上下文无关语言句子的反向自然枚举

在测试基于复杂数据结构的程序时,需要用到上下文无关语言句子的枚举.基于上下文无关语言按推导树高度的分层构造,提出了句子的反向自然枚举算法.通过堆、层、簇和长方体将句子划分为有穷集合序列,该算法的时间效率为O(n),n是被枚举句子的长度.实验数据表明,该算法是高效的,且应用更加便利.     

基于强制对齐的层次短语模型过滤和优化

该文提出一种层次短语模型过滤和优化方法。该方法在采用传统方法训练得到层次短语规则的基础上,通过强制对齐同时构建源语言和目标语言的解析树,从中过滤并抽取对齐的层次短语规则,最后利用这些规则重新估计翻译模型的翻译概率。该方法不需要引入任何语言学知识,适合大规模语料训练模型。在大规模中英翻译评测任务中,采用该方法训练的模型与传统层次短语模型相比,不仅能够过滤50%左右规则,同时获得0.8～1.2 BLEU值的提高。     

Surface tree languages and parallel derivation trees

The surface tree languages obtained by top-down finite state transformation of monadic trees are exactly the frontier-preserving homomorphic images of sets of derivation trees of ETOL systems. The corresponding class of tree transformation languages is therefore equal to the class of ETOL languages.     

On a relationship between 2-3 brother trees and dense ternary trees

Dense multiway trees have been introduced recently. The question arises whether they can be compared to classes of balanced trees which are already known. This is the case for dense ternary trees and 2–3 brother trees. We prove that the minimal trees of the class of 2–3 brother trees and of the class of strongly dense ternary trees have the same number of leaves. We also correct an error in the original derivation of the number of keys in a minimal 2–3 brother tree.     

基于 Web 的二元化合物结构推演自动流程实现

结构原型是通过任一定义的具有代表性的条件筛选出来的具有相同条件的结构类型。结构推演是一种理论上预测新化合物的方法。首先用特定元素替换大量的结构原型中的元素,得到很多化合物结构;然后,利用第一性原理对这些结构进行结构优化和能量计算,得到每个结构的能量值、形成焓等物理量;最终以形成焓为依据预测这些结构是否有可能形成稳定化合物。这一系列的处理和计算就是一次结构推演。结构推演是一种典型的高通量筛选,涉及到多通道,多任务,高并发的计算。MatCloud 材料计算平台是一个国内自主研发的高通量材料计算平台,其中一个主要功能就是实现了这些多通道,多任务,高并发作业的自动计算流程。为了更为有效及高效地开展结构推演,本论文的工作是在 MatCloud 材料计算平台已有的工作基础上,开发出一个基于 Web 的结构推演自动流程计算模块。     

Behaviours of concurrent systems

Concurrent systems and their behaviours are investigated. The behaviour of a system is understood as the set of processes which the system is capable to realize. The processes may be elementary (indivisible) or may consist of some components. Two ways of composing processes are considered: sequentially (one component is a continuation of another) and in parallel (the components are concurrent, i.e. independent).The behaviour of a system is defined as a set of processes which can be obtained by composing certain elementary processes. All information on the existing independence is reflected so that the system is completely determined by its behaviour.It is explained which sets of processes are the behaviours of concurrent systems.     

Composing recursive logic programs withClausal join

A family of methods for composing logic programs from simpler components is presented. Specifically, simple pairs of programs operating on lists, trees and other recursive structures are composed to generate a single program with composite functionality. The methods are based onclausal join, a specific sequence of unfold/fold transformations for deriving a new clause from a given pair of clauses and a join specification.Procedural join composes a new procedure from two given procedures by applyingclausal join to all pairs of their clauses.1-1 Join composes a new procedure from closely related procedures by applyingclausal join to corresponding pairs only.Meta join is a modification of1-1 join for composing meta-interpreters. The transformations are straightforward to implement in Prolog, as is demonstrated in the paper.     

Lossless fitness inheritance in genetic algorithms for decision trees

When genetic algorithms are used to evolve decision trees, key tree quality parameters can be recursively computed and re-used across generations of partially similar decision trees. Simply storing instance indices at leaves is sufficient for fitness to be piecewise computed in a lossless fashion. We show the derivation of the (substantial) expected speedup on two bounding case problems and trace the attractive property of lossless fitness inheritance to the divide-and-conquer nature of decision trees. The theoretical results are supported by experimental evidence.     

Formal Verification of Activity-Based Specification of Protocols

In this paper, we describe a method to formally verify activity-based specifications such as EBSDL. Starting from EBSDL-like specifications that specify engineering activities in terms of input and output behaviors, we derive programs in an asynchronous language CSP-R. CSP-R programs are then verified by the Maxpar method by composing them with the programs abstracting their environment. EBSDL-like specification and its verification using our method, is illustrated through the example of a fragment of LAPD protocol. The derivation of programs from the specification of activities of the underlying protocols through EBSDL-like specifications provides an important useful tool for formal verification of real-time protocols. We shall also discuss a translation of EBSDL-like specifications to synchronous languages such as Esterel. In the method proposed, it is possible for the user to choose asynchronous or synchronous formalisms depending upon the requirements of verification vis-a-vis logical specification.     

A Characterization of the Sets of Hypertrees Generated by Hyperedge-Replacement Graph Grammars

A characterization of the languages of hypertrees generated by hyperedge-replacement graph grammars is established. The characterization says that these languages are exactly those described by sets of derivation trees which are output languages of finite-copying top-down tree transducers. Furthermore, the statement remains valid if the tree transducers are required to generate derivation trees in which the right-hand sides of productions are hypertrees and the nonterminal hyperedges are at most as large as the hyperedges in the generated hypertrees. This result is closely related to a similar characterization that was obtained for the case of string graphs by Engelfriet and Heyker some years ago. In fact, the results of this paper also yield a new proof for the characterization by Engelfriet and Heyker. Received August 1997, and in final form May 1998.     

Modeling Trees with Hypertextures

In the paper, a new method for modelling trees at medium detail is presented. The method is based on a volumetric representation of trees, generated by an iterated function system (IFS). Alleviating the modeling restrictions of fractal techniques, extensions to the standard IFS are introduced. Practical aspects of modeling and rendering of trees, such as data structures and bounding volumes, are discussed. The advantages of the new method are described at the end together with some results.     

一种BPEL流程数据竞争检测方法

对使用业务流程执行语言(BPEL)组织流程中存在的数据竞争问题进行了研究,根据BPEL特性提出了一个静态分析和动态监控相结合的有效检测方法.实验结果表明,该方法能够准确检测出流程中存在的数据竞争,提高了流程的可靠性.     

基于逻辑行和最大接纳距离的网页正文抽取

网页正文抽取是很多互联网应用的基础工作和必须解决的问题。目前的主流方法是基于DOM树结构,此方法需要解析出网页的DOM树结构。对于目前互联网上的网页来源众多、结构众多的情形,基于DOM树的处理方法除了性能不足以外,还会遇到抽取精度上的问题。针对这些问题,该文提出了一个网页正文抽取的新方法,该方法不依赖DOM树,而是考虑人们编写网页的方式形成一些启发式规则,并结合相关的统计规律,以逻辑行为基本处理单位,基于最大接纳距离进行网页正文抽取。实验表明,论文的方法能够高效、高精度地抽取出网页正文。     

递归程序结构研究

对递归程序的结构进行了较为深入的研究 ,提出了递归树的概念 ,给出了递归程序的一般结构 ,把递归分为简单链结构、树状结构、复杂链结构三种情况 ,据此 ,给出了复杂的递归问题的程序设计方法 ,根据此方法 ,可方便地写出较为复杂的递归问题的递归程序 ,从而提高设计递归程序的效率。     

After Digital Storytelling: Video Composing in the New Media Age

This paper considers the composing processes for novice moviemakers and the disciplinary, pedagogical, and epistemological ramifications of these by using two student case studies. The current model of video composing as a sequential series of temporally discrete steps, made famous by the Center for Digital Storytelling, is predicated on the logic of old media and, thus, may no longer be sufficient to account for the new media composing processes that emerge with new media interfaces. In place of a linear and sequential process of video composing, we suggest expanding the predominant model to account for a simultaneous, multiply-recursive process that involves the semiotic channels of image, sound, and word.     

A domain method for shape design sensitivity analysis of built-up structures

A numerical method of obtaining accurate shape design sensitivity information for built-up structures is developed and demonstrated through examples. The method utilizes and basic character of the finite element method that gives accurate information not on the boundary, but in the domain. The method is shown to drastically simplify the derivation of shape design sensitivity formulas for complex built-up structures.     

Parsing electronic circuits in a logic grammar

Understanding circuits is a prerequisite for circuit design and trouble shooting. Circuit understanding by engineers is described as a process that starts with a structural analysis and then proceeds to a causal analysis. As a step toward automatic circuit understanding, a method for analyzing circuit structures is presented. In this method, a circuit is reviewed as a sentence and its elements as words. Circuit structures are defined by rules written in a logic grammar called definite clause set grammar (DCSG). Given circuits are decomposed into parse trees by the DCSG top-down parsing mechanism. These parse trees represent hierarchical structures of functional blocks. This representation is presented as one step in the process of automatic understanding of circuit structures     

Implied trees in illiquid markets: A Choquet pricing approach

Implied trees are necessary to implement the risk neutral valuation approach, and standard methodologies for their derivation are based on the validity of the put call parity. However, in illiquid markets the put call parity fails to hold, and the uniqueness of the artificial probabilities leaves room for an interval. The contribution of this article is twofold. First we propose a methodology for the derivation of implied trees in illiquid markets. Such a methodology, by contrast with standard ones, takes into account the information stemming both from call and put prices. Second, we set up a framework for pricing derivatives written on an underlying asset traded on an illiquid market. To this end we have extended the Choquet integral definition to account for interval payoffs of the underlying asset. The price interval we obtain may be interpreted as a bid‐ask price quoted by the intermediary issuing the derivative security. © 2002 Wiley Periodicals, Inc.     

An innovations approach to least-squares estimation--Part I: Linear filtering in additive white noise

The innovations approach to linear least-squares approximation problems is first to "whiten" the observed data by a causal and invertible operation, and then to treat the resulting simpler white-noise observations problem. This technique was successfully used by Bode and Shannon to obtain a simple derivation of the classical Wiener filtering problem for stationary processes over a semi-infinite interval. Here we shall extend the technique to handle nonstationary continuous-time processes over finite intervals. In Part I we shall apply this method to obtain a simple derivation of the Kalman-Bucy recursive filtering formulas (for both continuous-time and discrete-time processes) and also some minor generalizations thereof.     

Visual routines for eye location using learning and evolution

Eye location is used as a test bed for developing navigation routines implemented as visual routines within the framework of adaptive behavior-based AI. The adaptive eye location approach seeks first where salient objects are, and then what their identity is. Specifically, eye location involves: 1) the derivation of the saliency attention map, and 2) the possible classification of salient locations as eve regions. The saliency (“where”) map is derived using a consensus between navigation routines encoded as finite-state automata exploring the facial landscape and evolved using genetic algorithms (GAs). The classification (“what”) stage is concerned with the optimal selection of features, and the derivation of decision trees, using GAs, to possibly classify salient locations as eyes. The experimental results, using facial image data, show the feasibility of our method, and suggest a novel approach for the adaptive development of task-driven active perception and navigational mechanisms