语义相似度计算在语义标注中的应用

提出了一种词汇和本体概念间的语义相似度计算方法。该方法利用编辑距离和维基百科从语法和语义两方面综合考虑词汇和概念间的语义相似度。在领域本体的指导下,将方法应用于语义标注过程,建立词汇与本体概念之间的映射。在标注过程中建立知识库,提高算法性能,实验结果说明该方法是行之有效的。     

反馈日志与混合概率模型相结合的图像标注

为提高图像标注质量,提出一种反馈日志与混合概率模型相结合的图像标注方法。利用本体语义网计算标注词之间的相似性度,将相似度应用于日志分析,得到具体应用中的标注词间关系,结合标注词间的关系和图像底层特征,使用混合概率模型进行自动图像标注。实验结果表明,该方法能获得较好的查全率和查准率。     

融合骨架结构与相关传递性的三维模型标注方法

为了提高和改进三维模型自动语义标注技术,弥补已有的模型自动语义标注方法在精确性和鲁棒性方面的不足,在三维模型与语义词汇相关传递原理的基础上,提出一种融合相似词汇概率、模型语义直接相关性和模型整体相似性的三维模型与语义词汇之间的相似度计算算法。并结合模型骨架结构的拓扑属性与局部突起的几何形状属性在模型检索中的应用,设计出一种新的三维模型自动语义标注方法。一系列的实验结果验证了该方法的准确性与鲁棒性。     

基于加权的本体相似度计算方法

为优化基于本体的语义推理效果,提出了对本体中概念结点赋予权重的相似度计算方法.通过定义本体树中深度因子和密度因子,以解决本体中概念深度与密度对相似度计算的影响.利用Jena API、Lucene等开源工具包,提出了查询扩展方法.实验结果表明,提出的基于加权语义相似度计算模型与传统的计算法方法以及主观判断的方法相比,提高了相似度计算的准确性,效率有明显提高.     

一种计算组合词汇语义相似度的方法

对语义相似度的研究,已产生了许多计算模型和计算方法.这些模型和方法大多集中于简单词汇之间的相似度计算,很少有涉及两个或多个词汇组合的相似度计算.在本体匹配与服务发现的过程中,与本体有关的概念不可避免的会出现由两个或多个词汇组合成的组合词汇.在WordNet相似度计算基础上,综合考虑了词汇间的各类相似度,提出了一种计算组合词汇的语义相似度的方法.实验表明,该方法能有效提升相似度计算的有效性和准确性.     

基于本体的语义虚拟环境查询与推理模型

当前将本体引入到语义虚拟环境的研究,只是将领域本体的可视化信息用本体表示,并未发挥本体本身具有的优势。为此,提出一种基于本体的语义虚拟环境查询与推理模型。利用OWL语言统一描述虚拟场景图形内容与语义信息,并分别对两者进行查询,在图形内容查询过程中引入本体的推理方法推理出隐含的图形内容信息,然后查询需要的信息。在语义信息查询时引入语义搜索方法,利用基于语义距离计算本体概念相似度的方法计算语义虚拟环境本体中类之间的相似度,搜索与被查询实例语义相似度最大的实例,并借助推理找出其间的关系。对语义虚拟家具商店进行本体的查询与推理,结果证明了该模型的可行性。     

面向Web应用的语义标注方法

提出了一种语义标注的方法来支持用户在网上的浏览活动.采用了基于参考本体转换技术的语义转换,它能够从语义上同类型的标注资源中提取特征.随着获取标注资源的模式建立用户意向模型,利用概率的方法识别用户意向.然后利用启发式函数量化具体用户意向和资源之间的相似度,从而达到用户在浏览语义异构网络信息空间时获取相关信息的目的.     

融合深度特征和语义邻域的自动图像标注*

传统图像标注方法中人工选取特征费时费力,传统标签传播算法忽视语义近邻,导致视觉相似而语义不相似,影响标注效果.针对上述问题,文中提出融合深度特征和语义邻域的自动图像标注方法.首先构建基于深度卷积神经网络的统一、自适应深度特征提取框架,然后对训练集划分语义组并建立待标注图像的邻域图像集,最后根据视觉距离计算邻域图像各标签的贡献值并排序得到标注关键词.在基准数据集上实验表明,相比传统人工综合特征,文中提出的深度特征维数更低,效果更好.文中方法改善传统视觉近邻标注方法中的视觉相似而语义不相似的问题,有效提升准确率和准确预测的标签总数.     

基于个性化本体的图像语义标注和检索

针对目前图像检索系统较难实现语义检索的问题,提出了一种新的以本体为核心的图像语义标注和检索模型。构建个性化本体描述图像语义,继而提取基于概念集的图像语义特征并利用本体中“Is-A”关系设计相似性度量方法最终实现语义扩展检索。其难点在于顶级本体向个性化本体进化,以及基于概念集和“Is-A”关系实现语义相似度量的方法。通过系统的初步实现与相关实验的验证,该模型的检索准确度可达88.6%,明显高于传统的基于关键字和基于通用本体的图像检索,实现了图像智能检索功能。     

一种基于本体的概念相似度计算及其应用

概念的语义相似度研究,是知识表示以及信息检索领域中的一个重要内容。本文提出了基于语义相似度和相关度的综合概念相似度计算方法,考虑了语义距离和本体库特征,加入概念的信息重合度、概念的深度、概念的密度和不对称因子的辅助影响。通过实验和两种传统的语义相似度计算方法进行对比,本方法能更好地区分本体树中不同关系的概念对,验证了该方法的有效性。     

基于高斯过程分类器的三维模型多粒度语义检索

为解决三维模型语义检索中用户检索意图不一致问题,建立多粒度语义检索框架,使学习模型能够有效地适应用户的不同检索意图。首先对模型分类知识进行层次划分,形成语义概念的多粒度结构。然后提取一种多视图特征来描述三维模型的形状特性,并采用高斯过程分类器建立不同粒度层次上的学习模型,实现低层特征和查询概念之间的语义一致性描述。和已有研究相比,多粒度语义检索框架使用户可通过语义粒度级别变化进行检索意图设置,从而检索结果尽可能符合用户语义。在实验部分,采用三维模型基准数据库对框架进行算法性能测试。结果表明,检索准确率要明显提高,并且符合人类思维特点。     

BPMN 2.0编排的形式语义和分析

BPMN 2.0编排已成为描述业务流程间交互事实上的标准.BPMN 2.0编排面向流的特征,使之会产生控制流方面的语义错误.因此,检查编排语义正确性是BPMN 2.0编排建模工具所期望具有的功能.但是,BPMN 2.0标准规约中编排缺少形式语义及相应的分析技术,这阻碍了对BPMN 2.0编排的语义分析.本文提出了一种映射,用于将BPMN 2.0编排转换为工作流网,使用Petri网来形式定义BPMN 2.0编排的语义.借助Petri网的分析技术,这种定义的语义可用来分析BPMN 2.0编排的结构和控制流方面的错误.该映射和语义分析已被实现为一种工具.实验结果表明,这种形式化可以识别BPM AI过程模型库中编排的语义错误.     

An operator net model for distributed systems

An operator net is a graph consisting of nodes and directed arcs. While operator nets are syntactically similar to dataflow nets, they completely separate the operational semantics from the mathematical semantics. In this paper we define an operational semantics for operator nets that intuitively corresponds to communication in a distributed system. The operational semantics of operator ator nets provide a formal model for a distributed system that is an intermediate point between the actual system and a mathematical model. Abstract properties are expressed using relations on events and messages of an operator net. Corresponding operational specifications can be written using Lucid equations that define a node as a mathematical function on infinite history sequences. The operational specifications are executable and can be easily transformed into a practical implementation of the system. Examples of such specifications are included in the paper.Janice Glasgow is an associate professor in the Department of Computing and Information Science at Queen's University. She received her M. Math and Ph.D. degrees from the University of Waterloo. Dr. Glasgow's current research interests include programming language semantics and logics for reasoning about programming.Glenn H. MacEwen r received the B.Eng. degree in electrical engineering from McGill University, Montreal, P.Q., Canada, in 1962 and the M.Sc. and Ph.D. degrees in computer science from the University of Toronto, Toronto, Ont., Canada in 1967 and 1971, respectively. Since 1970 he has been with the Department of Computing and Information Science at Queen's University, Kingston, Ontario. He served as Head from 1982 to 1987 and is currently a Professor in the department. He is also a director and consultant to Andyne Computing Limited. His research interests include software engineering, computer security, and real-time systems. Dr. MacEwen is a senior member of the Institute of Electrical and Electronics Engineers, and a member of the Association for Computing Machinery and the Association of Professional Engineers of Ontario.     

形式语言基于Monads的语义计算模型

传统形式语言的语义建模方法在语义解释与规则描述等语义计算方面存在不足,应用范畴论方法的Monads对形式语言的语义计算进行了研究。基于Monads构造Kleisli范畴,在Kleisli范畴的形式化框架内建立语义计算模型,并对该模型进行了应用。与传统语义建模方法相比,所提语义计算模型具有普适性,其语义解释与规则描述的能力更强。     

A Semantics for Distributed Execution of Statemate

We present a semantics for the statechart variant implemented in the Statemate product of i-Logix. Our semantics enables distributed code generation for Statemate models in the context of rapid prototyping for embedded control applications. We argue that it seems impossible to efficiently generate distributed code using the original Statemate semantics. The new, distributed semantics has the advantages that, first, it enables the generation of efficient distributed code, second, it preserves many aspects of the original semantics for those parts of a model that are not distributed, and third, the changes made regarding the interaction of distributed model parts are similar to the interaction between the model and its environment in the original semantics, thus giving designers a familiar execution model. The semantics has been implemented in Grace, a framework for rapid prototyping code generation for embedded control applications.     

Enhanced Topic-based Vector Space Model for semantics-aware spam filtering

Spam has become a major issue in computer security because it is a channel for threats such as computer viruses, worms and phishing. More than 85% of received e-mails are spam. Historical approaches to combat these messages including simple techniques such as sender blacklisting or the use of e-mail signatures, are no longer completely reliable. Currently, many solutions feature machine-learning algorithms trained using statistical representations of the terms that usually appear in the e-mails. Still, these methods are merely syntactic and are unable to account for the underlying semantics of terms within the messages. In this paper, we explore the use of semantics in spam filtering by representing e-mails with a recently introduced Information Retrieval model: the enhanced Topic-based Vector Space Model (eTVSM). This model is capable of representing linguistic phenomena using a semantic ontology. Based upon this representation, we apply several well-known machine-learning models and show that the proposed method can detect the internal semantics of spam messages.     

基于深度卷积神经网络的三维模型检索

为了更有效地利用三维模型数据集进行特征的自主学习,提出一种使用自然图像作为输入源,以三维模型的较优视图集为基础,通过深度卷积神经网络的训练提取深度特征用于检索的三维模型检索方法。首先,从多个视点对三维模型进行视图提取,并根据灰度熵的排序选取较优视图;然后,通过深度卷积神经网络对视图集进行训练,从而提取较优视图的深度特征并进行降维,同时,对输入的自然图像提取边缘轮廓图,经过相似度匹配获得一组三维模型;最后,基于检索结果中同类模型总数占检索列表长度的比例对列表进行重排序,从而获得最终的检索结果。实验结果表明,该算法能够有效利用深度卷积神经网络对三维模型的视图进行深度特征提取,同时降低了输入源的获取难度,有效提高了检索效果。     

Extracting 3D model feature lines based on conditional random fields

We propose a 3D model feature line extraction method using templates for guidance. The 3D model is first projected into a depth map, and a set of candidate feature points are extracted. Then, a conditional random fields (CRF) model is established to match the sketch points and the candidate feature points. Using sketch strokes, the candidate feature points can then be connected to obtain the feature lines, and using a CRF-matching model, the 2D image shape similarity features and 3D model geometric features can be effectively integrated. Finally, a relational metric based on shape and topological similarity is proposed to evaluate the matching results, and an iterative matching process is applied to obtain the globally optimized model feature lines. Experimental results showed that the proposed method can extract sound 3D model feature lines which correspond to the initial sketch template.     

A Hierarchy of SOS Rule Formats

In 1981 Structural Operational Semantics (SOS) was introduced as a systematic way to define operational semantics of programming languages by a set of rules of a certain shape [G.D. Plotkin. A structural approach to operational semantics. Technical Report DAIMI FN-19, Computer Science Department, Aarhus University, Aarhus, Denmark, September 1981. Also published in: Journal of Logic and Algebraic Programming 60–61 (2004) 17–140]. Subsequently, the format of SOS rules became the object of study. Using so-called Transition System Specifications (TSS's) several authors syntactically restricted the format of rules and showed several useful properties about the semantics induced by any TSS adhering to the format. This has resulted in a line of research proposing several syntactical rule formats and associated meta-theorems. Properties that are guaranteed by such rule formats range from well-definedness of the operational semantics and compositionality of behavioral equivalences to security- and probability-related issues. In this paper, we provide an initial hierarchy of SOS rules formats and meta-theorems formulated around them.     

基于自然语言处理的中文科技论文特征提取研究

随着中国科技、经济的迅速发展,中文科技论文大量以电子文本的形式存在.因此,中文科技论文文本分类研究已经成为信息处理领域中的一个热门话题.本文建立了基于自然语言处理的文本特征提取模型,并从语法、语义和语用三个层面进行了详细地分析.