一种基于云模型的文摘单元选取方法研究

该文提出了一种基于云模型的文摘单元选取方法,利用云模型,全面考虑文摘单元的随机性和模糊性,提高面向查询的多文档自动文摘系统的性能。首先计算文摘单元和查询条件的相关性,将文摘单元和各个查询词的相关度看成云滴,通过对云的不确定性的计算,找出与查询条件真正意义相关的文摘单元;随后利用文档集合重要度对查询相关的结果进行修正,将文摘句和其他各文摘句的相似度看成云滴,利用云的数字特征计算句子重要度,找出能够概括尽可能多的文档集合内容的句子,避免片面地只从某一个方面回答查询问题。为了证明文摘单元选取方法的有效性,在英文大规模公开语料上进行了实验,并参加了国际自动文摘公开评测,取得了较好的成绩。
     

基于单词-章节关联的科技论文摘要

为科技论文生成自动摘要,这能够帮助作者更快撰写摘要,是自动文摘的研究内容之一.相比于常见的新闻文档,科技论文具有文档结构性强、逻辑关系明确等特点.目前,主流的编码-解码的生成式文摘模型主要考虑文档的序列化信息,很少深入探究文档的篇章结构信息.为此,文中针对科技论文的特点,提出了一种基于"单词-章节-文档"层次结构的自动摘要模型,利用单词与章节的关联作用增强文本结构的层次性和层级之间的交互性,从而筛选出科技论文的关键信息.除此之外,该模型还扩充了一个上下文门控单元,旨在更新优化上下文向量,从而能更全面地捕获上下文信息.实验结果表明,提出的模型可有效提高生成文摘在ROUGE评测方法上的各项指标性能.     

基于单词-章节关联的科技论文摘要

为科技论文生成自动摘要,这能够帮助作者更快撰写摘要,是自动文摘的研究内容之一.相比于常见的新闻文档,科技论文具有文档结构性强、逻辑关系明确等特点.目前,主流的编码-解码的生成式文摘模型主要考虑文档的序列化信息,很少深入探究文档的篇章结构信息.为此,文中针对科技论文的特点,提出了一种基于"单词-章节-文档"层次结构的自动摘要模型,利用单词与章节的关联作用增强文本结构的层次性和层级之间的交互性,从而筛选出科技论文的关键信息.除此之外,该模型还扩充了一个上下文门控单元,旨在更新优化上下文向量,从而能更全面地捕获上下文信息.实验结果表明,提出的模型可有效提高生成文摘在ROUGE评测方法上的各项指标性能.     

基于主位-述位结构理论的英文作文连贯性建模研究

该文在研究了有监督的基于实体和基于篇章关系网格的篇章连贯性模型的基础上,提出了一个无监督的基于主位-述位结构理论的篇章连贯性模型。该模型通过引入词语的词干、上下位、近义和复述等语义方面的信息来计算相邻句子中主位和述位的相似度,并利用此相似度值来描述篇章的连贯性。同时,该文提出了一种简单有效的基于篇章关系计数的连贯性模型,并采用线性组合方法将其与基于主位-述位结构理论的连贯性模型加以集成。上述模型在国际基准英文作文语料上进行试验,实验结果表明采用线性组合的连贯性模型后,作文连贯性检测准确率与目前基于实体和篇章关系网格的模型相比得到显著提升。     

自动文摘基集语句的提取与润色的数学模型*

针对统计和理解相结合的自动文摘方法,提出了一种新的内容词、有效词和特征词的动态加权函数以及句子重要性的动态加权函数.鉴于基于统计的自动文摘结果常常出现语句间缺乏连贯性及信息冗余的问题,设计了句间语义距离测试函数,并通过大量实验确定语句间语义距离的上限和下限.上限用于控制语句间的逻辑联系,下限用于解决文摘结果信息冗余的问题.实验结果证明,该模型能有效地提取文章中的重点语句,且很好地解决了统计文摘语句不连冠的瓶颈问题.     

主题模型LDA的多文档自动文摘

近年来使用概率主题模型表示多文档文摘问题受到研究者的关注.LDA (latent dirichlet allocation)是主题模型中具有代表性的概率生成性模型之一.提出了一种基于LDA的文摘方法,该方法以混乱度确定LDA模型的主题数目,以Gibbs抽样获得模型中句子的主题概率分布和主题的词汇概率分布,以句子中主题权重的加和确定各个主题的重要程度,并根据LDA模型中主题的概率分布和句子的概率分布提出了2种不同的句子权重计算模型.实验中使用ROUGE评测标准,与代表最新水平的SumBasic方法和其他2种基于LDA的多文档自动文摘方法在通用型多文档摘要测试集DUC2002上的评测数据进行比较,结果表明提出的基于LDA的多文档自动文摘方法在ROUGE的各个评测标准上均优于SumBasic方法,与其他基于LDA模型的文摘相比也具有优势.     

基于篇章修辞结构的自动文摘连贯性研究

尽管抽取式自动文摘方法是目前自动文摘领域的主流方法,并且取得了长足的进步,但抽取式自动文摘形成的摘要由于缺乏句子之间的合理指代或篇章结构,使得文摘缺乏连贯性而影响可读性。为提高自动摘要的可读性,该文尝试将篇章修辞结构信息应用于中文自动文摘。首先,基于汉语篇章修辞结构抽取摘要,然后使用基于LSTM的方法对文本连贯性进行建模,并使用该模型对文摘的连贯性做出评价。实验结果表明: 在摘要抽取方面,基于篇章修辞结构的自动文摘相比于传统的抽取方法具有更好的ROUGE评价值;在使用基于LSTM连贯性模型评价摘要连贯性方面,篇章结构信息在自动抽取文摘时可以很好地提炼出文章的主旨,同时使摘要具有更好的结果。     

基于主题概念抽取的多文档文摘方法

提出一种应用于多文档文摘的有效概念抽取方法。利用WordNet中词语的同义和上下义关系进行语义消歧和概念树构造,通过概念优化算法进行主题概念抽取,建立概念向量空间模型并通过最大边缘相关方法得到文摘句。采用语义概念统计来替代传统的词形统计,能更准确地提取文档中的重要信息。DUC2005的评测结果表明,该方法比传统方法能获得更好的效果。     

基于结构化网格自动划分的汽车发动机强度分析

使用三维设计软件CATIA设计了汽车发动机的复杂结构,然后在网格划分软件HARPOON中将设计的发动机零件自动划分为以结构化六面体网格为主的有限元网格,最后在大型非线性有限元软件ABAQUS中,分析了发动机在静力和动力作用下的应力分布.计算结果表明,采用结构化六面体网格,单元数量少,计算速度快,结果可靠;而采用四面体网格,单元数量大,对于同样配置的计算机,无法进行四面体网格的模型的计算.网格尺寸增加后,单元数量减少,计算机可以计算四面体网格模型,但计算结果与六面体网格偏差比较大.     

作文段落句间逻辑合理性等级评测

在作文评测中,句间逻辑合理性是评价语言运用能力的一项重要指标。从句间连贯性角度出发,采用句子排序的相关算法研究句间词汇中的潜在关联,并对作文段落逻辑合理性进行定量分析,达到对作文段落逻辑合理性等级评测的目的。以计算条件熵为基础,使用马尔科夫随机游走模型进行句子排序;计算过程中融入word2vec和同义词词林进行语义扩展;经ROUGE-L评分,该模型取得了较好的效果。提出带优化的句间逻辑合理性评测模型,实验证明其能够取得最优的效果。建立ROUGE-L分值与由专家评判的段落句间逻辑合理性等级的对应关系,确定ROUGE-L与逻辑合理性等级判定的分类边界,为辅助中文作文智能评测提供了新思路。     

Probing quantum coherence,uncertainty, steerability of quantum coherence and quantum phase transition in the spin model

In this paper, we study the relation among quantum coherence, uncertainty, steerability of quantum coherence based on skew information and quantum phase transition in the spin model by employing quantum renormalization-group method. Interestingly, the results show that the value of the local quantum uncertainty is equal to the local quantum coherence corresponding to local observable \(\sigma _z\) in XXZ model, and unlikely in XY model, local quantum uncertainty is minimal optimization of the local quantum coherence over local observable \(\sigma _x\) and this proposition can be generalized to a multipartite system. Therefore, one can directly achieve quantum correlation measured by local quantum uncertainty and coherence by choosing different local observables \(\sigma _x\), \(\sigma _z\), corresponding to the XY model and XXZ model separately. Meanwhile, steerability of quantum coherence in XY and XXZ model is investigated systematically, and our results reveal that no matter what times the QRG iterations are carried out, the quantum coherence of the state of subsystem cannot be steerable, which can also be suitable for block–block steerability of local quantum coherence in both XY and XXZ models. On the other hand, we have illustrated that the quantum coherence and uncertainty measure can efficiently detect the quantum critical points associated with quantum phase transitions after several iterations of the renormalization. Moreover, the nonanalytic and scaling behaviors of steerability of local quantum coherence have been also taken into consideration.     

Notes on modified trace distance measure of coherence

We investigate the modified trace distance measure of coherence recently introduced in Yu et al. [Phys. Rev. A 94, 060302(R), 2016]. We show that for any single-qubit state, the modified trace norm of coherence is equal to the \(l_{1}\)-norm of coherence. For any d-dimensional quantum system, an analytical formula of this measure for a class of maximally coherent mixed states is provided. The trade-off relation between the coherence quantified by the new measure and the mixedness quantified by the trace norm is also discussed. Furthermore, we explore the relation between the modified trace distance measure of coherence and other measures such as the \(l_{1}\)-norm of coherence and the geometric measure of coherence.     

Fuzzy coherence measures

Coherence measures are a tool to compare those fuzzy sets that are sensitive to their own similarity as well as to their fuzzy nature. Within this article we can find three generalizations made about the definition of coherence measures: a first one for any fuzzy set, a second one for any definition about strong negation, and a final one for an extension in those coherence measures that, as a result, do not cause a value in the unit interval, but a fuzzy set in that interval. Tools and properties are offered to create coherence measures. © 2005 Wiley Periodicals, Inc. Int J Int Syst 20: 1–11, 2005.     

基于神经网络的集句诗自动生成

集句诗是中国古典诗歌的一种特殊体裁。是从前人的诗篇中选取已有诗句,再将其巧妙组合形成一首新诗,是一种艺术的再创造形式。集句诗的生成要求集辑而成的诗不仅合辙押韵, 且有完整的内容、连贯的上下文和新颖的主旨意境,对创作者的知识储备和诗词鉴赏能力有极高的要求。该文基于计算机的海量存储和快速检索能力,以及神经网络模型对文本语义较强的表示和理解能力,提出一种新颖的集句诗自动生成模型。该模型以数十万首古诗作为基础,利用循环神经网络(RNN)自动学习古诗句的语义表示,并设计了多种方法自动计算两句诗句的上下文关联性。根据用户输入的首句,模型能够自动计算选取上下文语义最相关连贯的诗句进行集辑,从而形成一首完整的集句诗。自动评测和人工评测的实验结果都表明,该文模型能够生成质量较好的集句诗,远远超过基线模型的效果。     

音频高层语义分析

为跨越语义鸿沟,提出了一种提取音频中高层语义概念的方法。该方法先用隐马尔可夫模型(HMM)建立对应于分析窗口的低层语义概念,即基本声音语义事件(basic semantic-audio event,BE);然后以音框为单位将声音信号通过短时傅里叶变换及ICA处理来得到对应于HMM模型的可观察符号;接着用贝叶斯决策排除语义窗口对应声音段中的非预定义BE后,按贝叶斯公式所得最大后验概率为准则得到此语义窗口的一个基本声音语义事件组(group of BE,)GBE;最后采用高层语义逻辑定义来描述GBE与高层声音语义概念间的联系,结合由实例训练得到的高层语义逻辑定义最终得到相应语义窗口的高层语义声音概念(high level audio semantic concept,HC)。实验表明此方法能提取与人思维中相似的高层语义概念,在一定程度上可跨越语义鸿沟。     

Transactional coherence and consistency: simplifying parallel hardware and software

Transactional coherence and consistency (TCC) simplifies parallel hardware and software design by eliminating the need for conventional cache coherence and consistency models and letting programmers parallelize a wide range of applications with a simple, lock-free transactional model. TCC eases both parallel programming and parallel architecture design by relying on programmer-defined transactions as the basic unit of parallel work, communication, memory coherence, and memory consistency     

The number of terms in the superpositions upper bounds the amount of the coherence change

For the \(l_{1}\) norm of coherence, what is the relation between the coherence of a state and the individual terms that by superposition yield the state? We find upper bounds on the coherence change before and after the superposition. When every term comes from one Hilbert subspace, the upper bound is the number of terms in the superpositions minus one. However, when the terms have support on orthogonal subspaces, the coherence of the superposition cannot be more the double of the above upper bound than the average of the coherence of the all terms being superposed.     

A New view of the approximations in H v -groups

Let H is an H _v -group and the set of all finite products of elements of H. The relation β^* is the smallest equivalence relation on H such that the quotient H/ β^* is a group. The relation β^* is transitive closure of the relation β, where β is defined as follows: x β y if and only if for some . Based on the relation β, we define a neighborhood system for each element of H, and we presents a general framework for the study of approximations in H _v -groups. In construction approach, a pair of lower and upper approximation operators is defined. The connections between H _v -groups and approximation operators are examined.     

A programme management approach for ensuring curriculum coherence in IS (higher) education

This paper examines IS higher education, concentrating on issues of ‘coherence’ in IS curricula. While curriculum coherence can be jeopardized by poor curriculum design, misalignment between module content and/or misalignment between module or course aims can cause serious coherence issues over time. Misalignment of this type is exacerbated by the traditional processes of curriculum (re)design, which rely heavily on the (singular) interpretation of highly abstract documents, such as module syllabi and course specifications – often produced by curriculum designers in isolation. To improve curriculum coherence, this paper examines the use of a programme management framework as a means of (a) ‘humanizing’ the abstract aims and goals of curricula schemes and (b) managing the delivery and evolution of curricula in relation to the stakeholders in the process of delivery. The practical use of the framework is examined in the context of a Masters-level course in ‘Information Systems Management’. An action research approach is used to demonstrate the practical utility of the framework in terms of (a) improving communication of curricula, (b) improving the coherence between modules and between modules and course and (c) removing content redundancy. Analysis of outcomes demonstrates a significant positive impact of the use of the framework in relation to the above points from the perspective of all stakeholders. Guidelines are presented that generalize the findings in order that key practices may be adopted by others.     

The use of finite-element and boundary-element models for predicting the vibro-acoustic behaviour of layered structures

Layered structures made from elastic and porous materials are widely used as insulation systems in the automotive industry. They have a complex dynamic behaviour that is influenced by the various interaction mechanisms within the porous material. This paper concentrates on modelling these systems using a three-dimensional finite-element (FE) approach for the structure combined with a boundary-element (BE) procedure for the acoustic radiation process. The key part is a Biot model for the two-phase porous material. A mixed displacement formulation is selected. The model can be used for predicting the surface impedance and/or the transmission loss characteristics of layered material. The combined use of this FE/BE model enables evaluation of the acoustic response (radiated power, field pressure). Numerical applications are presented in order to show the capabilities of the developed procedures.