基于自动编码器的深度伪造图像检测方法

基于深度学习的图像伪造方法生成的图像肉眼难辨,一旦该技术被滥用于制作虚假图像和视频,可能会对国家政治、经济、文化造成严重的负面影响,也可能会对社会生活和个人隐私构成威胁。针对上述问题,提出了一种基于自动编码器的深度伪造Deepfake图像检测方法。首先,借助高斯滤波对图像进行预处理,提取高频信息作为模型输入;然后,利用自动编码器对图像进行特征提取,并在编码器中添加注意力机制模块以获取更好的分类效果;最后,通过消融实验证明,采用所提的预处理方法和添加注意力机制模块有助于伪造图像检测。实验结果表明,与ResNet50、Xception以及InceptionV3相比,所提方法在数据集样本量较小且包含的场景丰富时,可以有效检测多种生成方法所伪造的图像,其平均准确率可达97.10%,明显优于对比方法,且其泛化性能也明显优于对比方法。     

基于朴素贝叶斯算法的中文评论分类

针对中文评论分类问题,采用朴素贝叶斯算法进行深入研究.首先,根据中文评论分类的需求设计了朴素贝叶斯分类器;然后,使用WEKA以不同特征提取方式对其功能性进行了对比分析.通过一系列的实验数据的横向对比表明,在朴素贝叶斯分类器下采用集成特征选取时文本分类的准确率最佳,准确率达97.65％,验证了朴素贝叶斯分类器在处理中文评论分类问题的可应用性.     

Effects of foundation damage and water-level change on vibration modal parameters of gravity-type caisson structure

In this paper,the effects of foundation damage and water-level change on vibration characteristics of gravity-type caisson structure are examined by analyzing modal parameters extracted from output-only information.To achieve the objective,the following approaches are implemented.Firstly,vibration response analysis methods are selected to estimate power spectral density and modal parameters such as natural frequency,damping ratio and mode shape of a lab-scale caisson structural system.Secondly,vibration tests on the lab-scale caisson system are performed under a series of test scenarios which include three water-level changes and three damage levels.Thirdly,experimental modal parameters corresponding to the damaging cases as well as the water level cases are extracted by frequency domain decomposition method and stochastic subspace identification method.Finally,the effects of the water-level variation and foundation damage on the extracted modal parameters are examined to assess the feasibility of the vibration-based damage detection in gravity-type caisson structures under water-level uncertainty.     

Variable step-size affine projection algorithm based on global speech absence probability for adaptive feedbackcancellation简

A novel approach is proposed for improving adaptive feedback cancellation using a variable step-size affine projection algorithm（VSS-APA） based on global speech absence probability（GSAP）.The variable step-size of the proposed VSS-APA is adjusted according to the GSAP of the current frame.The weight vector of the adaptive filter is updated by the probability of the speech absence.The performance measure of acoustic feedback cancellation is evaluated using normalized misalignment.Experimental results demonstrate that the proposed approach has better performance than the normalized least mean square（NLMS） and the constant step-size affine projection algorithms.     

AN INVESTIGATION ON THE EFFECT OF MANGANESE SPECIFICATION IN BASIC OPEN HEARTH PROCESS

本文结合生产研究了鞍钢碱性平爐冶炼重轨和无缝等碳素镇静钢的锰制度.研究结果指出:就上述鋼种而言,熔炼过程中的锰制度对钢中氧、硫含量以及轧制的成品質量並没有实际影响,换句话说,并没有必要规定熔池含锰量高於一定水平(>0.15—0.2%Mn).相反地,熔池含锰较高或在熔炼过程中加入锰铁均使降碳速度趋於缓慢,从而延长了熔炼时间.因此,只要高炉铁水中含硫量能够合乎规定,用低锰生铁作原料对炼钢並无害处.通过高爐生产数据分析,说明在鞍钢的操作条件下完全有可能生产含锰低而含硫合乎上述规定的铁水,应该认为用低锰生鉄冶炼重轨和无缝类型的优質鋼是切实可行的.     

ON THE STICKING OF THIN STEEL SHEETS DURING PACK ROLLING

本文叙述关于热叠轧薄板粘结的研究结果。比较全面和深入的研究说明,薄板粘结与许多因素有关,因此必须较全面地控制轨料、工具和加热情况以及轧制条件,才能较有效地减少粘结。 实验结果指出,温度为影响粘结的主要因素之一,粘结随温度升高而增加。粘结随压下率的增加亦有增加。轧制速度适当地提高可以减少粘结。辊面光洁度对粘结亦有一定的影响,经常保持辊面的光洁度可以减少粘结。薄板轨后的冷却速度和方式,对粘结也有很大影响,故板堆不宜过高,且应使之加速冷却。 氧化铁皮亦为影响粘结的主要因素之一,毫无氧化铁皮的表面可以在轧制过程中结为牢固的整体。加热和轧制均能改变薄板内磷的分布’而使薄板表面层的磷含量增加,因此少量的磷对减少粘结可以起显著的作用。 粘结区的显微结构表现有五种不同的特征,足以说明造成粘结的多种原因。     

Pechini法制备的CeO2纳米晶的晶化动力学

通过Pechini法制备CeO2粉末.XRD分析表明,该粉末在300℃下煅烧1 h后成为纳米晶体.用DT-40热分析仪研究其晶化过程,在热分析过程中,CeO2的晶化过程符合Kissinger方程,计算得到其相变激活能为86.42 kJ/mol,且其晶化过程所需能量比固相法的晶化过程所需能量显著降低.根据Qzawa方程得到,在250℃下Avrami指数n为1.5,频率因子为1.75×108.并由此推断该反应机理的形核过程为长程扩散控制型生长(初期阶段),从小尺寸开始的各种形状的生长形核率随时间下降.     

A large-scale multi-agent system for navy personnel distribution

In the US Navy, at the end of each sailor's tour of duty, he or she is assigned to a new job. The Navy employs some 280 people, called detailers, full time to effect these new assignments. The intelligent distribution agent (IDA) prototype was designed and built to automate, in a cognitively plausible manner, the job of the human detailers. That model is being redesigned to function as a multi-agent system. This is not a trivial matter due to the fact that there would need to be approximately 350 ‐ 000 individual agents. There are also many issues relating to how the agents interact and how all entities involved, including humans, exercise their autonomy. This paper describes both the IDA prototype and the multi-agent IDA system being created from it. We shall also discuss several of the major issues regarding the design, interaction and autonomy of the various agents involved.     

ON THE IMPROVEMENT OF THE QUALITY OF BALL-BEARING STEEL L THE EFFECT OF SMELTING AND CASTING PRACTICE ON NON-METALLIC INCLUSIONS IN BALL-BEARING STEEL

本文是提高滚珠轴承钢质量研究工作的第一部分。研究了碱性电弧炉内不同脱氧制度和浇铸条件对于钢中非金属夹杂物类型和数量的影响设及冶炼和浇铸过程中夹杂物的变化。 试验结果指出,采用下列操作制度可以显著改善滚珠钢中非金属夹杂物: 扒氧化渣后用硅锰合金预脱氧,然后用电石直接造电石渣,保持40分钟以上,还原后期用硅铁粉或电石继续进行扩散脱氧,出钢前炉渣必须变白,采用先出渣后出钢的办法,使渣钢在盛钢桶内得到很好的混合,进一步改善钢中非金属夹杂物。加铝量应在0.5公斤/吨以上。 结果证实,使用焦油处理过的优质盛钢桶内衬砝和汤道砝可以减少外来夹杂物。钢中氮化钛的来源是铁合金中含钛,所以在冶炼优质滚珠钢时应该选用不含钛的铁合金。     

Compensation of hysteresis in piezoelectric actuator with iterative learning control

This paper presents the application of iterative learning control (ILC) to compensate hysteresis in a piezoelectric actuator. The proposed controller is a hybrid of proportional-integral-differential (PID) control, whose main function is for trajectory tracking, and a chatter-based ILC, whose main function is for hysteresis compensation. Stability analysis of the proposed ILC is presented, with the PID included in the dynamic of the piezoelectric actuator. The performance of the proposed controller is analysed through simulation and verified with experiment with a piezoelectric actuator.