Algorithmic Fusion for More Robust Feature Tracking

We present a framework for merging the results of independent feature-based motion trackers using a classification based approach. We demonstrate the efficacy of the framework using corner trackers as an example. The major problem with such systems is generating ground truth data for training. We show how synthetic data can be used effectively to overcome this problem. Our combined system performs better in both dropouts and errors than a correspondence tracker, and had less than half the dropouts at the cost of moderate increase in error compared to a relaxation tracker.     

一种基于算法机制设计的社会法则合成方法

社会法则是在多Agent系统中为确立某种目标属性而对各个Agent实施的行为限制集合.在Agent具有“个体理性”及“私有信息”的“策略情况”下,社会法则合成问题不应建模成通常的优化问题,而应建模成算法机制设计问题.“最小化副作用”经常是社会法则需要满足的基本要求.从博弈论的角度来看,“最小化副作用”与“最大化社会福利”的概念紧密相关,可以将“最小化副作用的社会法则合成”建模为一种效率机制设计问题.不仅需要为给定目标属性找到有效且社会福利最大的社会法则,还需要向Agent支付适当的金额,以实现激励相容性和个体理性.首先基于VCG机制设计一种名叫VCG-SLM的效率机制,证明它可满足所有必需的形式属性.然而,由于发现可证明该机制的计算是一个FP^NP-完全问题,针对性地提出该机制的一种基于整数规划的实现方式VCG-SLM-ILP,基于ATL语义将分配及支付的计算转化为整数规划,并严格地证明其正确性,从而可有效利用目前已非常成熟的工业级整数规划求解器,成功解决棘手的机制计算问题.     

Algorithmic Techniques for Maintaining Shortest Routes in Dynamic Networks

In this paper, we survey algorithms for shortest paths in dynamic networks. Although research on this problem spans over more than three decades, in the last couple of years many novel algorithmic techniques have been proposed. In this survey, we will make a special effort to abstract some combinatorial and algebraic properties, and some common data-structural tools that are at the base of those techniques. This will help us try to present some of the newest results in a unifying framework so that they can be better understood and deployed also by non-specialists.     

位移传感器特性区域直线拟合算法分析

在标定位移传感器静态特性时,原始测量数据通常为非线性分布,应选取其中能表征传感器线性特性的区域作为传感器的工作量程.为提高位移传感器特性区域的准确性和可靠性,本文提出一种识别特性区域的计算方法,利用迭代搜索法和最小二乘法拟合直线,求取特性区域的起始点和终止点.这样不仅可以获得位移传感器的最大工作量程,而且为提高传感器静态标定精度提供了保障.计算实例表明了该算法的实用性.     

Heterogeneous Similarity Learning for More Practical Kinship Verification

Neural Processing Letters - Kinship verification via facial images is a relatively new and challenging problem in computer vision. Prior studies in the literature have focused solely on...     

浮点加法器电路设计算法的研究

介绍了浮点加法器电路设计的基本算法,阐述了近年来有关浮点加法器电路设计算法研究的成果。对目前所普遍采用的Ｔｗｏ－Ｐａｔｈ算法及其改进算法进行了详细地分析。描述了快速规格化的关键技术——前导１的预判的基本原理。最后提出了一种进一步改进Ｔｗｏ—Ｐａｔｈ算法的方案。     

基于分布式集群的线性方程组求解

在很多应用领域中都需要求解大规模的线性方程组,然而由于单机和多处理器系统的计算能力有限,这个问题没有得到很好地解决。该文提出了一种求解这类方程组的并行算法及其在Linux集群上的实现。在介绍了该算法的矩阵分块、通信机制和改进的高斯消元法之后,描述了该算法的实现流程,并进一步对其通信开销、计算开销和存储需求进行了分析。     

稀疏线性方程组求解中的预处理技术综述

稀疏线性方程组的高效求解是数值计算方向的研究热点之一,其中包括预处理技术的研究。本文从技术分类的角度,总结了稀疏线性方程组求解中的预处理技术。首先,介绍了填充元缩减策略,旨在减少求解过程中存储量的同时,仍能保持矩阵的稀疏结构;其次,介绍了不同结构系数矩阵的多种匹配技术,旨在获得矩阵的对角优势性;最后,介绍了具有天然并行性的因子分解近似逆预条件子构造方法和不完全分解预条件中的并行求解技术等。     

线性查询的一种近似最优差分隐私机制

在差分隐私保护程度确定的条件下使数据的有用性最大化的问题称为差分隐私的最优机制问题.最优机制问题是差分隐私理论中的一个重要问题,与差分隐私模型的理论基础及应用前景有直接联系.与已有的研究不同,本文提出一种新的不基于敏感度的分析方法,来寻找最优机制.首先,本文将最优机制问题构造为一个多目标函数优化问题并提出了一种新的差分隐私机制构造方法.在此基础上,本文对线性查询问题给出了一种近似最优差分隐私机制（定理2）,该机制达到了差分隐私不等式的边界.此外,本文的大部分分析方法也可对非线性查询的最优机制问题进行分析.本文的研究揭示了敏感度方法的不足之处,发现其无法刻画数据集的邻居集合对应的查询函数值集合的特性,而该集合包含了差分隐私的一些深层特征.     

Ellipsoidal Techniques for Reachability Analysis of Discrete-Time Linear Systems

This paper describes the computation of reach sets for discrete-time linear control systems with time-varying coefficients and ellipsoidal bounds on the controls and initial conditions. The algorithms construct external and internal ellipsoidal approximations that touch the reach set boundary from outside and from inside. Recurrence relations describe the time evolution of these approximations. An essential part of the paper deals with singular discrete-time linear systems     

Towards an Efficient Path-Oriented Tool for Bounded Reachability Analysis of Linear Hybrid Systems using Linear Programming

The existing techniques for reachability analysis of linear hybrid automata do not scale well to problem sizes of practical interest. Instead of developing a tool to perform reachability check on all the paths of a linear hybrid automaton, a complementary approach is to develop an efficient path-oriented tool to check one path at a time where the length of the path being checked can be made very large and the size of the automaton can be made large enough to handle problems of practical interest. This approach of symbolic execution of paths can be used by design engineers to check important paths and thereby, increase the faith in the correctness of the system. Unlike simple testing, each path in our framework represents a dense set of possible trajectories of the system being analyzed. In this paper, we develop the linear programming based techniques towards an efficient path-oriented tool for the bounded reachability analysis of linear hybrid systems.     

调整一类部分线性混沌系统输出的实用方案

利用部分线性混沌系统耦合后内在的投影同步性质对耦合响应系统的状态输出进行反馈控制，得到与驱动系统相位一致(或反相)，振幅成比例的混沌输出信号。该方案所需控制代价较小，为设计任意功率输出的混沌信号源提供了理论依据，在利用混沌电路系统进行保密通信方面有着极大的应用前景；仿真实例进一步表明了控制方案的有效性。     

ENHANCING THE CREATIVITY OF REENGINEERING Techniques for Making IS More Creative

Industries in the US continue to lose ground in international competition. If IS is to help change that trend, it will have to become more creative. Creativity improvement programs are being introduced in other areas of the company— now is the time for IS organizations to follow suit.     

图形处理器低功耗设计技术研究

图形处理器(GPU)以其强大的图形加速性能以及在通用计算领域的出色表现正在被越来越广泛地应用。但随着芯片规模和集成度的不断提升,单个GPU芯片的功耗已经高达376W,是高端通用处理器的2～3倍。高功耗带来的可靠性、稳定性以及芯片成本问题使“功耗墙”已经成为未来GPU设计过程中需要突破的关键问题之一。立足于体系结构层次,结合图形处理器的渲染流水线的结构特点,从深度测试和消隐、染色器数据通路、纹理映射和压缩、渲染策略、寄存器文件和片上Cache等角度描述了图形处理器的低功耗设计技术,并指出了GPU低功耗设计技术的进一步研究方向。     

一类不确定线性系统的鲁棒线性控制器设计

针对一类不确定性不满足匹配条件的线性系统,利用Lyapunov方程和不确定项的范数界,分别设计了具有可调参数和鲁棒线性状态和输出控制器。这些可调参数可以依据不确定项的范数界的大小来选取,具有一定的灵活性。     

A First Step Towards Adaptive Control for Linear Systems in Max Algebra

Alinear system theory has been developed for discrete event systemssubject to synchronization. We are interested in the just intime control of such systems in presence of a possible mismatchbetween the system and its model. Taking such mismatch into accountin controller synthesis prevents us from using the usual openloop control structure. The approach we propose to synthesizethe controller is inspired by the principle of the indirect adaptivecontrol well known in conventional system theory. Its designleads to solve an optimal tracking problem under an equalityconstraint. Preliminary results about the estimation of modelsare presented.     

嵌入式软件实用设计模式

针对嵌入式软件的特点,介绍代理、适配器、中介者、观察者4种实用的设计模式及使用场景,并利用UML模型解析实现方法。     

A More Practical Approach for Single-Packet IP Traceback using Packet Logging and Marking

Tracing IP packets to their origins is an important step in defending Internet against denial-of-service attacks. Two kinds of IP traceback techniques have been proposed as packet marking and packet logging. In packet marking, routers probabilistically write their identification information into forwarded packets. This approach incurs little overhead but requires large flow of packets to collect the complete path information. In packet logging, routers record digests of the forwarded packets. This approach makes it possible to trace a single packet and is considered more powerful. At routers forwarding large volume of traffic, the high storage overhead and access time requirement for recording packet digests introduce practicality problems. In this paper, we present a novel scheme to improve the practicality of log-based IP traceback by reducing its overhead on routers. Our approach makes an intelligent use of packet marking to improve scalability of log-based IP traceback. We use mathematical analysis and simulations to evaluate our approach. Our evaluation results show that, compared to the state-of-the-art log-based approach called hash-based IP traceback, our approach maintains the ability to trace single IP packet while reducing the storage overhead by half and the access time overhead by a factor of the number of neighboring routers.