网络对高等体育院校大学生思想政治工作的影响

针对网络在当代大学生中的普遍运用而引起的系列问题,结合本职工作,深刻剖析了高等体育院校大学生思想和行为特点,重点阐述了网络对体育院校大学生的负面影响,并由此引申出如何更好的利用网络来加强高等体育院校大学生思想政治工作的问题。     

基于GAN序列的卫星数据接收系统故障定位

现有的卫星数据接收系统故障定位方法受困于专家知识的难以获取。提出一种基于生成对抗网络(GAN)序列的故障定位方法GANseq。利用故障传播特性,将接收系统分解为M个信号处理单元(SPU),形成一个按照信号处理顺序排列的SPU序列;将故障定位问题转化为有序排列的M个异常检测子问题,其中第m个子问题为对前m个SPU的联合状态的异常检测;然后针对每个子问题采用基于GAN的状态参数检测模型,进而形成GAN序列;综合分析该序列在所有子问题上的检测结果以定位故障SPU。利用实际接收系统进行的实验证明所提出的GANseq不仅从数据驱动的角度完成了故障定位,而且在总体上降低了虚警率,并提高了定位精度。     

基于深度强化学习方法的无线多跳网络能量高效机会路由

由于机会路由能够利用无线信道的广播特性和有损特性,因此一直是提高无线网络路由性能的一个很有效的途径。提出一种基于深度强化学习的无线多跳网络能量高效机会路由算法,该算法使得智能体能够通过训练学习最优的路由策略,以通过机会路由的方式减少传输时间,同时平衡能耗延长网络寿命。此外,本算法还可以极大地缓解冷启动问题并获得较好的初始性能。仿真结果表明,与现有算法相比,该算法具有更好的性能。     

基于深度强化学习的低轨卫星下行功率分配方案

当前的卫星资源分配方案大多为同步轨道卫星设计,针对低轨卫星的高动态特性,以及存在频率和功率资源受限的问题,提出一种基于深度强化学习的功率分配算法。首先对低轨卫星功率分配场景进行建模,引入一种时隙划分方案来简化低轨卫星的动态特性模型,进一步提出一种基于深度强化学习算法的功率分配策略,该策略通过调节单颗低轨卫星各个波束中子载波的功率值,降低同频干扰,能达到提升低轨卫星频谱效率的目的。仿真结果表明,所提算法能够在较短时间内收敛并达到稳定状态,在总功率一定的条件下,该方案能有效提升单颗低轨卫星的吞吐量,其频谱效率明显高于注水算法和Q学习算法。     

The Future of the Past: Archaeology and Anthropology on the World Wide Web

In the final scene of Raiders of the Lost Ark, a crate containing the object of Indiana Jones' quest is wheeled into an immense warehouse for indefinite storage and questionable research access. Unfortunately, this fate is not all that far from reality. Collections of archaeological and ethnographic materials ranging from stone axes, broken potsherds, and carved monuments to baskets, ceremonial masks, and skin canoes have been held by museums collections since the Renaissance. However, their inestimable value and unique conservation and curatorial requirements often conspire to remove them from the reach of all but the most diligent scholars. The potential of the Web to enhance the quality of research on archaeological and ethnographic collections is enormous. This paper will examine ways that one can use the Web to enhance research and improve access to a variety of materials; while there are many other resources for archaeology available on the Web, this paper focusses on museum-related sites. It will also explore the potential of the Web for innovative research strategies. Digitization of catalogs, associated documents, and images to help one locate and study collections and specific artifacts are just one approach. Others include the connection of devices to the Web, such as cameras and microscopes, the creation of virtual reference collections, and the establishment of research networks that will enhance the identification and analysis of material culture. This paper will also consider the role the Web could play in issues of cultural property, contributing to and in many ways intensifying ongoing debates of ownership, curation, conservation, and repatriation of sensitive materials.     

基于概率神经网络的文本自动分类研究

本文提出了将一种径向基网络的重要变形———概率神经网络应用于文本自动分类的研究 ,与常用的K 最邻近法相比 ,具有一定的优势 ,特别是在训练集中各类的训练样本数很不平衡时 ;与BP等其他神经网络相比 ,其特点是需调节的参数少 ,不需确定隐层数、隐层中的神经元数量等网络结构 ,比较容易使用。此外 ,从研究中的不同特征选择的评价函数来看 ,它们对分类有一定的影响 ,应用X2 统计进行特征选择的分类正确率最高 ,其次是文本证据权 ,而期望交叉熵的效果最差 ,说明特征选择在文本自动分类中也是非常重要的。     

网络环境下数字图书馆的安全与防范措施

阐述了网络环境下影响数字图书馆安全的因素，提出了保护数字图书馆信息安全的防范措施。     

Update: The Invisible Farmer: securing Australian farm women’s history

This article reviews the Invisible Farmer project – the largest study of Australian women on the land – a third of the way through its three year funding period. It introduces the aims of this collaborative and multidisciplinary project, focusing on the role of records and oral histories in the creation of new narratives, and the task of raising public awareness of the important historical and contemporary contributions made by Australian women farmers. We examine the way digital platforms work to curate records and promote research, and outline the important role that the Australian Women’s Register, a core project of the Australian Women’s Archives Project, will play in achieving outcomes.     

Network member support marshaling in response to another’s relational transgression experience

Relational transgressions are emotionally difficult for those who directly experience the transgression. Less is known about how transgressions affect those located within the personal network of a relational transgression victim. This study explored how network members engaged in support marshaling in response to another’s transgression experience. Results indicate that some participants marshaled support, citing four reasons for their actions: perceived insufficiency, camaraderie, personal ramifications, and personal values. We explore these reasons for marshaling support and discuss them within the broader context of support marshaling and relational transgressions.     

The density of social networks moderates effects of intergroup contact

This study tests how the density of the social network in which intergroup contact takes place might affect the extent to which contact improves intergroup attitudes. Having contact with more outgroup members in dense social networks, in which everybody knows each other, may reinforce contact's positive effect. In this case, outgroup contact is shared with ingroup members, which suggests positive ingroup norms toward the outgroup. Alternatively, more contact in denser networks may improve intergroup attitudes less because density may increase subtyping or reduce the salience of ethnic group memberships. These competing hypotheses are tested among white American adults in a nonprobability online sample (N = 305) and in a representative national sample (N = 1270). In both studies, contact is associated with more positive attitudes toward racial outgroups but the positive contact effect is weakened if that contact takes place in a denser social network.