基于注意力机制的多尺度仪表检测

为实现复杂场景下多尺度仪表检测,提出了一种基于注意力机制的视频多尺度仪表检测算法。首先,利用基于空间注意力机制的特征提取网络,建模特征的长距离依赖,增强特征的表达能力;其次,提出了一种自适应特征选择模块（Adaptive Feature Selection Module, AFSM）,对不同阶段的特征图进行权重调整,增强网络对多尺度目标的检测能力。在自建的仪表数据集上进行了实验。实验结果表明,相比较原来的Faster RCNN方法,所提出方法的检测精度提高了7.6%;与对比方法相比,检测精度也能达到95.4%。在对实际仪表监测视频的测试中,检测结果以及速度能够满足实际需要。所提方法通过改进特征提取网络和特征选择操作,增强了特征表达能力,有效降低了虚警,提升了网络对多尺度目标的检测性能。     

多段连续螺纹平滑过渡的插补方法

以线性加减速为例研究了螺纹过渡起始位置的确定方法;根据螺纹切削的基本原理,分析了通过重新计算加速度或螺距增量来确保过渡时间和过渡长度固定不变的方法;在此基础上,提出了一种多段连续螺纹平滑过渡的插补方法.仿真与加工实验表明,利用文中方法进行不同螺距或不同锥度多段连续螺纹切削时,加工轨迹没有明显的凸起和变化,可实现两段螺纹...     

硅锗异质结构中等离子-声子耦合模式

利用自洽场理论考虑半导体硅锗量子阱系统中电子-电子和电子-声子相互作用,得到由这些交互作用引起的介电函数.从而得到硅锗异质结中的等离子与声子的耦合模式.     

3个分布区蒜头果叶的解剖特征及其生态适应性

对广西境内的龙虎山、巴马、乐业三个蒜头果（Malania oleifera）自然分布区的蒜头果叶片进行解剖特征研究,探讨不同分布区蒜头果的叶片解剖特征及其表现出的生态适应性。结果表明,3个分布区蒜头果叶片在解剖结构上差异明显,但均表现出早生植物的特征。3个分布区中,巴马样地蒜头果表现出气孔密度较大、气孔较小而数目较多,角质层厚度较大,导管数目较多等特征,说明其具有更明显的生态适应性。     

留学、建设与展现:近代留学生群体的自我构建

洋务运动时期,洋务派拉开留学运动的帷幕,随后留学生队伍不断壮大。到了民国末年,留学生群体基本完成了近代化转型,这是留学生自我构建的结果。除了通过留学这一基本途径外,留学生群体还通过参与社会建设、搭建发展平台来实现自我构建。     

偏心搅拌槽内高浓度浆液颗粒的悬浮特性

以直径为0.34 m的无挡板平底圆筒形搅拌槽为研究对象,对偏心搅拌槽内高浓度浆液中颗粒的悬浮特性进行了数值研究,分析了45°的4斜叶开启涡轮式搅拌桨(PBT)和3窄叶整体板式螺旋桨(ZHX)2种桨型在不同偏心率和转速时,搅拌槽内的流型分布、颗粒体积分数分布、完全离底临界悬浮转速以及功率消耗,并与试验结果进行了对比.结果表明:对于高浓度浆液,偏心搅拌打破了中心搅拌时流场结构的对称性,提高了流体的轴向循环能力,颗粒悬浮效果优于中心搅拌;固体颗粒的悬浮效果与搅拌桨在槽内的偏心位置有关,当偏心率E=0.4时颗粒悬浮效果最佳,但偏心搅拌会增大颗粒的完全离底临界悬浮转速和设备的功率消耗,不利于节能降耗.     

Opposite effects of fear conditioning and extinction on dendritic spine remodelling

It is generally believed that fear extinction is a form of new learning that inhibits rather than erases previously acquired fear memories. Although this view has gained much support from behavioural and electrophysiological studies, the hypothesis that extinction causes the partial erasure of fear memories remains viable. Using transcranial two-photon microscopy, we investigated how neural circuits are modified by fear learning and extinction by examining the formation and elimination of postsynaptic dendritic spines of layer-V pyramidal neurons in the mouse frontal association cortex. Here we show that fear conditioning by pairing an auditory cue with a footshock increases the rate of spine elimination. By contrast, fear extinction by repeated presentation of the same auditory cue without a footshock increases the rate of spine formation. The degrees of spine remodelling induced by fear conditioning and extinction strongly correlate with the expression and extinction of conditioned fear responses, respectively. Notably, spine elimination and formation induced by fear conditioning and extinction occur on the same dendritic branches in a cue- and location-specific manner: cue-specific extinction causes formation of dendritic spines within a distance of two micrometres from spines that were eliminated after fear conditioning. Furthermore, reconditioning preferentially induces elimination of dendritic spines that were formed after extinction. Thus, within vastly complex neuronal networks, fear conditioning, extinction and reconditioning lead to opposing changes at the level of individual synapses. These findings also suggest that fear memory traces are partially erased after extinction.     

纯水相SEC法分离水溶性小金纳米簇的研究

以纯水为流动相,用PL aquagel-OH 40色谱柱与自动馏分收集器,在Agilent 1100高效液相色谱仪上,对巯基丁二酸(MSA)单层包裹的粒径约1-5 nm的水溶性小金纳米簇进行尺寸排阻色谱(SEC)分离,环境友好,操作简便,重现性好.TEM表征和粒径统计分析表明:在不同洗脱时间收集的金纳米簇馏分与相应的原始样品相比其粒径分布均得到改善,尺寸分布明显窄化.金纳米簇在所用色谱柱上的分离主要按尺寸排阻机制进行.     

The role of jumps in the agricultural futures market on forecasting stock market volatility: New evidence

In this study, we explore the effect of cojumps within the agricultural futures market, and cojumps between the agricultural futures market and the stock market, on stock volatility forecasting. Also, we take into account large and small components of cojumps. We have several noteworthy findings. First, large jumps may lead to more substantial fluctuations and are more powerful than small jumps. The effect of cojumps and their decompositions on future volatility are mixed. Second, a model including large and small cojumps between the agricultural futures market and the stock market can achieve a higher forecasting accuracy, implying that large and small cojumps contain more useful predictive information than cojumps themselves. Third, our conclusions are robust based on various robustness tests such as the realized kernel, expanding forecasts, different forecasting windows, different jump tests, and different threshold values.