Summarizing Vocabularies in the Global Semantic Web

In the Semantic Web, vocabularies are defined and shared among knowledge workers to describe linked data for scientific, industrial or daily life usage. With the rapid growth of online vocabularies, there is an emergent need for approaches helping users understand vocabularies quickly. In this paper, we study the summarization of vocabularies to help users understand vocabularies. Vocabulary summarization is based on the structural analysis and pragmatics statistics in the global Semantic Web. Local Bipa...     

语用视角下复述句生成方式的类型考察

该文将汉语母语者的160份复述文本与其原文进行以小句为单位的逐句比对,发现其中出现了6 484对复述句对。从其生成的方式来看,可以分为改换词语和重铸整句两大类。以语用学原理对这些复述句进行分析,发现与以往研究的复述现象不同的是: 句对间往往不具有相同的逻辑语义真值,但在特定语境下却能传达同一个语用意义,具有等效的语用功能。这说明在自然语言处理中,识别进入真实交际中的复述句不仅依赖语法、语义知识库,还需要借助含有语用知识和语境信息的知识库。     

Globally exponentially attractive sets of the family of Lorenz systems

In this paper, the concept of globally exponentially attractive set is proposed and used to consider the ultimate bounds of the family of Lorenz systems with varying parameters. Explicit estimations of the ultimate bounds are derived. The results presented in this paper contain all the existing results as special cases. In particular, the critical cases, b→ 1^＋ and a→0^＋, for which the previous methods failed, have been solved using a unified formula.     

基于图神经网络和语义知识的自然语言推理任务研究

自然语言推理任务的目的是推断两个句子之间的语义逻辑关系.该文通过模仿人类的推理过程构造模型,首先利用长短时记忆网络提取词的语境特征,模仿人类粗读句子的过程;然后依据外部语义知识,连接两个句子中有语义联系的词,构造一个以词为节点的语义图;接下来模仿人类比较两个句子的语义角色相似性的思维,用图卷积或图注意力神经网络聚合词在...     

On the 2<Superscript>m</Superscript>-variable symmetric Boolean functions with maximum algebraic immunity

The properties of the 2m-variable symmetric Boolean functions with maximum al- gebraic immunity are studied in this paper. Their value vectors, algebraic normal forms, and algebraic degrees and weights are all obtained. At last, some necessary conditions for a symmetric Boolean function on even number variables to have maximum algebraic immunity are introduced.     

Tracing application program execution on the CRAY X-MP and CRAY-2

Important insights into program operation can be gained by observing dynamic execution behavior. Unfortunately, many high-performance machines provide execution profile summaries as the only tool for performance investigation. We have developed a tracing library for the CRAY X-MP and CRAY-2 supercomputers that supports the low-overhead capture of execution events for sequential and multitasked programs. This library has been extended to use the automatic instrumentation facilities on these machines, allowing trace data from routine entry and exit, and other program segments, to be captured. To assess the utility of the trace-based tools, three of the Perfect Benchmark codes have been tested in scalar and vector modes with the tracing instrumentation. In addition to computing summary execution statistics from the traces, interesting execution dynamics appear when studying the trace histories. It is also possible to model application performance based on properties identified from traces. Our conclusion is that adding tracing support in Cray supercomputers can have significant returns in improved performance characterization and evaluation.An earlier version of this paper was presented at Supercomputing '90.Supported in part by the National Science Foundation under Grants No. NSF MIP-88-07775 and No. NSF ASC-84-04556, and the NASA Ames Research Center Grant No. NCC-2-559.Supported in part by the National Science Foundation under grant NSF ASC-84-04556.Supported in part by the National Science Foundation under grants NSF CCR-86-57696, NSF CCR-87-06653 and NSF CDA-87-22836 and by the National Aeronautics and Space Administration under NASA Contract Number NAG-1-613.     

面向信息处理的语境形式化研究

在自然语言处理研究领域中,句法研究已经取得了可喜的进展,语义研究也日益受到重视。但要真正实现计算机理解自然语言的目标,还必须进一步深入开展语用分析研究,目前还很少有这方面的研究成果。语境是语用学研究中的重要内容,本文首先讨论了现代语言学关于语境的定义,对计算语言学中的语境作了科学的界定,然后给出了语境及其相关概念的形式定义,并结合汉语实例进行了分析。最后指出“计算语用学”这一计算语言学领域中的新兴学科将大有可为。     

A prolog-like inference system for computing minimum-cost abductive explanations in natural-language interpretation

By determining what added assumptions would suffice to make the logical form of a sentence in natural language provable, abductive inference can be used in the interpretation of sentences to determine what information should be added to the listener's knowledge, i.e., what he should learn from the sentence. This is a comparatively new application of mechanized abduction. A new form of abduction — least specific abduction — is proposed as being more appropriate to the task of interpreting natural language than the forms that have been used in the traditional diagnostic and design-synthesis applications of abduction. The assignment of numerical costs to axioms and assumable literals permits specification of preferences on different abductive explanations. A new Prolog-like inference system that computes abductive explanations and their costs is given. To facilitate the computation of minimum-cost explanations, the inference system, unlike others such as Prolog, is designed to avoid the repeated use of the same instance of an axiom or assumption.This research is supported by the Defense Advanced Research Projects Agency, under Contract N00014-85-C-0013 with the Office of Naval Research, and by the National Science Foundation, under Grant CCR-8611116. The views and conclusions contained herein are those of the author and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the Defense Advanced Research Projects Agency, the National Science Foundation, or the United States government.     

Verifying functions in online stock trading systems

Temporal colored Petri nets, an extension of temporal Petri nets, are introduced in this paper. It can distinguish the personality of individuals (tokens), describe clearly the causal and temporal relationships betwee nevents in concurrent systems, and represent elegantly certain fundamental properties of concurrent systems, such as eventuality and fairness. The use of this method is illustrated with an example of modeling and formal verification of an online stock trading system. The functional correctness of the modeled system is formally verified based on the temporal colored Petri net model and temporal assertions. Also, some main properties of the system are analyzed. It has been demonstrated sufficiently that temporal colored Petri nets can verify efficiently some time-related properties of concurrent systems, and provide both the power of dynamic representation graphically and the function of logical inference formally. Finally. future work is described.     

Randomized competitive algorithms for the list update problem

We prove upper and lower bounds on the competitiveness of randomized algorithms for the list update problem of Sleator and Tarjan. We give a simple and elegant randomized algorithm that is more competitive than the best previous randomized algorithm due to Irani. Our algorithm uses randomness only during an initialization phase, and from then on runs completely deterministically. It is the first randomized competitive algorithm with this property to beat the deterministic lower bound. We generalize our approach to a model in which access costs are fixed but update costs are scaled by an arbitrary constantd. We prove lower bounds for deterministic list update algorithms and for randomized algorithms against oblivious and adaptive on-line adversaries. In particular, we show that for this problem adaptive on-line and adaptive off-line adversaries are equally powerful.A preliminary version of these results appeared in a joint paper with S. Irani in theProceedings of the 2nd Symposium on Discrete Algorithms, 1991 [17].This research was partially supported by NSF Grants CCR-8808949 and CCR-8958528.This research was partially supported by NSF Grant CCR-9009753.This research was supported in part by the National Science Foundation under Grant CCR-8658139, by DIMACS, a National Science Foundation Science and Technology center, Grant No. NSF-STC88-09648.     

句间回指模糊的多层次消解模型

句间回指消解是当前中文信息处理的一个重要研究课题,直接从语义和语用入手,以法律文本为语料来源,对句间回指进行形式化描述和消解,服务于计算机句群自动理解。概念基元是“显微镜”,看清指代语与先行语的微观语义联系;句类是“放大镜”,将指代语和先行语纳入57组基本句类中进行关联;语境单元则是“望远镜”,为指代语和先行语提供宏观的语境知识。语料考察结果表明,这一多层次的消解模型对实现句间回指消解是有效的。     

Redundancy check technique for designing optimal networks by branch-and-bound method

A redundancy check procedure is presented which is intended to reduce the computation time of the branch- and-bound algorithm for designing minimal NOR networks under various network constraints, such as fan-in, fan-out limitations and level restrictions. Several properties of NOR networks which form the theoretical basis for the redundancy check procedure are also discussed. A computational procedure incorporating some of these properties was implemented and tested. The results show that this procedure is quite effective in reducing the computation time of the branch-and-bound algorithm.This work was supported in part by the National Science Foundation under Grant No. NSF GJ-503.     

基于循环卷积神经网络的藏文句类识别

句子是语言的最小使用单位,句类识别是为了进一步细化句法和句义研究。由于藏文句尾通常没有特殊的标点符号来识别不同句类,因此这一藏文语言特性就变成了一大难题。该文提出了基于语境和功能特征为一体的句子用途分类方案。首先,该文介绍了文法中藏文句子分类及其特征。其次,收集了大量藏文句子并对其进行了人工标注。最后,采用循环卷积神经网络对藏文句类进行了自动识别。实验表明,该模型对藏文句类识别有较为显著的效果。     

Power allocation for non-orthogonal decode-and-forward cooperation protocol

This paper studies the power allocation problem for the non-orthogonal decode-and-forward (NDF) cooperation protocol with selection relaying. With the availability of the magnitudes of all channel gains at the source, the power allocation is explored that maximizes the mutual information between the source and destination subject to a total power constraint. The minimum power that avoids the outage of the relay is set as a condition, under which the power allocation problem becomes one of selecting the optimal one from several allocation factor triplets. It is shown that the power allocation scheme can provide considerable performance gain, and the non-orthogonal cooperation protocol is superior to the orthogonal protocol and direct transmission. Supported by the Project of Natural Science Foundation of Jiangsu Province (Grant No. BK2006701), the National Natural Science Foundation of China (Grant No. 60672079), and the National High-Tech Research & Development Program of China (Grant No. 2007AA01Z267)     

Efficient Optimization of Multiple Subspace Skyline Queries

We present the first efficient sound and complete algorithm （i.e., AOMSSQ） for optimizing multiple subspace skyline queries simultaneously in this paper. We first identify three performance problems of the na/ve approach （i.e., SUBSKY） which can be used in processing arbitrary single-subspace skyline query. Then we propose a cell-dominance computation algorithm （i.e., CDCA） to efficiently overcome the drawbacks of SUBSKY. Specially, a novel pruning technique is used in CDCA to dramatically decrease the query time. Finally, based on the CDCA algorithm and the share mechanism between subspaces, we present and discuss the AOMSSQ algorithm and prove it sound and complete. We also present detailed theoretical analyses and extensive experiments that demonstrate our algorithms are both efficient and effective.     

基于篇章理解的自动文摘研究

本文在自然语言理解的基础上, 对自动文摘系统进行研究。研究在两级上进行, 一级是基于中心名词, 中心动词同其修饰成分间关系研究, 实现句子成分过滤, 完成句子主干提取。一级是上下文级, 基于篇章文体结构, 句子语义关联, 分别实现摘要基集产生算法, 摘要基集扩充算法及摘要集合生成算法。摘要基集的产生依赖于中心段, 中心句选取, 摘要基集扩充和摘要集合产生依赖于模糊语义距离的测试。     

Composite Distance Transformation for Indexing and k-Nearest-Neighbor Searching in High-Dimensional Spaces

Due to the famous dimensionality curse problem, search in a high-dimensional space is considered as a ＂hard＂ problem. In this paper, a novel composite distance transformation method, which is called CDT, is proposed to support a fast κ-nearest-neighbor （κ-NN） search in high-dimensional spaces. In CDT, all （n） data points are first grouped into some clusters by a κ-Means clustering algorithm. Then a composite distance key of each data point is computed. Finally, these index keys of such n data points are inserted by a partition-based B^＋-tree. Thus, given a query point, its κ-NN search in high-dimensional spaces is transformed into the search in the single dimensional space with the aid of CDT index. Extensive performance studies are conducted to evaluate the effectiveness and efficiency of the proposed scheme. Our results show that this method outperforms the state-of-the-art high-dimensional search techniques, such as the X-Tree, VA-file, iDistance and NB-Tree.     

面向社交媒体评论的上下文语境讽刺检测模型

讽刺是日常交际中一种常见的语用现象,能够丰富说话者的观点并间接地表达说话者的深层含义。讽刺检测任务的研究目标是挖掘目标语句的讽刺倾向。针对讽刺语境表达变化多样以及不同用户、不同主题下的讽刺含义各不相同等特征,构建融合用户嵌入与论坛主题嵌入的上下文语境讽刺检测模型。该模型借助ParagraphVector方法的序列学习能力对用户评论文档与论坛主题文档进行编码,从而获取目标分类句的用户讽刺特征与主题特征,并利用一个双向门控循环单元神经网络得到目标句的语句编码。在标准讽刺检测数据集上进行的实验结果表明,与传统Bag-of-Words、CNN等模型相比,该模型能够有效提取语句的上下文语境信息,具有较高的讽刺检测分类准确率。