The potential structure of sample paths and performance sensitivities of Markov systems

We study the structure of sample paths of Markov systems by using performance potentials as the fundamental units. With a sample path-based approach, we show that performance sensitivity formulas (performance gradients and performance differences) of Markov systems can be constructed intuitively, by first principles, with performance potentials (or equivalently, perturbation realization factors) as building blocks. In particular, we derive sensitivity formulas for two Markov chains with possibly different state spaces. The proposed approach can be used to obtain flexibly the sensitivity formulas for a wide range of problems, including those with partial information. These formulas are the basis for performance optimization of discrete event dynamic systems, including perturbation analysis, Markov decision processes, and reinforcement learning. The approach thus provides insight on on-line learning and performance optimization and opens up new research directions. Sample path based algorithms can be developed.     

一种事件序列的加权变阶马尔可夫模型

变阶马尔可夫模型是对事件序列建模的一种简单且有效的模型,但经典变阶马尔可夫模型只考虑转移概率,未关注子序列本身出现的频率。为此,提出一种加权的变阶马尔可夫模型,在经典变阶马尔可夫模型基础上根据子序列的频率构建一棵加权概率后缀树。给出一种剪枝策略,在构建后缀树时根据结点相似程度剪除树枝,以提高模型的泛化能力,并在线性时间内完成加权概率后缀树的构建。通过将加权的模型应用于事件序列分类进行实验验证,结果表明,该模型可以对不同领域的实际序列数据进行有效分类。     

符号序列多阶Markov分类

针对基于固定阶Markov链模型的方法不能充分利用不同阶次子序列结构特征的问题,提出一种基于多阶Markov模型的符号序列贝叶斯分类新方法。首先,建立了基于多阶次Markov模型的条件概率分布模型;其次,提出一种附后缀表的n-阶子序列后缀树结构和高效的树构造算法,该算法能够在扫描一遍序列集过程中建立多阶条件概率模型;最后,提出符号序列的贝叶斯分类器,其训练算法基于最大似然法学习不同阶次模型的权重,分类算法使用各阶次的加权条件概率进行贝叶斯分类预测。在三个应用领域实际序列集上进行了系列实验,结果表明：新分类器对模型阶数变化不敏感;与使用固定阶模型的支持向量机等现有方法相比,所提方法在基因序列与语音序列上可以取得40%以上的分类精度提升,且可输出符号序列Markov模型最优阶数参考值。     

Markov离散事件动态系统参数灵敏度估计算法

给出了马尔可夫离散事件系统参数灵敏度估计的高效率仿真算法.既可以用于稳态性能测度的参数灵敏度估计,也可用于终止型性能测度的灵敏度估计.和现有的仿真算法相比,其数学描述和仿真流程比较简洁,易于编程实现.给出的各种仿真算例均验证了该方法的适用性.此外,还指出了马尔可夫系统性能测度参数灵敏度的精确估计为光滑扰动分析(Smoothed PerturbatinAnalysis)和似然比方法(Likelihood Ratios)得到的估计量之和,单独使用其中的任何一种均难以给出系统性能测度参数灵敏度的可靠、一致估计.     

Context-aware environments: from specification to implementation

Abstract: This paper deals with the problem of implementing a context model for a smart environment. The problem has already been addressed several times using many different data- or problem-driven methods. In order to separate the modelling phase from implementation, we first represent the context model by a network of situations. Then, different implementations can be automatically generated from this context model depending on user needs and underlying perceptual components. Two different implementations are proposed in this paper: a deterministic one based on Petri nets and a probabilistic one based on hidden Markov models. Both implementations are illustrated and applied to real-world problems.     

Design and implementation of a high‐performance CCA event service

Event services based on publish–subscribe architectures are well‐established components of distributed computing applications. Recently, an event service has been proposed as part of the common component architecture (CCA) for high‐performance computing (HPC) applications. In this paper we describe our implementation, experimental evaluation, and initial experience with a high‐performance CCA event service that exploits efficient communications mechanisms commonly used on HPC platforms. We describe the CCA event service model and briefly discuss the possible implementation strategies of the model. We then present the design and implementation of the event service using the aggregate remote memory copy interface as an underlying communication layer for this mechanism. Two alternative implementations are presented and evaluated on a Cray XD‐1 platform. The performance results demonstrate that event delivery latencies are low and that the event service is able to achieve high‐throughput levels. Finally, we describe the use of the event service in an application for high‐speed processing of data from a mass spectrometer and conclude by discussing some possible extensions to the event service for other HPC applications. Published in 2009 by John Wiley & Sons, Ltd.     

基于多时窗共空间模式的HMM运动想象脑电识别

运动想象脑电具有识别效果不佳及复杂时序信号建模困难的问题；提出一种基于多时窗共空间模式的隐马尔可夫模型运动想象脑电识别方法,首先将运动想象脑电划分为多个短时窗信号,然后使用共空间模式提取特征序列,以滤除脑电通道间的冗余信息,最后采用前向-后相算法与Viterbi算法求解隐马尔可夫模型并完成分类识别；将本文方法在公开运动想象脑电数据集上进行实验，得到77.17%的分类正确率，相较隐马尔可夫模型算法提升了5.74%，验证了所提方法的有效性。     

基于隐马尔科夫模型的胎盘植入产前诊断方法*

胎盘植入是产科严重的并发症之一,作为金标准的产后病理检验存在的滞后性和局限性问题,文中将病史和彩超数据等产前多特征关联作为观测显状态序列,将产后病理诊断作为隐状态,构建基于隐马尔科夫模型的胎盘植入产前诊断方法.采用Gini方法提取患病关联因素的特征集合,通过转化特征集合构建隐马尔科夫模型,结合Baum-Welch和Viterbi算法计算求解,通过显隐状态关系,实现胎盘植入产前诊断.实验表明,文中方法具有较好的准确率、特异度和灵敏度     

Performance analysis of computational approaches to solve Multiple Sequence Alignment

In the biotechnology field, the deployment of the Multiple Sequence Alignment (MSA) problem, which is a high performance computing demanding process, is one of the new challenges to address on the new parallel systems. The aim of this problem is to find similar regions on biological sequences. Furthermore, the goal of MSA applications is to align as much sequences as possible with a level of quality that makes the alignment biologically meaningful. An efficiency study of different MSA implementations, based on T-Coffee (one of the most used MSA aligners), has been performed in order to find new optimizations that may improve the average execution time on multi-core systems. We found that the current parallel implementations have some performance issues, affecting negatively the scalability of the process. Finally, the proposed implementation based on the usage of threads in conjunction with a message-passing library is presented, with the aim to optimize the execution of the MSA problem in multi-core-based clusters.     

A Guided Tour through some Extensions of the Event Calculus

Kowalski and Sergot's Event Calculus (EC) is a simple temporal formalism that, given a set of event occurrences, derives the maximal validity intervals (MVIs) over which properties initiated or terminated by these events hold. In this paper, we conduct a systematic analysis of EC by which we gain a better understanding of this formalism and determine ways of augmenting its expressive power. The keystone of this endeavor is the definition of an extendible formal specification of its functionalities. This formalization has the effects of casting determination of MVIs as a model checking problem, of setting the ground for studying and comparing the expressiveness and complexity of various extensions of EC, and of establishing a semantic reference against which to verify the soundness and completeness of implementations. We extend the range of queries accepted by EC, which is limited to Boolean combinations of MVI verification or computation requests, to support arbitrary quantification over events and modal queries. We also admit specifications based on preconditions. We demonstrate the added expressive power by encoding a number of diagnosis problems. Moreover, we provide a systematic comparison of the expressiveness and complexity of the various extended event calculi against each other. Finally, we propose a declarative encoding of these enriched event calculi in the logic programming language λProlog and prove the soundness and completeness of the resulting logic programs.     

Bayesian event detection for sport games with hidden Markov model

Event detection can be defined as the problem of detecting when a target event has occurred, from a given data sequence. Such an event detection problem can be found in many fields in science and engineering, such as signal processing, pattern recognition, and image processing. In recent years, many data sequences used in these fields, especially in video data analysis, tend to be high dimensional. In this paper, we propose a novel event detection method for high-dimensional data sequences in soccer video analysis. The proposed method assumes a Bayesian hidden Markov model with hyperparameter learning in addition to the parameter leaning. This is in an attempt to reduce undesired influences from ineffective components within the high-dimensional data. Implemention is performed by Markov Chain Monte Carlo. The proposed method was tested against an event detection problem with sequences of 40-dimensional feature values extracted from real professional soccer games. The algorithm appears functional.     

Distributed Markov Chain Monte Carlo kernel based particle filtering for object tracking

Particle filters are computationally intensive and thus efficient parallelism is crucial to effective implementations, especially object tracking in video sequences. Two schemes for pipelining particles under high performance computing environment, including an alternative Markov Chain Monte Carlo (MCMC) resampling algorithm and kernel function, are proposed so as to improve tracking performance and minimize execution time. Experimental results on a network of workstations composed of simple off-the-shelf hardware components show that global parallelizable scheme provides a promising resolution to clearly reduce execution time with increasing particles, compared with generic particle filtering.     

基于Markov 逻辑网的事件表象统一方法

为解决Web数据集成中大量事件表象语句共指现实世界同一事件,导致数据冗余问题,提出一种基于Markov逻辑网的事件表象统一方法。方法从共指事件表象集合中获得较准确详细的一条表象,作为统一的事件表象对应现实事件,为数据集成提供高质量数据。本文将事件表象使用8个维度的形式表示,训练Markov逻辑网从共指事件表象集合中推理出准确详细的维度内容,重新组合后形成一条事件表象。本文使用少量一阶谓词从维度内容、事件表象和数据源等多角度制定相应规则,通过推理解决数据不一致、不完整、不详细问题。实验结果表明基于Markov逻辑网的事件表象统一方法能获得较准确详细的统一事件表象。     

Prediction suffix trees for supervised classification of sequences

This paper presents a statistical test and algorithms for patterns extraction and supervised classification of sequential data. First it defines the notion of prediction suffix tree (PST). This type of tree can be used to efficiently describe variable order chain. It performs better than the Markov chain of order L and at a lower storage cost. We propose an improvement of this model, based on a statistical test. This test enables us to control the risk of encountering different patterns in the model of the sequence to classify and in the model of its class. Applications to biological sequences are presented to illustrate this procedure. We compare the results obtained with different models (Markov chain of order L, Variable order model and the statistical test, with or without smoothing). We set out to show how the choice of the parameters of the models influences performance in these applications. Obviously these algorithms can be used in other fields in which the data are naturally ordered.     

Performance vs. cost of redundant arrays of inexpensive disks

Data redundancy has been widely used to increase data availability in critical applications and several methods have been proposed to organize redundant data across a disk array. Data redundancy consists of either total data replication or the spreading of the data across the disk array along with parity information which can be used to recover missing data in the event of disk failure. In this paper we present an extended comparative analysis, carried out by using discrete event simulation models, between two disk array architectures: the Redundant Arrays of Inexpensive Disks (RAID) level 1 architecture, based on data replication; and the RAID level 5 architecture, based on the use of parity information. The comparison takes both performance and cost aspects into account. We study the performance of these architectures simulating two application environments characterized by different sizes of the data accessed by I/O operations. In addition, several scheduling policies for I/O requests are considered and the impact of non-uniform access to data on performance is investigated.     

基于CNN和MRF的运动目标分割

针对动态图像序列中背景成像过程因各种因素而变化存在复杂性,提出了一种基于细胞神经网络(CNN)和马尔可夫随机场(MRF)的目标分割方法.首先根据细胞神经网络与马尔可夫随机场能量函数的相似性,将马尔可夫随机场的最大后验概率模型映射到细胞神经网络近邻系统模型中.然后建立图像每一像素点的邻域系统模型,并且构造相应的能量函数.为使能量函数达到快速收敛,再利用模拟退火算法实现能量函数的最小值,以达到对运动目标的提取.由于CNN是由局部互连的细胞组成,因此易于用VLSI实现.实验的结果表明,该方法能够有效地抑制图像的噪声,对于运动目标的提取有较好的分割效果.     

True-concurrency probabilistic models: Markov nets and a law of large numbers

We introduce the model of Markov nets, a probabilistic extension of safe Petri nets under the true-concurrency semantics—this means that traces, not firing sequences, are given a probability. This model builds upon our previous work on probabilistic event structures. We use the notion of a branching cell for event structures, and show that the latter provides an adequate conception of local state for nets. We prove a Law of Large Numbers (LLN) for Markov nets, which constitutes the main contribution of the paper. This LLN allows for the characterization, in a quantitative way, of the asymptotic behavior of Markov nets.     

Implementation and modeling of two-phase locking concurrency control—a performance study

Two-phase locking (2PL) is the concurrency control mechanism that is used in most commercial database systems. In 2PL, for a transaction to access a data item, it has to hold the appropriate lock (read or write) on the data item by issuing a lock request. While the way transactions set their lock requests and the way the requests are granted would certainly affect a system's performance, such aspects have not received much attention in the literature. In this paper, a general transaction-processing model is proposed. In this model, a transaction is comprised of a number of stages, and in each stage the transaction can request to lock one or more data items. Methods for granting transaction requests and scheduling policies for granting blocked transactions are also proposed. A comprehensive simulation model is developed from which the performance of 2PL with our proposals is evaluated. Results indicate that performance models in which transactions request locks on an item-by-item basis and use first-come-first-served (FCFS) scheduling in granting blocked transactions underestimate the performance of 2PL. The performance of 2PL can be greatly improved if locks are requested in stages as dictated by the application. A scheduling policy that uses global information and/or schedules blocked transactions dynamically shows a better performance than the default FCFS.     

Semi-Markov decision problems and performance sensitivity analysis

Recent research indicates that Markov decision processes (MDPs) can be viewed from a sensitivity point of view; and the perturbation analysis (PA), MDPs, and reinforcement learning (RL) are three closely related areas in optimization of discrete-event dynamic systems that can be modeled as Markov processes. The goal of this paper is two-fold. First, we develop the PA theory for semi-Markov processes (SMPs); and then we extend the aforementioned results about the relation among PA, MDP, and RL to SMPs. In particular, we show that performance sensitivity formulas and policy iteration algorithms of semi-Markov decision processes can be derived based on the performance potential and realization matrix. Both the long-run average and discounted-cost problems are considered. This approach provides a unified framework for both problems, and the long-run average problem corresponds to the discounted factor being zero. The results indicate that performance sensitivities and optimization depend only on first-order statistics. Single sample path-based implementations are discussed.     

Implementing mixed-criticality synchronous reactive programs upon uniprocessor platforms

Model-based design methodologies based on the synchrony assumption are widely used in many safety-critical application domains. The synchrony assumption asserts that actions (such as the execution of code) occur instantaneously; however, physical platforms obviously do not possess this property. This paper considers the problem of obtaining resource-efficient implementations of programs that are written under the synchrony assumption on actual platforms, such that these implementations execute in a manner that is semantically consistent with the execution sequences that would occur if the synchrony assumption were to hold.