Using multithreshold quadratic sigmoidal neurons to improveclassification capability of multilayer perceptrons

This letter proposes a new type of neurons called multithreshold quadratic sigmoidal neurons to improve the classification capability of multilayer neural networks. In cooperation with single-threshold quadratic sigmoidal neurons, the multithreshold quadratic sigmoidal neurons can be used to improve the classification capability of multilayer neural networks by a factor of four compared to committee machines and by a factor of two compared to the conventional sigmoidal multilayer perceptrons.     

线性复杂度为2n-2m的2n-周期序列的k-错线性复杂度*

线性复杂度和k-错线性复杂度是衡量密钥序列随机性的两个重要标准,运用Chan-Games算法,得到线性复杂度为2n-2m的2n-周期二元序列的k-错线性复杂度的所有可能的值,LCk(s)=0或2n－2m－2r+1+c,2n-2r+1+c。这一结果对于进一步探讨流密码密钥序列的安全性有重要的应用价值。     

Design of circular dot pattern code (CDPC) for maximum information capacity and robustness on geometric distortion/noise

One dimensional or linear bar code has been used for distribution purposes such as product information and distribution channel identification. Those linear bar codes can support only one directional code layout and also support limited code error detection capability. Two dimensional bar codes (e.g., QR code) extending one dimensional bar codes were developed in database and index based types. Database type barcodes embed full information bits and show weak recognition rate with geometric distortion. Index-based embed only the index information and requires additional network servers to interpret the index information, which leads to limited information storage capacity. Instead of using visible bar codes, we propose CDPC (circular dot pattern code), which is a dot based codes which is more invisible than bar codes. We design CDPC to be more robust to geometric distortion and noise than previous coding schemes using circular template matching. To maximize the information capacity and robustness, we use a circular dot patterns which is more robust to affine transformation. Code can be easily extended according to the number of data circles. If the number of data circle is n, then we can embed \( \left( {5\sum {_{{k = 2}}^{{n + 1}}{\text{k}}} } \right) - {\text{n}} \) data bits. In our experimentation, we set the number of circle to three, and resulting information capacity can be 42 bits per one code. To extract information from a CDPC codes, we perform (1) image capture, (2) identification of dots, (3) graph based topological analysis of dot patterns, (4) template matching between topological graphs using position symbols, and (5) information bit extraction with error correction capability. To evaluate information capacity under various geometric distortions, we experiment our CDPC with StirMark Benchmark’s affine transformation (simulation of geometric and noise attacks) and with real cell phone image captures. Our experimental results also show that our CDPC scheme achieves more robust recognition performance than those proposed in previous research works including QR code.     

Shannon zero error capacity in the problems of state estimation and stabilization via noisy communication channels

The paper addresses state estimation and stabilization problems involving communication errors and capacity constraints. Discrete-time partially observed unstable linear systems perturbed by stochastic exogenous disturbances are studied. Unlike the classic theory, the sensor signals are transmitted to the estimator or controller over a noisy digital communication link modelled as a stochastic stationary discrete memoryless channel. It is shown that the capability of the noisy channel to ensure almost sure stabilizability/observability of the plant is identical to exactly its capability to transmit information with zero probability of error. Specifically, it is demonstrated that the standard numerical characteristic of the latter capability, i.e., the Shannon zero error capacity of the channel, constitutes the border separating the cases where the plant is and respectively, is not stabilizable/observable with probability 1.     

半Bent函数和多输出布尔函数的构造

半bent函数是一类非线性度几乎最优且平衡的布尔函数,它弥补了bent函数的一些不足,如变元个数可以是奇数,具有平衡性.半bent函数可用于对称密码系统的设计和CDMA系统中的正交可变扩频码的构造.本文利用不相交线性码构造了一类新的半bent函数,设输入维度为n,当n=2k+1时,将F2^n划分为2^k+1个[n,k]线性码和1个[n,k+1]线性码,通过从该码集中选取合适线性码作支撑集来构造新的半bent函数.另一方面,多输出布尔函数(向量值函数)在应用中的效率更高,因此其使用场景更为广泛.本文同时利用不相交线性码构造了(n,n-k)平衡的多输出布尔函数,其中n/3相似文献   

A fast parallel algorithm for routing unicast assignments in Benesnetworks

This paper presents a new parallel algorithm for routing unicast (one-to-one) assignments in Benes networks. Parallel routing algorithms for such networks were reported earlier, but these algorithms were designed primarily to route permutation assignments. The routing algorithm presented in this paper removes this restriction without an increase in the order of routing cost or routing time. We realize this new routing algorithm on two different topologies. The algorithm routes a unicast assignment involving O(k) pairs of inputs and outputs in O(lg ² k+lg n) time on a completely connected network of n processors and in O(lg⁴ k+lg² k lg n) time on an extended shuffle-exchange network of n processors. Using O(n lg n) professors, the same algorithm can be pipelined to route α unicast assignments each involving O(k) pairs of inputs and outputs, in O(lg² k+lg n+(α-1) lg k) time on a completely connected network and in O(lg⁴ k+lg² k lg n+(α-1)(lg ³ k+lg k lg n)) time on the extended shuffle-exchange network. These yield an average routing time of O(lg k) in the first case, and O(lg³ k+1g k lg n) in the second case, for all α⩾lg n. These complexities indicate that the algorithm given in this paper is as fast as Nassimi and Sahni's algorithm for unicast assignments, and with pipelining, it is faster than the same algorithm at least by a factor of O(lg n) on both topologies. Furthermore, for sparse assignments, i.e., when k=O(1), it is the first algorithm which has an average routing time of O(1g n) on a topology with O(n) links     

Delay and Capacity Trade-offs in Mobile Wireless Networks with Infrastructure Support

In this paper,we investigate the trade-offs between delay and capacity in mobile wireless networks with infrastructure support.We consider three different mobility models,independent and identically distributed （i.i.d） mobility model,random walk mobility model with constant speed and L’evy flight mobility model.For i.i.d mobility model and random walk mobility model with the speed θ（1/n～（1/2））,,we get the theoretical results of the average packet delay when capacityis θ（1）,θ（1/n～（1/2）） individually,where n is the number of nodes.We find that the optimal average packet delay is achieved whencapacity λ（n） <（1/（2.n.log₂（1/（（1-e）-（k/n））+1））,where K is the number of gateways.It is proved that average packet delay D（n） dividedby capacity λ（n） is bounded below by （n/（k·w））.When ω（n～（1/2））≤KO（n（（1-η）·（α+1））/2）ln n） when K=o（n～η）（0≤η<1）.We also provethat when ω（1/2）≤K相似文献   

Outlier mining in large high-dimensional data sets

A new definition of distance-based outlier and an algorithm, called HilOut, designed to efficiently detect the top n outliers of a large and high-dimensional data set are proposed. Given an integer k, the weight of a point is defined as the sum of the distances separating it from its k nearest-neighbors. Outlier are those points scoring the largest values of weight. The algorithm HilOut makes use of the notion of space-filling curve to linearize the data set, and it consists of two phases. The first phase provides an approximate solution, within a rough factor, after the execution of at most d + 1 sorts and scans of the data set, with temporal cost quadratic in d and linear in N and in k, where d is the number of dimensions of the data set and N is the number of points in the data set. During this phase, the algorithm isolates points candidate to be outliers and reduces this set at each iteration. If the size of this set becomes n, then the algorithm stops reporting the exact solution. The second phase calculates the exact solution with a final scan examining further the candidate outliers that remained after the first phase. Experimental results show that the algorithm always stops, reporting the exact solution, during the first phase after much less than d + 1 steps. We present both an in-memory and disk-based implementation of the HilOut algorithm and a thorough scaling analysis for real and synthetic data sets showing that the algorithm scales well in both cases.     

基于分类超平面的非线性集成学习机

针对支持向量机面临的大规模数据分类问题, 提出基于分类超平面的非线性集成学习机NALM。该方法借鉴管理学中协同管理的思想, 将大规模数据分成规模较小的子集, 然后分别在子集上运行分类超平面算法, 最后将各子集上的分类结果进行非线性集成得到最终的分类结果。该方法不仅继承了分类超平面的优点, 而且还将分类超平面的适用范围从小规模数据扩展到中大规模数据, 从线性空间推广到Hilbert核空间。若干数据集上的实验表明：NALM能以较少的支持向量来解决大规模样本分类问题。     

On the number of multilinear partitions and the computing capacity of multiple-valued multiple-threshold perceptrons

We introduce the concept of multilinear partition of a point set V/spl sub/R/sup n/ and the concept of multilinear separability of a function f:Vtwo head right arrowK={0,...,k-1}. Based on well-known relationships between linear partitions and minimal pairs, we derive formulae for the number of multilinear partitions of a point set in general position and of the set K(2). The (n,k,s)-perceptrons partition the input space V into s+1 regions with s parallel hyperplanes. We obtain results on the capacity of a single (n,k,s)-perceptron, respectively, for V subset R(n) in general position and for V=K(2). Finally, we describe a fast polynomial-time algorithm for counting the multilinear partitions of K(2).     

Alternating hashing for expansible files

Proposes a generalized approach for designing a class of dynamic hashing schemes which require no index and have the growth of a file at a rate of (n+1)/n per full expansion, where n is the number of pages of the file, as compared to a rate of 2 in linear hashing. Based on this generalized approach, we derive a new dynamic hashing scheme called alternating hashing, in which, when a split occurs in page k, the data records in page k are redistributed to page k and page (k+1), or to page k and page (k-1), according to whether the value of level d is even or odd, respectively (d is defined as the number of full expansions that have happened so far). From our performance analysis, given a fixed load control, the proposed scheme can achieve nearly 97% storage utilization, as compared to 78% by using linear hashing     

Bipanconnectivity and Bipancyclicity in k-ary n-cubes

In this paper we give precise solutions to problems posed by Wang, An, Pan, Wang and Qu and by Hsieh, Lin and Huang. In particular, we show that Qnk is bipanconnected and edge-bipancyclic, when k ≥ 3 and n ≥ 2, and we also show that when k is odd, Qnk is m-panconnected, for m=(n(k-1)+2k-6)/2, and (k-1)-pancyclic (these bounds are optimal). We introduce a path-shortening technique, called progressive shortening, and strengthen existing results, showing that when paths are formed using progressive shortening then these paths can be efficiently constructed and used to solve a problem relating to the distributed simulation of linear arrays and cycles in a parallel machine whose interconnection network is Qnk, even in the presence of a faulty processor.     

Fast Algorithms for Computing the Smallest k-Enclosing Circle

For a set $P$ of $n$ points in the plane and an integer $k \leq n$, consider the problem of finding the smallest circle enclosing at least $k$ points of $P$. We present a randomized algorithm that computes in $O( n k )$ expected time such a circle, improving over previously known algorithms. Further, we present a linear time $\delta$-approximation algorithm that outputs a circle that contains at least $k$ points of $P$ and has radius less than $(1+\delta)r_{opt}(P,k)$, where $r_{opt}(P,k)$ is the radius of the minimum circle containing at least $k$ points of $P$. The expected running time of this approximation algorithm is $O(n + n \cdot\min((1/k\delta^3) \log^2 (1/\delta), k))$.     

K带实时图灵机计算能力与图灵机带数的关系

本文证明了对任意整数k,至少存在一个语言能被k带实时图灵机接受,但不能被(k—1)带实时图灵机所接受,从而证明了k带图灵机计算能力严格强于(k-1)带实时图灵机。     

Minimizing Makespan in Batch Machine Scheduling

We study the scheduling of a set of n jobs, each characterized by a release (arrival) time and a processing time, for a batch processing machine capable of running at most B jobs at a time. We obtain an O(n log n)-time algorithm when B is unbounded. When there are only m distinct release times and the inputs are integers, we obtain an O(n(BR_max)^m-1(2/m)^m-3)-time algorithm where R_max is the difference between the maximum and minimum release times. When there are k distinct processing times and m release times, we obtain an O(n log m + k^k+2 B^k+1 m² log m)-time algorithm. We obtain even better algorithms for m=2 and for k=1. These algorithms improve most of the corresponding previous algorithms for the respective special cases and lead to improved approximation schemes for the general problem.     

Fuzzy kernel perceptron

A new learning method, the fuzzy kernel perceptron (FKP), in which the fuzzy perceptron (FP) and the Mercer kernels are incorporated, is proposed in this paper. The proposed method first maps the input data into a high-dimensional feature space using some implicit mapping functions. Then, the FP is adopted to find a linear separating hyperplane in the high-dimensional feature space. Compared with the FP, the FKP is more suitable for solving the linearly nonseparable problems. In addition, it is also more efficient than the kernel perceptron (KP). Experimental results show that the FKP has better classification performance than FP, KP, and the support vector machine.     

Adaptive probabilistic neural networks for pattern classification in time-varying environment

In this paper, we propose a new class of probabilistic neural networks (PNNs) working in nonstationary environment. The novelty is summarized as follows: 1) We formulate the problem of pattern classification in nonstationary environment as the prediction problem and design a probabilistic neural network to classify patterns having time-varying probability distributions. We note that the problem of pattern classification in the nonstationary case is closely connected with the problem of prediction because on the basis of a learning sequence of the length n, a pattern in the moment n+k, k/spl ges/1 should be classified. 2) We present, for the first time in literature, definitions of optimality of PNNs in time-varying environment. Moreover, we prove that our PNNs asymptotically approach the Bayes-optimal (time-varying) decision surface. 3) We investigate the speed of convergence of constructed PNNs. 4) We design in detail PNNs based on Parzen kernels and multivariate Hermite series.     

基于分类间隔的特征选择算法

对于二类目标特征选择问题,首先讨论了特征空间的线性可分性问题,并给出了其判别条件;其次,通过借鉴支撑矢量机原理,分析了特征可分性判据的基本性质;最后,依据各特征对分类间隔的贡献大小定义了特征有效率,并以此进行特征选择和特征空间降维.实测数据与网络公开UCI(University of california,Irvine)数据库的实验结果表明,与经典的Relief特征选择算法相比,该算法在识别性能和推广能力上明显有所提高.     

An energy-efficient permutation routing protocol for single-hop radio networks

A radio network (RN) is a distributed system where each station or node is a small hand-held commodity device called a station. Typically, each station has access to a few channels for transmitting and receiving messages. By RN(p, k), we denote a radio network with p stations, where each station has access to k channels. In a single-hop RN, every station is within the transmission range of every other station. Each station consumes power while transmitting or receiving a message, even when it receives a message that is not destined for it. It is extremely important that the stations consume power only when it is necessary since it is not possible to recharge batteries when the stations are on a mission. We are interested in designing an energy-efficient protocol for permutation routing, which is one of the most fundamental problems in any distributed system. An instance of the permutation routing problem involves p stations of an RN, each storing n/p items. Each item has a unique destination address which is the identity of the destination station to which the item should be sent. The goal is to route all the items to their destinations while consuming as little energy as possible. We show that the permutation routing problem of n packets on an RN(p, k) can be solved in 2n/k+(p/k)/sup 2/+p+2k/sup 2/ slots and each station needs to be awake for at most 6n/p+2p/k+8k slots. When k/spl Lt/p/spl Lt/n, our protocol is more efficient, both in terms of total number of slots and the number of slots each station is awake compared to a previously published protocol by Nakano et al. (2001).     

基于特征可分性的核函数自适应构造

核函数的选择与构造是支撑向量机研究中的关键问题和难点.该文针对该问题,首先讨论了特征空间的线性可分性,推导了其判别条件.然后,根据特征完全可分条件,基于函数逼近论和核函数的基本性质,提出了自适应多项式核函数和B-样条核函数模型,给出了模型参数的估计算法.实测数据仿真实验结果表明,与经典的核函数相比,该文提出的算法在分类性能上取得了明显改善.