模型的方法是指利用模型优化特征提取或者提高分类精度等。总结了当前小样本条件下故障诊断方法的不足,并展望了小样本轴承故障诊断的未来。     

物理域中针对人脸识别系统的对抗样本攻击方法

对抗样本攻击揭示了人脸识别系统可能存在不安全性和被攻击的方式。现有针对人脸识别系统的对抗样本攻击大多在数字域进行,然而从最近文献检索的结果来看,越来越多的研究开始关注如何能把带有对抗扰动的实物添加到人脸及其周边区域上,如眼镜、贴纸、帽子等,以实现物理域的对抗攻击。这类新型的对抗样本攻击能够轻易突破市面上现有绝大部分人脸活体检测方法的拦截,直接影响人脸识别系统的结果。尽管已有不少文献提出数字域的对抗攻击方法,但在物理域中复现对抗样本的生成并不容易且成本高昂。本文提出一种可从数字域方便地推广到物理域的对抗样本生成方法,通过在原始人脸样本中添加特定形状的对抗扰动来攻击人脸识别系统,达到误导或扮演攻击的目的。主要贡献包括:利用人脸关键点根据脸型构建特定形状掩膜来生成对抗扰动;设计对抗损失函数,通过训练生成器实现在数字域的对抗样本生成;设计打印分数损失函数,减小打印色差,在物理域复现对抗样本的生成,并通过模拟眼镜佩戴、真实场景光照变化等方式增强样本,改善质量。实验结果表明,所生成的对抗样本不仅能在数字域以高成功率攻破典型人脸识别系统VGGFace10,且可方便、大量地在物理域复现。本文方法揭示了人脸识别系统的潜在安全风险,为设计人脸识别系统的防御体系提供了很好的帮助。     

两批次番茄炒蛋蛋白质近红外光谱检测模型的维护方法

为解决光谱数据差异导致模型不稳定的问题,研究了不同批次中式菜肴营养素含量预测模型的传递方法。以间隔3个月制作的番茄炒蛋样本为例,采集光谱数据并利用理化方法测定蛋白质含量(每批次120个样本);选择预测效果较好的第二批次模型作为主模型,将分段直接标准化(Piecewise Direct Standardization, PDS)算法、模型更新(Model Updating, MP)和斜率/截距(Slope/Bias, S/B)修正法联合(PDS-MP-S/B)用于菜肴类模型传递,分析不同PDS窗口数和标准集数对预测结果的影响。当PDS窗口数为11且标准集数为100时,PDS-MP-S/B算法对蛋白质含量的预测结果明显优于无模型传递和单独使用3种算法时,预测模型的预测集决定系数R2(Pred)为0.9628,相对预测偏差(Relative Prediction Deviation, RPD)为5.6731,预测均方根误差(Root Mean Square Error of Prediction, RMSEP)为0.3157。从光谱、模型、结果三个方面实现了模型传递,提高了模型的通用性,减少了建模成本,为中式菜肴的快检提供了理论支持。     

Gold immunochromatographic sensor for the rapid detection of twenty-six sulfonamides in foods

A gold immunochromatographic sensor (GICS) was developed for the rapid detection of 26 sulfonamides in honey samples.The sensor was based on a group-specific monoclonal antibody (mAb) that can recognize all 26 sulfonamides.Three haptens (hapten 1 with a thiazole ring,hapten 2 with a benzene ring,and hapten 3 with a straight carbon chain) were used for antigen preparation.With hybridoma technology,a group-specific mAb was screened with a 50％ maximal inhibitory concentration (IC50) against sulfathizole (STZ) and the other 25 analogues ranging from 0.08 to 90.18 ng/mL.Mono-dispersed gold nanoparticles were conjugated with the mAb to develop the lateral immunochromatographic strip.A labeled antibody concentration of 0.1 μg/mL and a coating antigen concentration of 0.2 μg/mL in the test line were chosen for strip preparation.Under optimized conditions,the visual limits of detection (vLOD) for the concentrations of STZ,sulfamethoxazole,sulfamethizole,sulfadiazine,sulfamerazine,sulfadimethoxine,sulfamonomethoxine,sulfameter,sulfamethoxypyridazine,and sulfachloropyridazine were 5,0.25,0.25,10,5,10,25,2.5,5,0.25,and 10 μg/kg,respectively.Scanner analysis in honey samples revealed good performance for detection of the 26 sulfonamides.Commercial honey samples were tested with the sensor and positive results were confirmed with high-performance liquid chromatography.The proposed strip sensor provides a convenient method for the rapid and reliable determination of sulfonamides pollutants in honey samples.     

聚碳酸酯基聚氨酯/芳纶纤维复合材料的制备与性能

以聚碳酸酯二元醇为软段,利用预聚体法制备了系列聚碳酸酯基聚氨酯/芳纶浆粕纤维(Aramid pulp)复合材料,研究了芳纶浆粕纤维对聚碳酸酯基聚氨酯弹性体(PUE)力学性能的影响。采用扫描电镜、X射线衍射、差示扫描量热分析、热重分析、红外光谱及电子拉伸机等测试仪器对聚氨酯复合材料的结构与性能进行了表征和分析。结果表明,芳纶浆粕纤维能有效实现对聚氨酯弹性体的增强作用,其中对撕裂强度的增强尤为突出。当芳纶浆粕纤维用量为0.5%时,复合材料的撕裂强度由106 k N/m提高到120 k N/m,表现出最好的力学性能;芳纶浆粕纤维在聚氨酯中分散均匀,提高了复合材料的热稳定性。     

Semisupervised dimensionality reduction for hyperspectral images based on the combination of semisupervised learning and metric learning

This paper presents a semisupervised dimensionality reduction (DR) method based on the combination of semisupervised learning (SSL) and metric learning (ML) (CSSLML-DR) in order to overcome some existing limitations in HSIs analysis. Specifically, CSSML focuses on the difficulties of high dimensionality of hyperspectral images (HSIs) data, the insufficient number of labelled samples and inappropriate distance metric. CSSLML aims to learn a local metrics under which the similar samples are pushed as close as possible, and simultaneously, the different samples are pulled away as far as possible. CSSLML constructs two local-reweighted dynamic graphs in an iterative two-steps approach: L-step and V-step. In L-step, the local between-class and within-class graphs are updated. In V-step, the transformation matrix and the reduced space are updated. The algorithm is repeated until a stopping criterion is satisfied. Experimental results on two well-known hyperspectral image data sets demonstrate the superiority of CSSLML algorithm compared to some traditional DR methods.     

一种纸浆模塑蜂窝板材缓冲性能的研究

目的 为了解决因技术水平制约,纸蜂窝材料芯纸构型单一、单层厚度受限等瓶颈问题,开发一种通用型缓冲结构板材。方法 以纸浆模塑为原材料,以六棱柱结构为例,设计并加工出一种通过正反插方式组合而成的蜂窝板材,并采用有限元方法进行仿真,得到关键结构参数和相关最优参数组合。结果 正反插结构的纸浆模塑板材具有较好的耦合效应,承载能力相较于单层板材大幅提高;结合极限载荷、比吸能等评价指标对板材性能进行定量分析,初始压缩载荷最大可达到73.7kN,满足托盘等重型包装器具的承载需求。通过极差和方差分析可知,各结构参数对板材承载性能和缓冲性能的影响程度,研究表明单元壁厚、中截面边长对该板材承载性能和缓冲性能影响显著。结论 正反插结构解决了现有纸蜂窝材料高度、厚度受限,构型单一等瓶颈问题,提高了纸浆模塑蜂窝板材的承载性能和缓冲性能,可以满足重载要求。与之相关的其他正多边形锥形薄壁管单元亦可以使用同样的正反插设计理念和研究方法。     

New aspects in the microstructure of YBCO melt-textured high-Tc superconductors

The microstructure and phase constitution of bulk textured YBCO samples prepared by the modified MTG method in the longitudinal temperature gradient have been investigated. We found the existence of an unusual phase between stacked platelets. Microanalysis shows that this phase may be characterized as Y-143. This phase forms ellipsoid-like particles with a typical size about 0.05–2 which can play the role of effective pinning centers. The reason for the phase formation is discussed. We assume that this phase exists commonly in melt-textured samples prepared in the temperature field gradient.     

2010年我国恶意代码新特点的研究

文章以公安部2010年计算机病毒疫情调查活动的数据作为基础,以国家计算机病毒应急处理中心国家级恶意代码样本库作为支撑,结合信息网络安全通报机制成员单位汇总的情况和各反病毒研究机构发布的安全报告,阐述了2010年我国恶意代码发展的总体情况,分析了2010年我国恶意代码出现的新特点,探索了今后恶意代码的发展趋势并提出对策建议。     

A distributed approach to enabling privacy-preserving model-based classifier training

This paper proposes a novel approach for privacy-preserving distributed model-based classifier training. Our approach is an important step towards supporting customizable privacy modeling and protection. It consists of three major steps. First, each data site independently learns a weak concept model (i.e., local classifier) for a given data pattern or concept by using its own training samples. An adaptive EM algorithm is proposed to select the model structure and estimate the model parameters simultaneously. The second step deals with combined classifier training by integrating the weak concept models that are shared from multiple data sites. To reduce the data transmission costs and the potential privacy breaches, only the weak concept models are sent to the central site and synthetic samples are directly generated from these shared weak concept models at the central site. Both the shared weak concept models and the synthetic samples are then incorporated to learn a reliable and complete global concept model. A computational approach is developed to automatically achieve a good trade off between the privacy disclosure risk, the sharing benefit and the data utility. The third step deals with validating the combined classifier by distributing the global concept model to all these data sites in the collaboration network while at the same time limiting the potential privacy breaches. Our approach has been validated through extensive experiments carried out on four UCI machine learning data sets and two image data sets.