语义分析和结构化语言模型

提出了一个语义分析集成系统,并在此基础上构建了结构化的语言模型.该语义分析集成系统能够自动分析句子中各个词的词义以及词之间的语义依存关系,达到90.85%的词义标注正确率和75.84%的语义依存结构标注正确率.为了描述语言的结构信息和长距离依存关系,研究并分析了两种基于语义结构的语言模型.最后,在中文语音识别任务上测试两类语言模型的性能.与三元语言模型相比,性能最好的语义结构语言模型--中心词三元模型,使绝对字错误率下降0.8%,相对错误率下降8%.     

融合语义类信息的句法分析统计模型

稀疏数据严重影响句子结构分析模型的结果, 而句法结构是语义内容和句法分析形式的结合。本文在语义结构信息标注的基础上提出了一种基于语义搭配关系的词聚类模型和算法,建立基于语义类的头驱动句子结构分析统计模型。该语言模型不但比较成功地解决了数据稀疏问题, 而且句子结构分析系统性能也有了明显的提高。句子结构分析实验结果表明,基于语义类的头驱动的句子结构分析统计模型,其召回率和精确率的值相应为88.26%和88.73%,综合指标改进了8.39%。     

BERT跨语言词向量学习研究

随着互联网多语言信息的发展,如何有效地表示不同语言所含的信息已成为自然语言信息处理的一个重要子任务,因而跨语言词向量成为当下研究的热点.跨语言词向量借助迁移学习将单语词向量映射到一个共享的低维空间,在不同语言间进行语法、语义和结构特征的迁移,能够对跨语言语义信息进行建模.B E RT模型通过大量语料的训练,得到一种通用的词向量,同时根据具体的下游任务进一步动态优化,生成上下文语境敏感的动态词向量,解决了以往模型的聚义问题.通过对现有基于B E RT的跨语言词向量研究的文献回顾,综合阐述了基于B E RT的跨语言词向量学习方法、模型、技术的发展,以及所需的训练数据.根据训练方法的不同,分为有监督学习和无监督学习两类,并对两类方法的代表性研究进行详细的对比和总结.最后概述了跨语言词向量的评估方法,并以构建基于B E RT的蒙汉文跨语言词向量进行展望.     

基于描述约束的词表示学习

词语作为语言模型中的基本语义单元,在整个语义空间中与其上下文词语具有很强的关联性。同样,在语言模型中,通过上下文词可判断出当前词的含义。词表示学习是通过一类浅层的神经网络模型将词语和上下文词之间的关联关系映射到低维度的向量空间中。然而,现有的词表示学习方法往往仅考虑了词语与上下文词之间的结构关联,词语本身所蕴含的内在语义信息却被忽略。因此,该文提出了DEWE词表示学习算法,该算法可在词表示学习的过程中不仅考量词语与上下文之间的结构关联,同时也将词语本身的语义信息融入词表示学习模型,使得训练得到的词表示既有结构共性也有语义共性。实验结果表明,DEWE算法是一种切实可行的词表示学习方法,相较于该文使用的对比算法,DEWE在6类相似度评测数据集上具有优异的词表示学习性能。     

基于语义依存关系的汉语语料库的构建

语料库是自然语言处理中用于知识获取的重要资源。本文以句子理解为出发点,讨论了在设计和建设一个基于语义依存关系的汉语大规模语料库过程中的几个基础问题,包括:标注体系的选择、标注关系集的确定,标注工具的设计,以及标注过程中的质量控制。该语料库设计规模100万词次,利用70个语义、句法依存关系,在已具有语义类标记的语料上进一步标注句子的语义结构。其突出特点在于将《知网》语义关系体系的研究成果和具体语言应用相结合,对实际语言环境中词与词之间的依存关系进行了有效的描述,它的建成将为句子理解或基于内容的信息检索等应用提供更强大的知识库支持。     

汉语语句形式结构到语义结构的理解自明度

计算机理解语句的方法之一是从形式结构出发得到语义结构。本文从计算机的角度给出了语句形式结构的表示方法,基于概念层次网络理论说明了语句语义结构的描述单位,介绍了从语句形式结构得到语句语义结构的分析方法。用理解自明度来反映分析结果的可靠性,并给出了描述分析过程中各子过程的理解自明度的计算公式。理解自明度是衡量系统分析性能的一个重要指标,也有助于系统解模糊能力的改进和提高。     

基于知识图的汉语基本名词短语分析模型

本文提出了一种基于知识图的汉语baseNP分析模型。它以知识图为知识表示方法,利用《知网》为语义知识资源,采用以语义为主、语法为辅的策略,先为短语中的每一个实词构造“词图”,然后合并“词图”而组成“短语图”,最后得到一个关于汉语baseNP结构信息和语义信息的知识图。因此它不仅分析了汉语baseNP结构的内部句法关系,而且分析了汉语baseNP结构成分间的语义关系并以知识图的形式表示出了这种语义关系。实验结果表明这个模型对于汉语baseNP的分析是有效的。     

现代汉语语义知识库用于句法分析的研究

汉语在词类这个语言层次上存在着许多歧义结构,仅仅依靠句子中词的词性标记无法正确获得汉语词之间的句法关系.针对这种情况,在已有的基于规则与概率模型的句法分析器基础上,使用现代汉语语义知识库作为语义资源,进行了在句法分析中引入语义信息以提高分析正确性的研究.     

一种改进的带有情感信息的词向量学习方法

词语的情感信息对于情感分析任务至关重要,现有大多数基于词向量的无监督学习方法只能对词语的语法语境建模,但忽略了词语的情感信息。本文提出了一种结合监督学习和非监督学习的词向量学习方法：既能够获得词语的语义信息又能够获得情感内容。在相关实验中,论文对词向量分析做了直观地举例对比,并将该方法用于情感分类任务中,通过引入新的评论数据集对本文方法进行验证。实验结果表明,融合了语义与情感的词向量方法效果良好,能更为精确地对情感信息进行分类,更为客观地对用户信息进行评价,助力社交网络良性发展。     

多原型词向量与文本主题联合学习模型

常见的词嵌入向量模型存在每个词只具有一个词向量的问题,词的主题值是重要的多义性条件,可以作为获得多原型词向量的附加信息。在skip-gram(cbow)模型和文本主题结构基础上,该文研究了两种改进的多原型词向量方法和基于词与主题的嵌入向量表示的文本生成结构。该模型通过联合训练,能同时获得文本主题、词和主题的嵌入向量,实现了使用词的主题信息获得多原型词向量,和使用词和主题的嵌入式向量学习文本主题。实验表明,该文提出的方法不仅能够获得具有上下文语义的多原型词向量,也可以获得关联性更强的文本主题。     

Transactions and updates in deductive databases

In this paper, we develop a new approach that provides a smooth integration of extensional updates and declarative query languages for deductive databases. The approach is based on a declarative specification of updates in rule bodies. Updates are not executed as soon as evaluated. Instead, they are collected and then applied to the database when the query evaluation is completed. We call this approach nonimmediate update semantics. We provide a top-down and equivalent bottom-up semantics which reflect the corresponding computation models. We also package set of updates into transactions and we provide a formal semantics for transactions. Then, in order to handle complex transactions, we extend the transaction language with control constructors still preserving formal semantics and semantics equivalence     

An algebraic semantics for MOF

In model-driven development, software artifacts are represented as models in order to improve productivity, quality, and cost effectiveness. In this area, the meta-object facility (MOF) standard plays a crucial role as a generic framework within which a wide range of modeling languages can be defined. The MOF standard aims at offering a good basis for model-driven development, providing some of the building concepts that are needed: what is a model, what is a metamodel, what is reflection in the MOF framework, and so on. However, most of these concepts are not yet fully formally defined in the current MOF standard. In this paper we define a reflective, algebraic, executable framework for precise metamodeling based on membership equational logic (mel) that supports the MOF standard. Our framework provides a formal semantics of the following notions: metamodel, model, and conformance of a model to its metamodel. Furthermore, by using the Maude language, which directly supports mel specifications, this formal semantics is executable. This executable semantics has been integrated within the Eclipse modeling framework as a plugin tool called MOMENT2. In this way, formal analyses, such as semantic consistency checks, model checking of invariants and LTL model checking, become available within Eclipse to provide formal support for model-driven development processes.     

Mechanized Semantics for the Clight Subset of the C Language

This article presents the formal semantics of a large subset of the C language called Clight. Clight includes pointer arithmetic, struct and union types, C loops and structured switch statements. Clight is the source language of the CompCert verified compiler. The formal semantics of Clight is a big-step operational semantics that observes both terminating and diverging executions and produces traces of input/output events. The formal semantics of Clight is mechanized using the Coq proof assistant. In addition to the semantics of Clight, this article describes its integration in the CompCert verified compiler and several ways by which the semantics was validated.     

SQL语言的形式语义

对SQL查询的形式语义的研究有助于形式地证明两条SQL语句是否等价,从而消除了自然语言的二义性。SQL标准对SQL的语义规则进行了定义,但是并没有很好地处理不完全信息问题。文中以中介逻辑谓词演算系统MFM为基础,构造一个形式的三值谓词演算模型EPMC,然后通过语法转化规则把SQL查询转化为EPMC,从而完整地定义了SQL查询的形式语义。     

面向方面软件重构等价性形式化证明方法

软件重构在不改变程序行为的情况下通过对代码进行小的改进以提升设计,使之更容易理解和维护,面向方面的程序设计是软件开发的新技术,为了有效实施面向方面的软件重构,需要开发者识别面向方面程序的转化规则。然而,由于使用的AOP语言没有形式化的语义定义,难以确认转化和重构的程序运行行为。本文对MCI操作语义的面向方面的扩展使之支持程序的方面特征的描述,定义了两个程序的观测等价,讨论了AspectJ的形式语义模型的建立,在MCI的语义下形式化地精确证明了Add Before-executing编程规则的观测等价性,其基本原理和方法可以适用于其他规则的证明,通过上述工作提出了面向方面重构的程序和它的面向对象程序原型等价性的证明方法。     

On the tree-transformation power of XSLT

XSLT is a standard rule-based programming language for expressing transformations of XML data. The language is currently in transition from version 1.0 to 2.0. In order to understand the computational consequences of this transition, we restrict XSLT to its pure tree-transformation capabilities. Under this focus, we observe that XSLT 1.0 was not yet a computationally complete tree-transformation language: every 1.0 program can be implemented in exponential time, using a DAG representation of trees. A crucial new feature of version 2.0, however, which allows node sets over temporary trees, yields completeness. We provide a formal operational semantics for XSLT programs, and establish confluence for this semantics.Wim Janssen is a research assistant of the Fund for Scientific Research, Flanders. Alexandr Korlyukov sadly passed away shortly after we agreed to write a joint paper.     

AADL进程子集行为语义研究

AADL是一种基于组件的半形式化建模语言,采用结构化方法对大型复杂软件系统的软硬件进行统一建模,有效描述系统的功能行为、非功能属性以及运行时的体系结构动态演化,但其许多问题需要进一步研究与完善。本文首先分析了AADL形式语义研究现状,然后定义了AADL进程子集的形式语言,建立了AADL进程子集通信模型,通过对事件的形式化定义和分析体现了事件在系统状态转变过程中的重要作用,对AADL进程子集行为语义进行了研究。与相关研究成果的对比说明了本文的优势。本文为AADL语言及其形式语义的发展提供了一种有益的参考,进一步完善大型复杂软件系统体系结构建模与分析技术。     

Alliance of model-driven engineering with a proof-based formal approach

Model-driven engineering (MDE) promotes the use of models throughout the software development cycle in order to increase abstraction and reduce software complexity. It favors the definition of domain-specific modeling languages (DSMLs) thanks to frameworks dedicated to meta-modeling and code generation like EMF (Eclipse Modeling Framework). The standard semantics of meta-models allows interoperability between tools such as language analysers (e.g., XText), code generators (e.g., Acceleo), and also model transformation tools (e.g., ATL). However, a major limitation of MDE is the lack of formal reasoning tools allowing to ensure the correctness of models. Indeed, most of the verification activities offered by MDE tools are based on the verification of OCL constraints on instances of meta-models. However, these constraints mainly deal with structural properties of the model and often miss out its behavioral semantics. In this work, we propose to bridge the gap between MDE and the rigorous world of formal methods in order to guarantee the correctness of both structural and behavioral properties of the model. Our approach translates EMF meta-models into an equivalent formal B specification and then injects models into this specification. The equivalence between the resulting B specification and the original EMF model is kept by proven design steps leading to a rigorous MDE technique. The AtelierB prover is used to guarantee the correctness of the model’s behavior with respect to its invariant properties, and the ProB model-checker is used to animate underlying execution scenarios which are translated back to the initial EMF model. Besides the use of these automatic reasoning tools in MDE, proved B refinements are also investigated in this paper in order to gradually translate abstract EMF models to concrete models which can then be automatically compiled into a programming language.

     

SEMANOL (73) a metalanguage for programming the semantics of programming languages

Summary SEMANOL is a practical programming system for writing readable formal specifications of the syntax and semantics of programming languages. SEMANOL is based on a theory of semantics which embraces algorithmic (operational) and extensional (input/output) semantics. Specifications for large contemporary languages have been constructed in the formal language, SEMANOL (73), which is a readable high-level notation. A SEMANOL (73) specification can be executed (by an existing interpreter program); when given a program from the specified language, and its input, the execution of the SEMANOL (73) specification produces the program's output. The demonstrated executability of SEMANOL (73) provides important practical advantages. This paper includes discussions of the theory of semantics underlying SEMANOL, the syntax and semantics of the SEMANOL (73) language, the use of the SEMANOL (73) language in the SEMANOL method for describing programming languages, and the contrast between the Vienna definition method (VDL) and SEMANOL.     

ODRL权利描述语言逻辑实施机制研究

针对缺乏正式语义使基于XML的ODRL等权利描述语言的确切含义依赖应用程序的特定理解、易产生二义性和不确定性等问题,将ODRL,语言转换成一种基于逻辑的权利描述语言,为ODRL语言提供了正式语义和策略实施的逻辑框架,为ODRL语言实施提供可信和形式化分析基础.在ODRL规范基础上对转换方法进行了说明和示例,最后描述了实施平台.