一个用于数据并行语言计算划分的时序优化模型

一个程序中数据并行语句的计算划分(CP)对该程序的运行性能有决定性的作用.尽管人们对这一问题已经进行了广泛的研究,但这些研究的重点都集中在如何提高被选择计算划分的空间局部性上.针对并行循环结构的计算划分问题,提出了一个时序优化模型.在该模型中,一个计算划分被表示成一个有向图,在把并行语句中的操作映射到各个处理器的同时,给出了被分配到不同处理器上的操作之间的相关性.对于一条数据并行语句,时序优化模型对它的每个计算划分选择方案分别采用多种有效的优化策略进行优化;并综合考虑各个计算划分选择方案的负载平衡性、处理器间的操作依赖性、数据访问的空间局部性和时间局部性四个方面的因素,估算每个方案的执行效率;最后从这些方案中选择一个执行效率最优的方案作为该语句的计算划分.作者已在HPF编译器p-HPF采用时序优化模型实现了对FORALL结构的支持.实验结果表明,该模型具有非常好的通用性,对不同领域多种数据并行问题均取得了理想的加速比.同时,只需略微改动,该模型也可用于其他类型数据并行语句的计算划分.     

统计句法分析建模中基于信息论的特征类型分析

统计句法分析利用概率评价模型评价每棵选句法树存在的可能性,选择概率值最高的候选句法树作为最终的句法分析结果。因此,统计句法分析的核心是一个概率评价模型,而各种概率评价模型的本质区别主要在于它们分别是根据上下文中的哪些特征来赋予句法树概率的。在统计句法分析研究领域,虽然已经提出了大量的概率评价模型,然而,不同的模型用得到了不同类型的特征,如何评价这些特征类型对于句法分析的作用呢？针对以上的问题,本研究为统计句法分析提出了一种特征类型的分析模型,该模型可以从信息论的角度量化地分析不同类型的上下文特征对于句法结构的预测作用。其基本思想是利用信息论中熵与条件熵的度量来显示一个特征类型是否抓住了预测句法结构的主要信息。如果加入某个特征类型之后当前句法结构的不确定性（熵）明显下降,则认为该特征类型抓住了上下文中影响句法结构的某些主要信息。特征类型分析的信息论模型利用预测信息量、预测信息增益、预测信息关联度以及预测信息总量四种度量从不同的仙量化地分析各种特征类型及特征类型组合对于当前目标的预测作用。实验以Penn TreeBank为训练集,将上下文中不同的特征类型对于句法分析规则的预测作用进行了系统的量化分析,得出了一系列有关不同特征类型及特征类型组合对句法结构的预测作用的结论。     

基于检测点设置依赖图和属性表的卷回恢复算法

为了解决检测点设置过程中的Domino效应问题及卷回恢复过程中的活锁问题,并最大限度地减小时间开销,提出了基于检测点设置依赖图和属性表的卷回恢复算法。同以前的算法相比较,该算法一方面节省了用于进程之间同步的时间开销,另一方面检测点设置及卷回过程中涉及少量的相关进程。对该算法的正确性进行了证明。     

The contingent role of dependency in predicting the intention to adopt B2B e-commerce

ABSTRACT

A review of existing Electronic Commerce adoption literature reveals that the prevailing theoretical frameworks implicitly assume that organizations are independent and have complete control over their decisions to adopt innovations. Given the fact that organizations are often constrained by and depend on their environment and exchange partners to survive, this paper proposes that Resource Dependency can either be a facilitator or inhibitor of the role of Technological, Organizational, and Environmental (TOE) motivating factors. Using 239 valid responses from supply chain managers in Jordan, the study found partial support for the role of Resource Dependency as a moderator. The findings provide insights on how Resource Dependency between partners drives different patterns of innovation acceptance among potential adopters, regardless of several technological and organizational motivations to innovate. The results suggest that predicting how innovation is diffused within a market needs a detailed understanding of the inter-dependence structure among the supply chain members.     

CONTEXTUAL LANGUAGE MODELS FOR RANKING ANSWERS TO NATURAL LANGUAGE DEFINITION QUESTIONS

Question–answering systems make good use of knowledge bases (KBs, e.g., Wikipedia) for responding to definition queries. Typically, systems extract relevant facts from articles regarding the question across KBs, and then they are projected into the candidate answers. However, studies have shown that the performance of this kind of method suddenly drops, whenever KBs supply narrow coverage. This work describes a new approach to deal with this problem by constructing context models for scoring candidate answers, which are, more precisely, statistical n‐gram language models inferred from lexicalized dependency paths extracted from Wikipedia abstracts. Unlike state‐of‐the‐art approaches, context models are created by capturing the semantics of candidate answers (e.g., “novel,”“singer,”“coach,” and “city”). This work is extended by investigating the impact on context models of extra linguistic knowledge such as part‐of‐speech tagging and named‐entity recognition. Results showed the effectiveness of context models as n‐gram lexicalized dependency paths and promising context indicators for the presence of definitions in natural language texts.     

Updating method for the computation of orbits in parallel and sequential dynamical systems

In this article, we provide a matrix method in order to compute orbits of parallel and sequential dynamical systems on Boolean functions. In this sense, we develop algorithms for systems defined over directed (and undirected) graphs when the evolution operator is a general minterm or maxterm and, likewise, when it is constituted by independent local Boolean functions, so providing a new tool for the study of orbits of these dynamical systems.     

Time-bounded controlled bidirectional grammars

We study regularly controlled bidirectional (RCB) grammars from the viewpoint of time-bounded grammars. RCB-grammars are context-free grammars of which the rules can be used in a productive and in a reductive fashion, while the application of these rules is controlled by a regular language. Several modes of derivation can be distinguished for this kind of grammar. A time bound on such a grammar is a measure of its derivational complexity. For some families of time bounds and for some modes of derivation we establish closure properties and a normal form theorem. In addition parsing algorithms are given for some modes of derivation. We conclude with considering generalizations with respect to the family of control languages and the family of bounding functions..     

Quality aspects for component‐based systems: A metrics based approach

In component‐based development, software systems are built by assembling components already developed and prepared for integration. To estimate the quality of components, complexity, reusability, dependability, and maintainability are the key aspects. The quality of an individual component influences the quality of the overall system. Therefore, there is a strong need to select the best quality component, both from functional and nonfunctional aspects. The present paper produces a critical analysis of metrics for various quality aspects for components and component‐based systems. These aspects include four main quality factors: complexity, dependency, reusability, and maintainability. A systematic study is applied to find as much literature as possible. A total of 49 papers were found suitable after a defined search criteria. The analysis provided in this paper has a different objective as we focused on efficiency and practical ability of the proposed approach in the selected papers. The various key attributes from these two are defined. Each paper is evaluated based on the various key parameters viz. metrics definition, implementation technique, validation, usability, data source, comparative analysis, practicability, and extendibility. The paper critically examines various quality aspects and their metrics for component‐based systems. In some papers, authors have also compared the results with other techniques. For characteristics like complexity and dependency, most of the proposed metrics are analytical. Soft computing and evolutionary approaches are either not being used or much less explored so far for these aspects, which may be the future concern for the researchers. In addition, hybrid approaches like neuro‐fuzzy, neuro‐genetic, etc., may also be examined for evaluation of these aspects. However, to conclude that one particular technique is better than others may not be appropriate. It may be true for one characteristic by considering different set of inputs and dataset but may not be true for the same with different inputs. The intension in the proposed work is to give a score for each metric proposed by the researchers based on the selected parameters, but certainly not to criticize any research contribution by authors. Copyright © 2012 John Wiley & Sons, Ltd.     

Video events recognition by improved stochastic parsing based on extended stochastic context-free grammar representation

Video events recognition is a challenging task for high-level understanding of video sequence. At present,there are two major limitations in existing methods for events recognition. One is that no algorithms are available to recognize events which happen alternately. The other is that the temporal relationship between atomic actions is not fully utilized. Aiming at these problems,an algorithm based on an extended stochastic contextfree grammar ( SCFG) representation is proposed for events recognition. Events are modeled by a series of atomic actions and represented by an extended SCFG. The extended SCFG can express the hierarchical structure of the events and the temporal relationship between the atomic actions. In comparison with previous work,the main contributions of this paper are as follows: ① Events ( include alternating events) can be recognized by an improved stochastic parsing and shortest path finding algorithm. ② The algorithm can disambiguate the detection results of atomic actions by event context. Experimental results show that the proposed algorithm can recognize events accurately and most atomic action detection errors can be corrected simultaneously.