Short-term memory circuit using hardware ring neural networks

A number of studies have recently been made on various neuron models and neural networks. This research is studied for applications to engineering problems and an understanding of the information processing functions of living organisms. We are studying an asynchronous neural network using a pulse-type hardware neuron model (P-HNM). Recently, we have been trying to construct a short-term memory circuit using hardware ring neural networks (RNN) with P-HNM. In this article, we discuss the construction of a short-term memory circuit using the hardware RNN, and conduct experiments that explain the characteristics of the network through circuit simulation using PSpice. As a result, we verify that the RNN which is proposed in this article can be used as the short-term memory circuit.This work was presented, in part, at the 9th International Symposium on Artificial Life and Robotics, Oita, Japan, January 28–30, 2004     

一种基于K-Means局部最优性的高效聚类算法

K-Means聚类算法只能保证收敛到局部最优,从而导致聚类结果对初始代表点的选择非常敏感.许多研究工作都着力于降低这种敏感性.然而,K-Means的局部最优和结果敏感性却构成了K-MeanSCAN聚类算法的基础.K-MeanSCAN算法对数据集进行多次采样和K-Means预聚类以产生多组不同的聚类结果,来自不同聚类结果的子簇之间必然会存在交集.算法的核心思想是,利用这些交集构造出关于子簇的加权连通图,并根据连通性合并子簇.理论和实验证明,K-MeanScan算法可以在很大程度上提高聚类结果的质量和算法的效率.     

H∞ model reduction for discrete-time singular systems

This paper investigates the problem of H_∞ model reduction for linear discrete-time singular systems. Without decomposing the original system matrices, necessary and sufficient conditions for the solvability of this problem are obtained in terms of linear matrix inequalities (LMIs) and a coupling non-convex rank constraint set. When these conditions are feasible, an explicit parametrization of the desired reduced-order models is given. Particularly, a simple LMI condition without rank constraint is derived for the zeroth-order H_∞ approximation problem. Finally, an illustrative example is provided to demonstrate the applicability of the proposed approach.     

A binary neural k-nearest neighbour technique

K-Nearest Neighbour (k-NN) is a widely used technique for classifying and clustering data. K-NN is effective but is often criticised for its polynomial run-time growth as k-NN calculates the distance to every other record in the data set for each record in turn. This paper evaluates a novel k-NN classifier with linear growth and faster run-time built from binary neural networks. The binary neural approach uses robust encoding to map standard ordinal, categorical and real-valued data sets onto a binary neural network. The binary neural network uses high speed pattern matching to recall the k-best matches. We compare various configurations of the binary approach to a conventional approach for memory overheads, training speed, retrieval speed and retrieval accuracy. We demonstrate the superior performance with respect to speed and memory requirements of the binary approach compared to the standard approach and we pinpoint the optimal configurations.     

一种半监督K均值多关系数据聚类算法

提出了一种半监督K均值多关系数据聚类算法.该算法在K均值聚类算法的基础上扩展了其初始类簇的选择方法和对象相似性度量方法,以用于多关系数据的半监督学习.为了获取高性能,该算法在聚类过程中充分利用了标记数据、对象属性及各种关系信息.多关系数据库Movie上的实验结果验证了该算法的有效性.     

一种无线传感器网络中的多维K-NN查询优化算法

提出了一种基于过滤器的无线传感器网络多维K-NN查询优化算法PREDICTOR.过滤器是设置在节点端的取值分布区间,用来屏蔽节点发送属于区间内的数据,从而节省节点能耗.在服务器端保存有各节点的历史样本数据,根据K-NN查询请求和样本数据的分布范围为节点定义过滤器.提出了3种优化策略:(1) 过滤器覆盖区间大小分配策略的动态调整方法,使得进入最终查询结果可能性小的节点拥有较大的覆盖区间;(2) 节点间过滤器共享方法,使得历史样本数据相近的节点使用相同的过滤器;(3) 过滤器压缩传输方法,减少为不同K-NN查询更新过滤器的代价.通过实验评价,验证了PREDICTOR算法的能量有效性,与朴素算法相比,极大地降低了数据传输量.     

基于K近邻和优化分配策略的密度峰值聚类算法

密度峰值聚类(density peak clustering, DPC)是一种简单有效的聚类分析方法.但在实际应用中,对于簇间密度差别大或者簇中存在多密度峰的数据集,DPC很难选择正确的簇中心;同时,DPC中点的分配方法存在多米诺骨牌效应.针对这些问题,提出一种基于K近邻(K-nearest neighbors,KNN)和优化分配策略的密度峰值聚类算法.首先,基于KNN、点的局部密度和边界点确定候选簇中心;定义路径距离以反映候选簇中心之间的相似度,基于路径距离提出密度因子和距离因子来量化候选簇中心作为簇中心的可能性,确定簇中心.然后,为了提升点的分配的准确性,依据共享近邻、高密度最近邻、密度差值和KNN之间距离构建相似度,并给出邻域、相似集和相似域等概念,以协助点的分配;根据相似域和边界点确定初始聚类结果,并基于簇中心获得中间聚类结果.最后,依据中间聚类结果和相似集,从簇中心到簇边界将簇划分为多层,分别设计点的分配策略;对于具体层次中的点,基于相似域和积极域提出积极值以确定点的分配顺序,将点分配给其积极域中占主导地位的簇,获得最终聚类结果.在11个合成数据集和27个真实数据集上进行仿真...     

基于熵加权K-means全局信息聚类的高光谱图像分类

目的 高光谱图像波段数目巨大,导致在解译及分类过程中出现“维数灾难”的现象。针对该问题,在K-means聚类算法基础上,考虑各个波段对不同聚类的重要程度,同时顾及类间信息,提出一种基于熵加权K-means全局信息聚类的高光谱图像分类算法。方法 首先,引入波段权重,用来刻画各个波段对不同聚类的重要程度,并定义熵信息测度表达该权重。其次,为避免局部最优聚类,引入类间距离测度实现全局最优聚类。最后,将上述两类测度引入K-means聚类目标函数,通过最小化目标函数得到最优分类结果。结果 为了验证提出的高光谱图像分类方法的有效性,对Salinas高光谱图像和Pavia University高光谱图像标准图中的地物类别根据其光谱反射率差异程度进行合并,将合并后的标准图作为新的标准分类图。分别采用本文算法和传统K-means算法对Salinas高光谱图像和Pavia University高光谱图像进行实验,并定性、定量地评价和分析了实验结果。对于图像中合并后的地物类别,光谱反射率差异程度大,从视觉上看,本文算法较传统K-means算法有更好的分类结果;从分类精度看,本文算法的总精度分别为92.20%和82.96%, K-means算法的总精度分别为83.39%和67.06%,较K-means算法增长8.81%和15.9%。结论 提出一种基于熵加权K-means全局信息聚类的高光谱图像分类算法,实验结果表明,本文算法对高光谱图像中具有不同光谱反射率差异程度的各类地物目标均能取得很好的分类结果。     

保护位置隐私和查询内容隐私的路网K近邻查询方法

位置隐私和查询内容隐私是LBS兴趣点(point of interest,简称POI)查询服务中需要保护的两个重要内容,同时,在路网连续查询过程中,位置频繁变化会给LBS服务器带来巨大的查询处理负担,如何在保护用户隐私的同时,高效地获取精确查询结果,是目前研究的难题.以私有信息检索中除用户自身外其他实体均不可信的思想为基本假设,基于Paillier密码系统的同态特性,提出了无需用户提供真实位置及查询内容的K近邻兴趣点查询方法,实现了对用户位置、查询内容隐私的保护及兴趣点的精确检索;同时,以路网顶点为生成元组织兴趣点分布信息,进一步解决了高强度密码方案在路网连续查询中因用户位置变化频繁导致的实用效率低的问题,减少了用户的查询次数,并能确保查询结果的准确性.最后从准确性、安全性及查询效率方面对本方法进行了分析,并通过仿真实验验证了理论分析结果的正确性.     

基于Temporal rule的忆阻神经网络电路

忆阻器是一种动态特性的电阻,其阻值可以根据外场的变化而变化,并且在外场撤掉后能够保持原来的阻值,具有类似于生物神经突触连接强度的特性,可以用来存储突触权值。在此基础上,为了实现基于Temporal rule对IRIS数据集识别学习的功能,建立了以桥式忆阻器为突触的神经网络SPICE仿真电路。采用单个脉冲的编码方式,脉冲的时刻代表着数据信息,该神经网络电路由48个脉冲输入端口、144个突触、3个输出端口组成。基于Temporal rule学习规则对突触的权值修改,通过仿真该神经网络电路对IRIS数据集的分类正确率最高能达到93.33%,表明了此神经系统结构设计在类脑脉冲神经网络中的可用性。     

Robust H∞ control for linear discrete-time systems with norm-bounded time-varying uncertainty

The problem of robust H_∞ analysis and synthesis for linear discrete-time systems with norm-bounded time-varying uncertainty is studied in this paper. It will be shown that this problem is equivalent to the problem of H_∞ analysis and synthesis of an auxiliary system. The necessary and sufficient conditions for the equivalency are proved. Thus the original problem can be solved by existing H_∞ control methods.     

Similarity K-d tree method for sparse point pattern matching with underlying non-rigidity

We propose a method for matching non-affinely related sparse model and data point-sets of identical cardinality, similar spatial distribution and orientation. To establish a one-to-one match, we introduce a new similarity K-dimensional tree. We construct the tree for the model set using spatial sparsity priority order. A corresponding tree for the data set is then constructed, following the sparsity information embedded in the model tree. A matching sequence between the two point sets is generated by traversing the identically structured trees. Experiments on synthetic and real data confirm that this method is applicable to robust spatial matching of sparse point-sets under moderate non-rigid distortion and arbitrary scaling, thus contributing to non-rigid point-pattern matching.     

K-means聚类方法下的复数图像群组稀疏编码降噪算法

稀疏编码已经广泛应用于复数图像的降噪问题,其中,近些年提出的分组稀疏编码由于能够充分利用同一分组图像块的相似性,在滤除噪声和提高降噪信噪比方面具有更大的优势.研究了一种基于K-means聚类方法的复数图像分组稀疏降噪算法,通过改进聚类算法,验证了K-means算法对分组稀疏编码算法的分组有效性.采用在线复数词典训练算法快速获取编码字典,并运用分组正交匹配追踪算法,实现了分组图像块的稀疏编码.通过限制每一分组图像块中编码的相似性,有效抑制了对图像块中噪声的编码,提高了对复数图像的降噪效果.为验证算法的有效性,对模拟和真实的干涉合成孔径雷达图像的仿真噪声进行了定量分析,证明了所提算法相对于以前的分组稀疏编码算法在峰值信噪比指标上有一定的提升.最后对真实的干涉合成孔径雷达图像进行了降噪,进一步验证了所提降噪算法对于真实噪声的降噪能力.     

Quadratic stabilisation with H∞-norm bound of non-linear discrete-time uncertain systems with bounded control

The paper addresses the problem of quadratic stabilisability with H_∞-norm bound of uncertain discrete-time control-affine systems by norm-bounded controls. Both structured parameter uncertainties and unstructured exogenous disturbances are taken into account. The given definition of quadratic stabilisability is a generalisation of that used for linear systems so far. A necessary condition of the stabilisability is formulated. A state feedback control satisfying an a priori constraint is proposed for the solution of the formulated H_∞ problem. The proposed method may be applicable even in such cases when the linearisation technique cannot be used.     

Efference copies in neural control of dynamic biped walking

In the early 1950s, von Holst and Mittelstaedt proposed that motor commands copied within the central nervous system (efference copy) help to distinguish ‘reafference’ activity (afference activity due to self-generated motion) from ‘exafference’ activity (afference activity due to external stimulus). In addition, an efference copy can be also used to compare it with the actual sensory feedback in order to suppress self-generated sensations. Based on these biological findings, we conduct here two experimental studies on our biped “RunBot” where such principles together with neural forward models are applied to RunBot’s dynamic locomotion control. The main purpose of this article is to present the modular design of RunBot’s control architecture and discuss how the inherent dynamic properties of the different modules lead to the required signal processing. We believe that the experimental studies pursued here will sharpen our understanding of how the efference copies influence dynamic locomotion control to the benefit of modern neural control strategies in robots.     

An LMI approach to periodic discrete-time unbiased filtering

The problem of unbiased filtering for a discrete-time linear periodic system is faced by means of linear matrix inequality techniques. As leading case, we derive the synthesis conditions to obtain an unbiased filter and an unbiased fixed-lag smoother enforcing a bound on the H_∞ performance on the error dynamics.     

Analysing the localisation sites of proteins through neural networks ensembles

Scientists involved in the area of proteomics are currently seeking integrated, customised and validated research solutions to better expedite their work in proteomics analyses and drug discoveries. Some drugs and most of their cell targets are proteins, because proteins dictate biological phenotype. In this context, the automated analysis of protein localisation is more complex than the automated analysis of DNA sequences; nevertheless the benefits to be derived are of same or greater importance. In order to accomplish this target, the right choice of the kind of the methods for these applications, especially when the data set is drastically imbalanced, is very important and crucial. In this paper we investigate the performance of some commonly used classifiers, such as the K nearest neighbours and feed-forward neural networks with and without cross-validation, in a class of imbalanced problems from the bioinformatics domain. Furthermore, we construct ensemble-based schemes using the notion of diversity, and we empirically test their performance on the same problems. The experimental results favour the generation of neural network ensembles as these are able to produce good generalisation ability and significant improvement compared to other single classifier methods.     

基于图压缩的k可达查询处理

研究了基于图压缩的k可达查询处理,提出了一种支持k可达查询的图压缩算法k-RPC及无需解压缩的查询处理算法,k-RPC算法在所有基于等价类的支持k-reach查询的图压缩算法中是最优的.由于k-RPC算法是基于严格的等价关系,因此进一步又提出了线性时间的近似图压缩算法k-GRPC.k-GRPC算法允许从原始图中删除部分边,然后使用k-RPC获得更好的压缩比.提出了线性时间的无需解压缩的查询处理算法.真实数据上的实验结果表明,对于稀疏的原始图,两种压缩算法的压缩比分别可以达到45%,对于稠密的原始图,两种压缩算法的压缩比分别可以达到75%和67%;与在原始图上直接进行查询处理相比,两种基于压缩图的查询处理算法效率更好,在稀疏图上的查询效率可以提高2.5倍.     

结合调整差值变换的(K,N)有意义图像分存方案

目的 针对传统有意义分存方法存在的像素扩张和分发掩体图像视觉质量不高等问题,提出一种结合调整差值变换的(K,N)有意义图像分存方案,该方案可用于与掩体等大且同为自然图像的密图分存.方法 在分存阶段,首先用调整差值变换将密图转换为差值图和位置图;其次将差值图和位置图进行(K,N)分存,分别嵌入到掩体图像中,并使用密钥确定位置图分存信息的嵌入位置和根据位置图中不同的差值类型选择不同的差值嵌入方法;再次对密钥进行(K,N)分存,将子密钥和分发掩体对应的MD5值公布到第3方公信方作为认证码;最后将子密钥和分发掩体分发给参与者进行保管.在恢复阶段,首先核对参与者的子密钥和分发掩体对应的MD5值,若认证通过的人数小于K,则恢复失败;否则,使用认证通过的子密钥还原出密钥,然后根据密钥提取并恢复出位置图;其次根据位置图中的差值类型来提取和恢复出差值图;最后使用逆调整差值变换还原出最终密图.结果 同现有方法相比,所提策略不存在像素扩张且分发掩体图像视觉质量较高,具有较强的恶意参与者检测能力.结论 本文方法的掩体图像与密图等大且同为自然图像,同经典有意义图像分存方案相比,克服了像素扩张问题,嵌入信息后的掩体图像具有较高的视觉质量,使用第3方公信方存储的MD5值作为认证码,具有较强的恶意参与者识别能力.