基于保局部核RVM的说话人识别方法

针对说话人语音特征随音量、情绪、健康等因素变化呈现出的复杂分布结构,提出一种基于保局部核相关向量机(RVM)的说话人识别方法。在RVM模型所采用的高斯核函数中引入相似度因子,以保留数据局部结构,构成保局部核RVM模型。在模型训练过程中采用快速算法以避免大型矩阵逆操作,减少计算量,可适用于大样本场合。应用结果表明,该方法能加快测试速度,提高分类精度。     

基于三维体的多分辨率视频分割算法

传统视频分割算法无法处理视频中庞大的数据量,且分割结果需要保持时空连续性。针对该问题,提出一种基于三维体的多分辨率视频分割算法。在三维空间中对整个视频体进行分割,以保证时空连续性,采用多分辨率的思想,减少计算开销。实验结果表明,该算法能减少传统分割算法的时空耗费,修正连续帧的局部分割结果。     

BitTorrent流量控制方案

设计并实现了一个克隆Tracker以减少互联网中BitTorrent (BT)流量占用过多进出口带宽的问题。当BT客户端向Tracker发出列表请求时,Peers列表请求数据包被路由到克隆Tracker。克隆Tracker使用不同于原Tracker的列表选择算法,代替原Tracker向BT客户端返回Peers列表。克隆Tracker从原Tracker中获得并更新列表。在校园网的实地测试结果表明：在不修改BT客户端和原Tracker服务器的条件下,克隆Tracker能够有效控制所测网段BT流量占用的进出口带宽。     

基于局部多核支持向量机的视频镜头边界检测

提出了一种基于局部多核支持向量机的视频镜头边界检测方法.利用视频图像相邻帧的时空信息构建视频中间特征,在此基础上利用局部多核支持向量机将视频帧划分为边界帧和非边界帧.为了提高基于全局优化的多核支持向量机的检测精度,利用局部敏感哈希算法将视频帧投影全哈希子空间,结合多核学习方法为各个哈希子空间构建局部多核支持向量机,利用SMOTE上采样技术解决了视频图像边界帧和普通帧的不平衡问题.试验结果表明,本文提出的镜头边界检测方法的金全率和查准率得到了提高.     

基于LATE的Hadoop数据局部性改进调度算法

调度问题是目前云计算研究中的热点问题,其目的是如何协同云计算资源,使其得到充分合理的利用。数据局部性是特定云平台Hadoop的主要特性之一。针对该特性,在Hadoop原有调度算法LATE的基础上提出了一种基于数据局部性的改进算法,以解决数据局部性带来的慢任务备份执行时读取数据要占用大部分时间而影响其处理速率的问题。最后,对该算法进行了实验及性能分析,并验证了算法在提高任务的响应时间和整个系统吞吐率方面有很大改进。     

基于流形学习的局部保持PCA算法在故障检测中的应用

提出一种新的基于流形学习的数据降维及特征提取方法:局部保持 PCA 算法(LPPCA).通过在 PCA 的优化目标中融入流形学习的思想,不仅使投影得到的低维空间和原始样本空间具有相似的全局结构,并且保持了相似的局部近邻结构,克服了传统 PCA 方法只关注全局结构特征而忽略局部流形特征的缺陷,同时给出了 LPPCA 在故障检测中的应用方法. S-Curve 和 Swiss-roll 曲面数值仿真和 TE 过程仿真验证了算法的有效性和优越性.     

一种新的蛋白质结构预测多模态优化算法

针对现阶段药物设计中对于蛋白质结构多模态的需求,提出了一种基于排挤差分进化策略的多模态优化算法.为了降低蛋白质构象空间求解的复杂度,算法采用能量极小化过程,有效缩小了可行域的搜索空间;同时,为了有效地平衡多模态优化问题的局部收敛性和模态多样性,在排挤差分进化算法的框架下,在保证算法收敛速度的前提下,算法采用空间局部性原理,同时随机选取不同交叉策略的集结思想又有效改善了种群的多样性.以脑啡肽为例,算法不仅得到了其全局最稳定结构,还获得了一系列局部最优结构.     

EFL learners’ use of online reading strategies and comprehension of texts: An exploratory study

This study investigated EFL learners’ online reading strategies and the effects of strategy use on comprehension. To fulfill the purposes of this study, a Web-based reading program, English Reading Online, was created. Thirty applied English majors, divided into a high group and a low group based on their proficiency levels, were asked to read four authentic online texts; two were appropriate to the students’ level of proficiency, and two were more difficult. Results from data analysis showed that the use of support strategies dominated the strategy use and contributed to most of the comprehension gains, but an exclusive dependence on support strategies did not successfully predict the increase in scores on main ideas and details when the students were reading more challenging texts. On the whole, the use of global strategies significantly contributed to better comprehension, especially for low proficiency students.     

Courseware development for semiconductor technology and its application into instruction

This study reports on the development of ESP (English for specific purposes) courseware for semiconductor technology and its integration as a “silent partner” into instruction. This kind of team-teaching could help overcome current problems encountered in developing ESP in Taiwan. The content of the material under discussion includes general knowledge about fundamental theories, process technologies and applications within the semiconductor industry. In the design of the whole courseware, five skills for learning English (listening, speaking, reading, writing, and translation) have been considered and a 3D multimedia technique has been used to promote learning interest, student engagement, and efficiency. The design of the courseware endeavors to mainly follow principles by [Mayer, R. E. (2001). Multimedia learning. New York: Cambridge University Press]. Students report they have benefited from the courseware implementation. Initial evaluations suggest that students are satisfied with practices for learning professional knowledge and English skills provided by the courseware. They report that the multimedia-assisted environment of the courseware promotes learning effectiveness. Students with higher achievement on the posttest showed better participation and motivation, made greater use of the multimedia and had a better understanding of the English content so that they are more competent to function in a professional and learner-centered ESP course using the courseware.     

Computer-supported negotiation of course content

Students learn more effectively with personally meaningful tasks. Thus, students learn more if they have a say in deciding what specific topics and examples are being discussed in class. Naturally, the instructor knows what topics are important to cover in a course and which ones might be optional. Finding the right balance between students’ preferences and the instructor’s requirements is not so easy and thus may prevent this kind of shared control of the classroom from being realized. This article describes how the instructor’s and students’ interests can be used to generate a list of course topics that satisfies both parties. However, instead of adding technology to the classroom, technology is used to improve the classroom experiences. Specifically, it is shown how course topics can be assigned to specific students maximizing what is meaningful to the students and satisfies the course parameters as defined by the instructor. This problem can be formulated as a variation of the linear assignment problem and solved with a binary linear program. Results from actual and simulated courses are discussed and generalizations of the topic assignment problem presented.