SINGULAR CROSS SPECTRUM APPLIED ON DIAGNOSTIC ANALYSIS OF AIR-SEA INTERACTION

The Singular Cross Spectrum Analysis (SCSA) method was employed to investigate the coupledperiods of air-sea/mid-low circulation interaction using 1951-1993 500-hpa geopotential heights and the seasurface temperature (SST) in the Northern Hemisphere. Results show that air-sea correlation is noticeable on the3-7 year scale which is similar to ENSO circle. In this sense, ENSO is a Strong signal. Quasi-ten/quasi-four yearperiods are prominent in the mid-low interaction, and quasi-if our year oscillation is uniform with the ENSO circle.Studies indicate that the mid-low interaction possess quasi-ten year oscillation beside being affected by ENSO.

Experimental study of cross-section ratios in the collisions of C^(q+) and O^(q+) (q=1-4) on atomic helium in strong-interaction region

We have measured the cross-section ratios of helium induced by Cq＋ and Oq＋ （q = 1-4） in an energy range from 20 keV/amu to 500 keV/amu, and obtained the two-dimensional spectra by employing the coincidence method combined with the MPA-3 data acquisition system. Hence, we obtain the ratios of total single-ionization cross-sections （SI, SC, SLSI, and DLSI）, total double-ionization cross-sections （DI, DC, TI, SLDI, and DLDI） and cross-sections of every process （SI, SC, SLSI, DLSI, DI, DC, TI, SLDI, and DLDI）, which induce the single-ionization and double-ionization, to the total cross sections respectively. The competitive relations between the reaction-channels and the experimental data law of each reaction-channel are revealed explicitly, and the qualitative explanations involved in those results are also presented accordingly.

Comparative study on wheel-rail dynamic interactions of side-frame cross-bracing bogie and sub-frame radial bogie

Improving freight axle load is the most effective method to improve railway freight capability; based on the imported technologies of railway freight bogie, the 27 t axle load side-frame cross-bracing bogie and sub-frame radial bogie are developed in China. In order to analyze and compare dynamic interactions of the two newly developed heavy-haul freight bogies, we establish a vehi- cle-track coupling dynamic model and use numerical calculation methods for computer simulation. The dynamic performances of the two bogies are simulated separately at various conditions. The results show that at the dipped joint and straight line running conditions, the wheel-rail dynamic interactions of both bogies are basically the same, but at the curve negotiation condition, the wear and the lateral force of the side-frame cross-bracing bogie are much higher than that of the sub-frame radial bogie, and the advantages become more obvious when the curve radius is smaller. The results also indicate that the sub- frame radial bogie has better low-wheel-rail interaction characteristics.

The Role of the Neutron-proton Interaction in Anomalous Band Crossing of Rare Earth Nuclei

Experimental estimate values of the n-p interaction between the odd protonand the aligning neutron pairs for odd proton nuclei in the first band crossing region areextracted based on the experimental quasi-particle routhians.The reliability of thismethod and the relation between the anomalous band crossing and the neutron-protoninteraction are discussed.

Controllable cross-Kerr interaction between microwave photons in circuit quantum electrodynamics

We propose a scheme to enable a controllable cross-Kerr interaction between microwave photons in a circuit quantum electrodynamics （QED） system. In this scheme we use two transmission-line resonators （TLRs） and one superconducting quantum interference device （SQUID） type charge qubit, which acts as an artificial atom. It is shown that in the dispersive regime of the eircuit-QED system, a controllable cross-Kerr interaction can be obtained by properly preparing the initial state of the qubit, and a large cross-phase shift between two microwave fields in the two TLRs can then be reached. Based on this cross-Kerr interaction, we show how to create a macroscopic entangled state between the two TLRs.

Gel-Based Chemical Cross-Linking Analysis of 20S Proteasome Subunit-Subunit Interactions in Breast Cancer

The ubiquitin-proteasome system plays a pivotal role in breast tumorigenesis by controlling transcription factors, thus promoting cell cycle growth, and degradation of tumor suppressor proteins. However, breast cancer patients have failed to benefit from proteasome inhibitor treatment partially due to proteasome heterogeneity, which is poorly understood in malignant breast neoplasm. Chemical crosslinking is an increasingly important tool for mapping protein three-dimensional structures and proteinprotein interactions. In the present study, two cross-linkers, bis（sulfosuccinimidyl） suberate（BS3） and its water-insoluble analog disuccinimidyl suberate（DSS）, were used to map the subunit-subunit interactions in 20 S proteasome core particle（CP） from MDA-MB-231 cells. Different types of gel electrophoresis technologies were used. In combination with chemical cross-linking and mass spectrometry, we applied these gel electrophoresis technologies to the study of the noncovalent interactions among 20 S proteasome subunits. Firstly, the CP subunit isoforms were profiled. Subsequently, using native/SDSPAGE, it was observed that 0.5 mmol/L BS^3 was a relatively optimal cross-linking concentration for CP subunit-subunit interaction study. 2-DE analysis of the cross-linked CP revealed that α1 might preinteract with α2, and α3 might pre-interact with α4. Moreover, there were different subtypes of α1α2 and α3α4 due to proteasome heterogeneity. There was no significant difference in cross-linking pattern for CP subunits between BS3 and DSS. Taken together, the gel-based characterization in combination with chemical cross-linking could serve as a tool for the study of subunit interactions within a multi-subunit protein complex. The heterogeneity of 20 S proteasome subunit observed in breast cancer cells may provide some key information for proteasome inhibition strategy.

DIAMOND PORT JET INTERACTION WITH SUPERSONIC FLOW

Interaction flow field of the sonic air jet through diamond shaped orifices at different incidence angles （10 degrees, 27.5 degrees, 45 degrees and 90 degrees） and total pressures （0.10 MPa and 0. 46 MPa） with a Mach 5.0 freestream was studied experimentally. A 90 degrees circular injector was examined for comparison. Crosssection Mach number contours were acquired by a Pitot-cone five-hole pressure probe. The results indicate that the low Mach semicircular region close to the wall is the wake region. The boundary layer thinning is in the areas adjacent to the wake. For the detached case, the interaction shock extends further into the freestream, and the shock shape has more curvature, also the low-Mach upwash region is larger. The vortices of the plume and the height of the jet interaction shock increase with increasing incidence angle and jet pressure. 90 degrees diamond and circular injector have stronger plume vorticity, and for the circular injector low-Mach region is smaller than that for the diamond injector. Tapered ramp increases the plume vorticity, and the double ramp reduces the level of vorticity. The three-dimensional interaction shock shape was modeled from the surface shock shape, the center plane shock shape, and crosssectional shock shape. The shock total pressure was estimated with the normal component of the Mach number using normal shock theory. The shock induced total pressure losses decrease with decreasing jet incidence angle and injection pressure, where the largest losses are incurred by the 90 degrees, circular injector.

An overview of structure-soil-structure dynamic interaction research

The concept of structure-soil-structure dynamic interaction was introduced and the research methods were summarized.Based on lots of documents,a systematic summary of the history and current situation of structure-soil-structure dynamic interaction research considering adjacent structures was proposed as reference for researchers.The existing matter and the prospect of future research trend in this field was also examined.

MECHANICAL RELAXATION AND INTERMOLECULAR INTERACTION IN EPOXY RESINS/POLY (ETHYLENE OXIDE) BLENDS CURED WITH PHTHALIC ANHYDRIDE

The miscibility of the blend,composed of a bisphenol A epoxy resins (Diglycidyl ether of bisphenol A) (DGEBA) and poly(ethylene oxide) (PEG) and crosslinked by phthalic anhydride (PA) was studied using dynamic mechanical method. Single glass transition temperatures intermediate between the two pure components were observed for all blend levels. The secondary relaxation mechanism should relate to not only diester linkage, but also hydroxyether structural unit in the system. Fourier transform infrared spectroscopy (FTIR) is applied to study the curing reaction and intermolecular specific interaction of the system. The results indicate the PEO participates the crosslinking reaction, accelerates the curing reaction and make the reaction more perfect. The shifts of the hydroxyl band and carbonyl band demonstrate the presence of the intermolecular interaction in the cured blend. Moreover, the molecular interaction between the side hydroxyl in the hydroxyether units and the ether bond in PEO macromolecules is stronger.

Cross electro-nape-acupuncture ameliorates cerebral hemorrhageinduced brain damage by inhibiting necroptosis

BACKGROUND Receptor interacting protein kinase 1(RIPK1)-mediated cell death,including apoptosis and necroptosis,belongs to programmed cell death.It has been reported that RIPK1-mediated necroptosis exists in lesions of cerebral hemorrhage(CH).Electroacupuncture,a treatment derived from traditional Chinese medicine,could improve neurological impairment in patients with brain injury.AIM To investigate the protective role of cross electro-nape acupuncture(CENA)in CH,and clarify the potential mechanism.METHODS CH rat models were established,and CENA was applied to the experimental rats.Neurological functions and encephaledema were then measured.Necrotic cells in the brain of rats with CH were evaluated by propidium iodide staining.Necroptosis was assessed by immunofluorescence.Activation of the necroptosisrelated pathway was detected by western blot.Extraction of brain tissue,cerebrospinal fluid and serum samples was conducted to measure the expression and secretion of inflammatory cytokines by quantitative real-time polymerase chain reaction and enzyme-linked immunosorbent assay.RESULTS The necroptotic marker p-MLKL was detectable in the brains of rats with CH.Next,we found that CENA could ameliorate neurological functions in rat models of CH.Moreover,the upregulation of RIPK1-mediated necroptosis-related molecules in the brains of rats with CH were inhibited by CENA.Further investigation revealed that CENA partially blocked the interaction between RIPK1 and RIPK3.Finally,in vivo assays showed that CENA decreased the expression of the inflammatory cytokines tumor necrosis factor-α,interleukin-6 and interleukin-8 in CH rat models.CONCLUSION These findings revealed that CENA exerts a protective role in CH models by inhibiting RIPK1-mediated necroptosis.

SiamCross:孪生交叉的目标跟踪对象感知网络

近来,基于孪生架构的方法因其能在保持良好速度的同时取得较显著的性能,引起了视觉跟踪领域的广泛关注.然而,孪生网络分支通常是独立的,缺乏信息交互,这限制了模型性能的进一步提升.为了增强孪生网络分支的协作能力,本文提出基于孪生架构的交叉感知网络模型——SiamCross(Siamese Cross Object-Aware Network).孪生网络双分支特征提取是提升模型性能的首要核心操作,区分目标和语义背景在很大程度上依赖模型挖掘的特征鲁棒性.在SiamCross中,我们首先基于孪生网络分支的互监督,设计了全新的孪生交叉感知子网络(Siamese Cross-Aware Network,SCAN)用来提取鲁棒特征.SCAN允许孪生分支彼此全方位高效协同工作,使模板分支可充分利用特征丰富的上下文语义信息,对目标产生更具有区分性的表示;搜索分支结合模板特征,也主动学习到了目标的本质信息.另一方面,无锚框算法将跟踪任务直接映射为对每个像素的分类和回归,网络分支特征可各自聚焦于目标的局部与全局空间信息.以上两种特征具有很好的潜在局部-全局互补性.具体而言,回归特征学习到了更多的目标全局尺寸信息,但同时也引入了周围背景信息,而分类分支专注于学习局部中心定位信息.二者结合,有利于抑制回归特征的背景信息表达.同时,回归特征会在目标周边位置进行突出响应,揭示目标所在区域,也为分类分支进行定位提供了有益参考.为充分利用以上不同的分支空间特征信息来获得更精确的跟踪结果,我们又提出了新型的目标注意力交互网络(Obejct-Attention Interaction Network,OAIN),并将其融入到SiamCross中.OAIN包含并行交叉注意力模块(Parallel Cross Attention Module,PCA)和自适应可形变交叉对齐模块(Adaptive Deformable Cross Align Module,ADCA).PCA模块通过对分支中局部与全局信息的巧妙融合,提升了目标状态估计的准确性.为了进一步使回归特征和目标区域对齐,缓解特征对齐失焦导致的分类分支参考信源可靠性大幅度降低,我们为ADCA模块设计了自适应空间转换操作,可以使得回归特征更好反映目标所在区域.最终,ADCA模块完善了无锚框网络的高效交互机制.最后,我们在OTB2015、VOT2018/2019、GOT-10k和LaSOT五个具有挑战性的公开基准中对SiamCross进行了详尽的实验评估.实验结果显示,SiamCross与当前先进的跟踪器SiamRPN++、ATOM及DiMP相比,均取得了更优异的综合表现,并且可实现实时跟踪.

Ar^+/Ar, O_2^+/O_2 and N_2^+/N_2 Elastic Momentum Collision Cross Sections: Calculation and Validation Using the Semi-Classical Model

The aim of this paper is to obtain relevant sets of collision cross sections of the parent ions in low pressure discharges in argon, oxygen, and nitrogen, i.e., Ar＋ in Ar, O2＋ in O2 and N2＋ in N2. These ion data are first discussed and then validated from comparisons between the calculated transport coefficients and those measured in the literature. The elastic momentum transfer collision cross sections are determined from a semi-classical approximation for the phase shift calculation based on a 12-6-4 inter-particle potential while ion transport coefficients are determined versus the reduced electric field from Monte Carlo simulations.

Panel data models with cross-sectional dependence: a selective review

In this review, we highlight some recent methodological and theoretical develop- ments in estimation and testing of large panel data models with cross-sectional dependence. The paper begins with a discussion of issues of cross-sectional dependence, and introduces the concepts of weak and strong cross-sectional dependence. Then, the main attention is primarily paid to spatial and factor approaches for modeling cross-sectional dependence for both linear and nonlinear （nonparametric and semiparametric） panel data models. Finally, we conclude with some speculations on future research directions.

改进YOLOv4的遥感图像目标检测算法

为有效解决遥感图像目标检测算法在复杂背景下的检测效果不佳的问题,提出一种改进YOLOv4的目标检测算法。设计一种跨阶段残差结构,替换原主干网络的简单残差结构,降低模型参数量和计算负担;引入CBAM注意力机制,加强CSP模块间有效特征交互;使用跨阶段分层卷积模块重构特征融合阶段对深层特征图的处理方式,防止网络退化和梯度消失;采用Mish激活函数,增强融合网络对非线性特征的提取能力。在RSOD、DIOR数据集上的实验结果表明,改进YOLOv4算法的测试mAP相比原YOLOv4算法分别高出4.5%、7.3%,其检测速度分别达到48 fps、45 fps,在保证实时性的同时检测精度有较大提升。

基于复合跨模态交互网络的时序多模态情感分析

针对多模态情感分析中存在的不同模态间语义特征差异性导致模态融合不充分、交互性弱等问题,通过研究分析不同模态之间存在的潜在关联性,搭建一种基于复合跨模态交互网络的时序多模态情感分析(CCIN-SA)模型。该模型首先使用双向门控循环单元和多头注意力机制提取具有上下文语义信息的文本、视觉和语音模态时序特征;然后,设计跨模态注意力交互层,利用辅助模态的低阶信号不断强化目标模态,使得目标模态学习到辅助模态的信息,捕获模态间的潜在适应性;再将增强后的特征输入到复合特征融合层,通过条件向量进一步捕获不同模态间的相似性,增强重要特征的关联程度,挖掘模态间更深层次的交互性;最后,利用多头注意力机制将复合跨模态强化后的特征与低阶信号做拼接融合,提高模态内部重要特征的权重,保留初始模态独有的特征信息,将得到的多模态融合特征进行最终的情感分类任务。在CMU-MOSI和CMUMOSEI数据集上进行模型评估,结果表明,CCIN-SA模型相比其他现有模型在准确率和F1指标上均有提高,能够有效挖掘不同模态间的关联性,做出更加准确的情感判断。

知识交流生态视角下的用户跨平台知识交流行为偏好研究

[目的/意义]从知识交流生态视角出发,探究基于视频资源的用户跨平台知识交流行为偏好规律,为用户利用多平台开展交互、实现高价值知识交流提供有效指导。[方法/过程]通过对齐在B站和微博上进行跨平台知识交流的用户,并获取他们在两个平台发布的全量视频内容7215条和同质化视频的330367条评论文本,基于知识生态系统框架,分别针对知识生产者和知识分解者进行内容发布和评论互动等跨平台知识交流偏好的探究。[结果/结论]知识生产者会在不同平台规律性地发布同质化视频以获得知识消费者更好的互动反馈,且根据平台定位和平台功能差异进行标题差异化展示;跨平台知识交流中,知识分解者的评论互动偏好与平台特性相关。[局限]仅选择B站和微博中高影响力知识生产者群体进行研究,后续研究中可对更为多元的知识交流平台以及更丰富的大众知识生产者行为规律进行拓展分析。

基于XFLOW-ABAQUS联合仿真的十字形伞绕流流柔耦合数值模拟

以降落伞为代表的流柔耦合问题是一个涉及柔性结构大变形与周围流场相互耦合的复杂非线性问题.针对此问题,应用XFLOW-ABAQUS联合仿真对流柔耦合问题进行数值模拟.采用该方法对不同透气性的旗帜在Re=2.2×10^(4)气流中的拍动进行数值模拟,并与实验数据进行对比,验证该方法的有效性.然后对Re=6.2×10^(5)来流中柔性十字形伞的气动特性进行数值模拟,并与实验数据以及使用LS-DYNA的数值模拟结果进行比较,分析了十字形伞周围的流场特征,进一步证明了XFLOW-ABAQUS联合仿真在模拟流柔耦合问题上具有良好的表现.

跨域无人集群协同反潜搜索方法研究

针对海上反潜搜索问题,提出一种将无人艇作为无人机的通信中继的跨域无人集群协同搜索方法,采用数字网格地图表征任务区域,构建了跨域平台的运动学模型;提出了跨域无人系统的协作方法,设计了无人系统之间的分布式信息融合机制;设计了异构平台的搜索目标函数,引导无人系统在搜索任务中实时决策。仿真结果表明:所提方法能够有效适用于文中反潜搜索任务想定,跨域平台之间能够通过优势互补提高搜索效能。

基于结构功能交叉神经网络的多模态医学图像融合

针对多模态医学图像融合中存在纹理细节模糊和对比度低的问题,提出了一种结构功能交叉神经网络的多模态医学图像融合方法。首先,根据医学图像的结构信息和功能信息设计了结构功能交叉神经网络模型,不仅有效地提取解剖学和功能学医学图像的结构信息和功能信息,而且能够实现这两种信息之间的交互,从而很好地提取医学图像的纹理细节信息。其次,利用交叉网络通道和空间特征变化构造了一种新的注意力机制,通过不断调整结构信息和功能信息权重来融合图像,提高了融合图像的对比度和轮廓信息。最后,设计了一个从融合图像到源图像的分解过程,由于分解图像的质量直接取决于融合结果,因此分解过程可以使融合图像包含更多的细节信息。通过与近年来提出的7种高水平方法相比,本文方法的AG,EN,SF,MI,QAB/F和CC客观评价指标分别平均提高了22.87%,19.64%,23.02%,12.70%,6.79%,30.35%,说明本文方法能够获得纹理细节更清晰、对比度更好的融合结果,在主观视觉和客观指标上都优于其他对比算法。

面向目标交互图神经网络的多模态方面级情感分析

对于多模态方面级情感分析任务,除了需要提取出文本和图像的表示,还需要将它们与方面语义信息相结合处理。然而,以往的相关方法对方面与文本和图像信息之间的交互处理不够充分,即使使用注意力机制建立起模态全局之间的关联,也难以在细粒度表达出它们的交互。为了充分进行多模态之间细粒度上的信息交互,提出一种面向目标交互图神经网络,围绕文本、图像和方面三者的关系建模,采用交叉注意力获取面向方面目标的文本和图像全局表示;建立多模态交互图,以连接不同模态的局部及全局表示节点;使用图注意力网络在粗细两个粒度上充分融合特征。在两个基准数据集上进行实验,结果表明该模型相比于仅使用注意力机制的模型,具有更佳的情感分类效果。