基于流水光总线阵列上的快速可扩展并行排序算法

在Y．Pan提出的基于流水光总线阵列模型(LARPBS)上使用N个处理器对N个元素进行排序在最好情况下以O(logN)时间，最坏情况下以O(N)时间完成的并行排序算法的基础上，提出了一种LARPBS模型上的可扩展的快速并行排序算法，对N个元素进行排序，使用p(1≤P≤N)个处理器在最好情况下以O(NlogN／p)时间，最坏情况下以O(N^2／p)时间完成排序。另外还提出了一种LARPBS模型上改进的快速高效并行排序算法，该算法对N个元素进行排序使用N个处理器在最好情况下以O(log√N)时间、最坏情况下以O(√N)时间完成排序。     

Forces of Change: Ten Trends that Will Impact the Internet Over the Next Five Years

Regardless of how organizations measure Internet growth and usage, it is certain that the forces of change that affect the Internet will continue to evolve. in evaluating the economics of competing technologies and determining how organizations can get a return on their technology investments, this article reviews ten technology and consumption trends on the Internet. They are (1) multiple bandwidth delivery, (2) linguistic and cultural diversity, (3) privacy, (4) laws, (5) restricting access, (6) taxation of Internet commerce, (7) information age law enforcement requirements, (8) surviving on a crowded Web, (9) new sources of competition, and (10) information warfare and economic terrorism.     

Automatic speaker age and gender recognition using acoustic and prosodic level information fusion

The paper presents a novel automatic speaker age and gender identification approach which combines seven different methods at both acoustic and prosodic levels to improve the baseline performance. The three baseline subsystems are (1) Gaussian mixture model (GMM) based on mel-frequency cepstral coefficient (MFCC) features, (2) Support vector machine (SVM) based on GMM mean supervectors and (3) SVM based on 450-dimensional utterance level features including acoustic, prosodic and voice quality information. In addition, we propose four subsystems: (1) SVM based on UBM weight posterior probability supervectors using the Bhattacharyya probability product kernel, (2) Sparse representation based on UBM weight posterior probability supervectors, (3) SVM based on GMM maximum likelihood linear regression (MLLR) matrix supervectors and (4) SVM based on the polynomial expansion coefficients of the syllable level prosodic feature contours in voiced speech segments. Contours of pitch, time domain energy, frequency domain harmonic structure energy and formant for each syllable (segmented using energy information in the voiced speech segment) are considered for analysis in subsystem (4). The proposed four subsystems have been demonstrated to be effective and able to achieve competitive results in classifying different age and gender groups. To further improve the overall classification performance, weighted summation based fusion of these seven subsystems at the score level is demonstrated. Experiment results are reported on the development and test set of the 2010 Interspeech Paralinguistic Challenge aGender database. Compared to the SVM baseline system (3), which is the baseline system suggested by the challenge committee, the proposed fusion system achieves 5.6% absolute improvement in unweighted accuracy for the age task and 4.2% for the gender task on the development set. On the final test set, we obtain 3.1% and 3.8% absolute improvement, respectively.     

Three-dimensional representations of photo-induced electron transfer rates in pyrene-(CH2)n-N,N'-dimethylaniline systems obtained by three electron transfer theories

The observed rates of photo-induced electron transfer (ET) from N,N'-dimethylaniline (DMA) to the excited pyrene (Py) in confined systems of pyrene-(CH(2))(n)-N,N'- dimethylaniline (PnD: n=1-3) were studied by molecular dynamic simulation (MD) and three kinds of electron transfer theories. ET parameters contained in Marcus theory (M theory), Bixon and Jortner theory (BJ theory) and Kakitani and Mataga theory (KM theory) were determined so as to fit the calculated fluorescence intensities with those obtained by the observed ET rates, according to a non-linear least squares method. Three-dimensional profiles of logarithm of calculated ET rates depending on two of three ET parameters, R, epsilon(0) and -DeltaG degrees were systematically examined with best-fit ET parameters of P1D. Bell shape dependencies of ET rate were predicted on R and on epsilon(0), and on -DeltaG degrees as well, by M theory and KM theory. The profiles of logarithm of ET rate calculated by BJ theory exhibited oscillatory dependencies not only on -DeltaG degrees , but also on R and on epsilon(0). Relationship between ET state and charge transfer complex was discussed with BJ theory.     

SUPANET基于OAM的保护交换研究

单层用户交换平台体系结构(Single-layer User-data switching Platform Architecture, SUPA)是基于面向以太网的物理帧时槽交换(Ethernet oriented Physical Frame Timeslot Switching,EPFTS)技术的一种未来Internet体系结构。研究了SUPA用户平台的OAM(Operation and Management or Operation,Administration and Maintenance)机制,以支持SUPANET域内的连通性诊断、故障诊断和故障恢复等功能。基于SUPA虚线路交换(Virtual Line Switching, VLS)服务,重点研究了基于OAM的保护交换机制。最后,基于QVL(QoS Virtual Line)和SVL(Shared Virtual Line)服务,仿真比较了SUPA用户平台中基于OAM的故障恢复的保护效果,验证了基于QVL的保护交换比基于SVL的保护交换具有更好的QoS保障能力。     

Combined effects of working environmental conditions in VDT work

The combined effects of city noise and luminance of the computer display were evaluated from the changes in lymphocytes and mental activities of participants. Healthy male students were tested under the following four experimental conditions: (1) a calculating task on a video display terminal (VDT) with luminance of 90 cd m(-2) without city noise; (2) a calculating task on a VDT with luminance of 20 cd m(-2) without city noise; (3) a calculating task on a VDT with luminance of 90 cd m(-2) with city noise of 70 dB(A); and (4) a calculating task on a VDT with luminance of 20 cd m(-2) with city noise of 70 dB(A). A visual reaction test (VRT) was performed, and critical flicker fusion frequency (CFF), heart rate (HR), numbers of circulating white blood cells (WBCs), lymphocyte subsets and subjective symptoms of fatigue were measured (1) before; (2) just after; and (3) 30 min after each 60 min test. Subjective symptoms of fatigue significantly increased just after experiments conducted under the two noisy conditions. VRT and CFF showed significant changes in the case of the high-luminance display with noise. WBCs and neutrophils showed significant increases in the two quiet conditions. These results suggested that high luminance with noise had the most effect on subjective fatigue and mental activities.     

CALL FOR PAPERSEI北大核心CSCD

2012 IEEE Multi-conference on Systems and Control(IEEE MSC 2012) October 3-5,2012 Dubrovnik Palace Hotel,Dubrovnik,CROATIA The 2012 IEEE Multi-Conference on Systems and Control(MSC 2012) will take place in Dubrovnik Palace Hotel,Dubrovnik,Croatia,on October 3 5,2012.MSC 2012 includes two international conferences sponsored and promoted by the IEEE Control Systems Society: The IEEE Conference on Control Applications(CCA) The IEEE International Symposium on Intelligent Control(ISIC)     

CUDA-enabled Sparse Matrix–Vector Multiplication on GPUs using atomic operations

Existing formats for Sparse Matrix–Vector Multiplication (SpMV) on the GPU are outperforming their corresponding implementations on multi-core CPUs. In this paper, we present a new format called Sliced COO (SCOO) and an efficient CUDA implementation to perform SpMV on the GPU using atomic operations. We compare SCOO performance to existing formats of the NVIDIA Cusp library using large sparse matrices. Our results for single-precision floating-point matrices show that SCOO outperforms the COO and CSR format for all tested matrices and the HYB format for all tested unstructured matrices on a single GPU. Furthermore, our dual-GPU implementation achieves an efficiency of 94% on average. Due to the lower performance of existing CUDA-enabled GPUs for atomic operations on double-precision floating-point numbers the SCOO implementation for double-precision does not consistently outperform the other formats for every unstructured matrix. Overall, the average speedup of SCOO for the tested benchmark dataset is 3.33 (1.56) compared to CSR, 5.25 (2.42) compared to COO, 2.39 (1.37) compared to HYB for single (double) precision on a Tesla C2075. Furthermore, comparison to a Sandy-Bridge CPU shows that SCOO on a Fermi GPU outperforms the multi-threaded CSR implementation of the Intel MKL Library on an i7-2700 K by a factor between 5.5 (2.3) and 18 (12.7) for single (double) precision.     

A new algorithm for analysis of the homology in protein primary structure

A new algorithm for analysis of the homology and genetic semihomology in protein sequence is described. It assumes the close relation between the compared amino acids and their codons in related proteins. The algorithm is based on the network of the genetic relationship between amino acids and, thus differs from the commonly used statistical matrices. The results obtained by using this method are more comprehensive than used at present, and reflect the actual mechanism of protein differentiation and evolution. They concern: (1) location of homologous and semihomologous sites in compared proteins; (2) precise estimation of insertion/deletion gaps in non-homologous fragments; (3) analysis of internal homology and semihomology; (4) precise location of domains in multidomain proteins; (5) estimation of genetic code of non-homologous fragments; (6) construction of genetic probes; (7) studies on differentiation processes among related proteins; (8) estimation of the degree of relationship among related proteins; (9) studies on the evolution mechanism within homologous protein families and (10) confirmation of actual relationship of sequences showing low degree of homology.     

视频内容分析技术

概述了基于内容的视频检索的方法和工作过程,研究的重点是突变镜头探测(基于像素的方法、模板匹配法、基于直方图的方法、基于视频特征的方法),渐变镜头探测(双重比较法、基于模型的方法、基于压缩域的方法),关键帧提取的关键技术(基于颜色特征的方法,基于镜头边界的方法、基于镜头的方法、基于运动分析的方法等),在综合分析了各类方法的工作机理和优缺点的基础上,提出了一个优化的视频内容分析检索框架.     

Inter-relationships among dominance, energetic and tense arousal, and pleasure, and differences in their impacts under online vs. offline environment

Only a few studies have investigated dominance and observed mixed results. The impact of arousal on pleasure and behavior has also produced inconsistent or no meaningful results. In addition, emotional responses such as dominance, arousal, and pleasure have been assumed to be independent, which have resulted in a void in the relationships research among these constructs. Based on this research tradition, the present study proposed an inter-relationship among dominance, energetic and tense arousal, pleasure, and their impact on intention. A two-dimensional concept of energetic arousal and tense arousal has never been adopted and studied in previous retail research. Dominance, energetic arousal, and tense arousal are introduced to explain inconsistent relationships found in previous retail consumer studies. Results from a survey of 406 consumers (217 from offline store customers and 189 online store users) using PLS demonstrated that that (1) while dominance has a significant positive and/or negative effect on both energetic and tense arousal, it has no impact on pleasure and intention under both offline and online environment; (2) the effect of dominance on tense arousal is not statistically significant in an online shopping environment; (3) both energetic and tense arousal has a positive and/or negative impact on pleasure; (4) energetic arousal affects intention directly and indirectly mediated by pleasure, whereas tense arousal has an indirect effect on intention; and (5) the results found in testing differences in path coefficients under the two different shopping environment suggested that (1) the impact of dominance on energetic arousal is stronger under offline shopping context; (2) the effect of dominance on tense arousal is stronger in online environment; (3) the effect of energetic arousal on pleasure is stronger in offline shopping context; and (4) the impact of tense arousal on pleasure is stronger under offline shopping environment. Theoretical and practical implications are discussed in conclusion.     

Guest Editorial on Vehicular Networking Protocols

Vehicular communications is receiving an increased interest by the research community both from industry and academia. These technologies known as Vehicular Ad hoc Networks (VANET) include vehicle-to-infrastructure (V2I), and vehicle-to-vehicle (V2V) communications and can be based on various communications technologies. This special issue on Vehicular Networking Protocols technologies highlights new trends on networking protocols and their issues, further focusing on the Medium Access Control (MAC) layer and Physical (PHY) layer challenges. It also underlines on algorithms and protocols for data aggregation and group-based networking in vehicular networks.     

Effects of alcohol and caffeine on maritime navigational skills

Twelve experienced navigators participated in an experiment to determine the effects of alcohol, caffeine, and an alcohol+caffeine mixture on performance during the following tasks: visual search, the search and location of items on a navigational chart (chartsearch) and the solving of maritime navigational problems. Alcohol (75 ml) produced impairment in performance on visual search (p<0.05) and navigational problem-solving (p<0.01). Caffeine was found to enhance performance on visual search (p<0.05) but not on the chartsearch, although a significant correlation was found between performance on the two tests (p<0.05). Caffeine was not found to improve the accuracy of navigational problem-solving (p>0.05). Neither alcohol nor caffeine had any significant effect on the speed of problem-solving (p>0.05).     

Determination of sunglint location and its characteristics on observation from a METEOSAT 9 satellite

On the basis of observations from a geostationary satellite, a method for the determination of sunlight location on the ocean surface, based on two parameters – (1) the Greenwich time and (2) the longitude of the satellite – was developed. The problem is solved in three stages: first, the position of the Earth in its orbit for any given point of time is determined; second, for this point of time, the relative position of Sun–Earth–satellite is defined; and, third, the latitude and longitude of the principal point of sunglint (PPS) was found. The outputs of the program based on this method are: (1) the geographical coordinates of PPS; (2) the boundary of a simultaneously illuminated and observed region of the Earth; (3) the contour of the sunglint (disk) on the smooth ocean surface; and (4) the distribution of the sunglint brightness on the rough (waved) ocean surface. This method is applied to detect sunglint characteristics in the images gathered from the METEOSAT 9 satellite.     

Analysis of the interaction between the dendritic conductance density and activated area in modulating alpha-motoneuron EPSP: statistical computer model

Five reconstructed alpha-motoneurons (MNs) are simulated under physiological and morphological realistic parameters. We compare the resulting excitatory postsynaptic potential (EPSP) of models, containing voltage-dependent channels on the dendrites, with the EPSP of a passive MN and an active soma and axon model. In our simulations, we apply three different distribution functions of the voltage-dependent channels on the dendrites: a step function (ST) with uniform spatial dispersion; an exponential decay (ED) function, with proximal to the soma high-density location; and an exponential rise (ER) with distally located conductance density. In all cases, the synaptic inputs are located as a gaussian function on the dendrites. Our simulations lead to eight key observations. (1) The presence of the voltage-dependent channels conductance (g(Active)) in the dendrites is vital for obtaining EPSP peak boosting. (2) The mean EPSP peaks of the ST, ER, and ED distributions are similar when the ranges of G (total conductance) are equal. (3) EPSP peak increases monotonically when the magnitude of g(Na_step) (maximal g(Na) at a particular run) is increased. (4) EPSP kinetics parameters were differentially affected; time integral was decreased monotonically with increased g(Na_step), but the rate of rise (the decay time was not analyzed) does not show clear relations. (5) The total G can be elevated by increasing the number of active dendrites; however, only a small active area of the dendritic tree is sufficient to get the maximal boosting. (6) The sometimes large variations in the parameters values for identical G depend on the g(Na_step) and active dendritic area. (7) High g(Na_step) in a few dendrites is more efficient in amplifying the EPSP peak than low g(Na_step) in many dendrites. (8) The EPSP peak is approximately linear with respect to the MNs' R(N) (input resistance).     

基于等腰归一化距离的模糊粒度空间研究

本文将等腰归一化距离引入到模糊商空间中,提出了基于等腰归一化距离的模糊粒度空间理论.研究了它的结构和性质,并得到了四个重要结论.首先,下面3个叙述是等价的(定理3.2): (1) 给定X上的一个模糊等价关系;(2)给定X上的一个等腰归一化距离;(3)给定X上的一个分层递阶结构(或有序的粒度空间).其次,讨论了等腰归一化距离与Fuzzy等价关系间相互确定的对应关系,且都是一对多的关系(定理2.2,定理2.3).最后,给出了通过X上的模糊等价关系R诱导的等腰归一化距离d确定其引导的粒度上的度量dλ,且dλ正好是d在粒度X(λ)上压缩的等腰归一化距离(定理4.1), 同时给出了确定粒度空间上等腰归一化距离的方法.这些研究结论为模糊粒度计算的理论研究和应用提供了强有力的数字模型和工具, 同时表明模糊商空间的粒度计算可以在等腰归一化距离的范畴内进行,为模糊粒度计算提供了更为直观的几何解释.     

Efficient 3-dimensional GLV method for faster point multiplication on some GLS elliptic curves

We discover that two distinct efficient endomorphisms can both exist on some Galbraith-Lin-Scott (GLS) elliptic curves Galbraith et al. (2009) [4]. By using them we generalize the Gallant-Lambert-Vanstone (GLV) method Gallant et al. (2001) [5] for faster point multiplication on these curves to dimension 3, and give some implementation result which shows that our 3-dimensional GLV (3GLV) method runs in 0.897 the time of 2-dimensional GLV (2GLV) method as Galbraith et al. did in Galbraith et al. (2009) [4] for the point multiplication on these curves.     

基于GA-RBF网络的乙烯裂解炉在线操作优化

基于OPC技术实现过程数据的实时采集,并对所需变量进行数据滤波与异常检测,再利用RBF神经网络建立乙烯裂解炉过程多输入多输出(MIMO)裂解产物收率在线软测量模型以及模型校正方法,以乙烯和丙烯收率之和最大为目标,基于遗传算法对RBF神经网络模型进行操作优化,得到裂解过程的最优操作条件以指导生产.实际的工业应用表明,该方法提高了乙烯和丙烯的收率,具有良好的适应性和稳定性,对实际生产有重要的指导意义.     

Selected tree classifier combination based on both accuracy and error diversity

This paper proposes a method for combining multiple tree classifiers based on both classifier ensemble (bagging) and dynamic classifier selection schemes (DCS). The proposed method is composed of the following procedures: (1) building individual tree classifiers based on bootstrap samples; (2) calculating the distance between all possible two trees; (3) clustering the trees based on single linkage clustering; (4) selecting two clusters by local region in terms of accuracy and error diversity; and (5) voting the results of tree classifiers selected in the two clusters. Empirical evaluation using publicly available data sets confirms the superiority of our proposed approach over other classifier combining methods.