流计算模式下概率粗糙集三支决策的快速计算

概率粗糙集三支决策是不确定问题求解的一种重要理论,流计算模式是一种新型的动态内存计算形式,实施流计算模式下三支决策的快速动态计算是一项具有挑战性的新议题。本研究以流计算模式中的两个核心计算步骤即动态增量与动态减量作为研究对象,提出了一种流计算模式下概率粗糙集三支决策域的快速动态学习方法。首先对流计算模式中三支决策动态增量和动态减量的不同变化情况进行了数据建模。然后基于不同数据变化情况分别讨论了数据增量与数据减量时三支决策域的变化推理,并且基于上述理论给出了流计算模式下的三支决策动态增减学习算法。该算法能够以更低的时间复杂度获得与经典三支决策算法相同决策效果。最后通过八种UCI数据集的实验证明了流计算模式下三支决策动态增减学习算法在时间消耗上明显优于经典概率粗糙集三支决策算法,并且在不同阈值下具有稳定的决策效率。本研究表明了流计算模式下三支决策快速计算是可行的。     

大数据流式计算系统研究综述

互联网产业的迅速发展在带来数据规模爆炸式增长的同时,也使大数据呈现出越发鲜明的流式特征,传统的基于MapReduce的批处理模式难以满足流式大数据处理对于计算实时性的要求,因此,更为高效的流式计算系统得到了越来越广泛的应用。首先介绍了流式大数据的相关概念和特征,并对几类典型的流式计算系统实例进行了介绍和比较,分析了它们的系统架构、性能特点以及主要应用场景。最后总结并展望了大数据分析处理技术的发展前景。     

2n周期优秀二元序列生成及其特性分析简

The 2ⁿ-periodic binary sequence with high linear complexity and high k-error linear complexity is defined as an excellent sequence. We design a genetic algorithm for generating excellent sequences and studying their features. Choosing the N-periodic binary sequences, where N=8, 16, 32, k=N/4, we search the resulted sequences by the genetic algorithm with various parameters, and compute the linear complexity profiles of results sequences by using the Lauder-Paterson algorithm, to confirm that the obtained sequences are the real excellent sequences. By numerous experiments, we speculate that the k-error linear complexity of the N-periodic binary excellent sequence meets the formula LC_k(S)≤N-2k+1, when k=N/4、N/8 (we also do experiments on sequences with periods 64, 128 and 256). By the brute-force method we obtain that the proportion of the excellent sequence in all binary sequences of the same period is 1/4.     

Probabilistic Bit Stream Neural Chip: Theory

A design for a neural network chip is proposed using probabilistic bit streams to represent real values. This paper analyzes the performance of the proposed neurons in this design and demonstrates that very simple operations can be used to obtain the desired functionality. It is also shown that a suitable ‘activation function’ for neurons of this type can be obtained using the interaction of two probability distributions. Finally, the paper introduces a variant of the back-propagation learning algorithm which involves computing the derivatives of the output with respect to individual weights in a network of such units.     

基于辽河中下游来水来沙特点的分析研究

辽河是中国七大江河之一,同时辽河也是一条多沙河流。据统计,1954—2010年系列辽河下游六间房站多年平均含沙量达3.32 kg/m3,辽河中下游的泥沙淤积是一个亟待解决的问题。为对辽河中下游水沙特点更好的研究,摸清辽河中下游泥沙冲淤演变规律,特对辽河中下游各主要支流选取一个水文站作为控制,得出了辽河来水来沙特点是东部水量大,且西辽河是辽河干流泥沙最主要的来源。     

Seasonal Effects of Hydropeaking on Growth,Energetics and Movement of Juvenile Atlantic Salmon (Salmo Salar)

The performance and movements of juvenile Atlantic salmon Salmo salar exposed to variable water discharge (simulating hydropeaking) but with a stable water‐covered area were studied in six experimental stream channels, both during the winter and summer. Thirty fish were stocked into each channel, and the growth, body fat and movements of the fish were followed for about 2.5 months in each season. During the winter, no effect of hydropeaking was documented on performance or movement. In the summer, fish experiencing hydropeaking had lower body mass, lower body fat, and higher movement rates than the control fish. In general, effect sizes were small, and the rapid and frequent changes in water discharge and water level in the present study had small effects on the performance of juvenile Atlantic salmon. The cumulative long‐term effect at the population level is unknown, but a reduced growth rate of 10% and a reduction in body fat of 16% in the hydropeaking experiments in the summer might to some extent translate into increased smolt age and lower overwintering survival. Copyright © 2014 John Wiley & Sons, Ltd.     

Mesh Denoising using Extended ROF Model with L1 Fidelity

This paper presents a variational algorithm for feature‐preserved mesh denoising. At the heart of the algorithm is a novel variational model composed of three components: fidelity, regularization and fairness, which are specifically designed to have their intuitive roles. In particular, the fidelity is formulated as an L₁ data term, which makes the regularization process be less dependent on the exact value of outliers and noise. The regularization is formulated as the total absolute edge‐lengthed supplementary angle of the dihedral angle, making the model capable of reconstructing meshes with sharp features. In addition, an augmented Lagrange method is provided to efficiently solve the proposed variational model. Compared to the prior art, the new algorithm has crucial advantages in handling large scale noise, noise along random directions, and different kinds of noise, including random impulsive noise, even in the presence of sharp features. Both visual and quantitative evaluation demonstrates the superiority of the new algorithm.     

Computing Minimum Area Homologies

Calculating and categorizing the similarity of curves is a fundamental problem which has generated much recent interest. However, to date there are no implementations of these algorithms for curves on surfaces with provable guarantees on the quality of the measure. In this paper, we present a similarity measure for any two cycles that are homologous, where we calculate the minimum area of any homology (or connected bounding chain) between the two cycles. The minimum area homology exists for broader classes of cycles than previous measures which are based on homotopy. It is also much easier to compute than previously defined measures, yielding an efficient implementation that is based on linear algebra tools. We demonstrate our algorithm on a range of inputs, showing examples which highlight the feasibility of this similarity measure.     

MIMO Ad Hoc网络中一种新的流控多址接入协议

针对PRP-MAC协议中分组冲突导致网络吞吐量减小的问题,提出一种带有冲突避免机制的流控多址接入协议SCMA-CA。该协议可用于MIMO Ad Hoc网络,其在PRP-MAC定义的控制分组基础上,新增CTS分组类型,并利用MIMO技术的流控能力实现冲突避免和最大化数据流数目。仿真结果表明, SCMA-CA协议不仅可避免发送节点数增加时控制分组和数据分组的冲突,提高网络吞吐量,同时又适用于天线多样化的环境,能有效增强网络性能。     

一种用于流交换的代理重签名方案

为了解决大规模复杂网络环境下流交换的安全问题,首次将代理重签名技术应用到安全流交换中,提出一种用于流交换的基于陷门Hash函数的代理重签名方案.首先,针对陷门Hash函数在流交换应用中存在的密钥泄露问题,提出一种新的基于椭圆曲线的无密钥泄露的陷门Hash函数(EDL-MTH),并对其安全性加以证明;然后,基于EDL-MTH构造了一个在随机预言模型下可证明适应性选择消息攻击安全的代理重签名方案;最后,通过一个示例分析了该方案在安全流交换中的应用和性能.