Approximate OLAP of document-oriented databases: A variety-aware approach

Schemaless databases, and document-oriented databases in particular, are preferred to relational ones for storing heterogeneous data with variable schemas and structural forms. However, the absence of a unique schema adds complexity to analytical applications, in which a single analysis often involves large sets of data with different schemas. In this paper we propose an original approach to OLAP on collections stored in document-oriented databases. The basic idea is to stop fighting against schema variety and welcome it as an inherent source of information wealth in schemaless sources. Our approach builds on four stages: schema extraction, schema integration, FD enrichment, and querying; these stages are discussed in detail in the paper. To make users aware of the impact of schema variety, we propose a set of indicators inspired by the definition of attribute density. Finally, we experimentally evaluate our approach in terms of efficiency and effectiveness.     

Interschema correspondence establishment in a cooperative OWL-based multi-information server grid environment

Establishing interschema semantic knowledge between corresponding elements in a cooperating OWL-based multi-information server grid environment requires deep knowledge, not only about the structure of the data represented in each server, but also about the commonly occurring differences in the intended semantics of this data. The same information could be represented in various incompatible structures, and more importantly the same structure could be used to represent data with many diverse and incompatible semantics. In a grid environment interschema semantic knowledge can only be detected if both the structural and semantic properties of the schemas of the cooperating servers are made explicit and formally represented in a way that a computer system can process. Unfortunately, very often there is lack of such knowledge and the underlying grid information servers (ISs) schemas, being semantically weak as a consequence of the limited expressiveness of traditional data models, do not help the acquisition of this knowledge. The solution to overcome this limitation is primarily to upgrade the semantic level of the IS local schemas through a semantic enrichment process by augmenting the local schemas of grid ISs to semantically enriched schema models, then to use these models in detecting and representing correspondences between classes belonging to different schemas. In this paper, we investigate the possibility of using OWL-based domain ontologies both for building semantically rich schema models, and for expressing interschema knowledge and reasoning about it. We believe that the use of OWL/RDF in this setting has two important advantages. On the one hand, it enables a semantic approach for interschema knowledge specification, by concentrating on expressing conceptual and semantic correspondences between both the conceptual (intensional) definition and the set of instances (extension) of classes represented in different schemas. On the other hand, it is exactly this semantic nature of our approach that allows us to devise reasoning mechanisms for discovering and reusing interschema knowledge when the need arises to compare and combine it.     

Is that my hand? An egocentric dataset for hand disambiguation

With the recent development of wearable cameras, the interest for research on the egocentric perspective is increasing. This opens the possibility to work on a specific object detection problem of hand detection and hand disambiguation. However, recent progress in egocentric hand disambiguation and even hand detection, especially using deep learning, has been limited by the lack of a large dataset, with suitable variations in subject, activity, and scene. In this paper, we propose a dataset that simulates daily activities, with variable illumination and people from different cultures and ethnicity to address daily life conditions. We increase the dataset size from previous works to allow robust solutions like deep neural networks that need a substantial amount of data for training. Our dataset consists of 50,000 annotated images with 10 different subjects doing 5 different daily activities (biking, eating, kitchen, office and running) in over 40 different scenes with variable illumination and changing backgrounds, and we compare with previous similar datasets.Hands in an egocentric view are challenging to detect due to a number of factors, such as shape variations, inconsistent illumination, motion blur, and occlusion. To improve hand detection and disambiguation, context information can be included to aid in the detection. In particular, we propose three neural network architectures that jointly learn the hand and context information, and we provide baseline results with current object/hand detection approaches.     

Symbolic connectionism in natural language disambiguation

Natural language understanding involves the simultaneous consideration of a large number of different sources of information. Traditional methods employed in language analysis have focused on developing powerful formalisms to represent syntactic or semantic structures along with rules for transforming language into these formalisms. However, they make use of only small subsets of knowledge. This article describes how to use the whole range of information through a neurosymbolic architecture which is a hybridization of a symbolic network and subsymbol vectors generated from a connectionist network. Besides initializing the symbolic network with prior knowledge, the subsymbol vectors are used to enhance the system's capability in disambiguation and provide flexibility in sentence understanding. The model captures a diversity of information including word associations, syntactic restrictions, case-role expectations, semantic rules and context. It attains highly interactive processing by representing knowledge in an associative network on which actual semantic inferences are performed. An integrated use of previously analyzed sentences in understanding is another important feature of our model. The model dynamically selects one hypothesis among multiple hypotheses. This notion is supported by three simulations which show the degree of disambiguation relies both on the amount of linguistic rules and the semantic-associative information available to support the inference processes in natural language understanding. Unlike many similar systems, our hybrid system is more sophisticated in tackling language disambiguation problems by using linguistic clues from disparate sources as well as modeling context effects into the sentence analysis. It is potentially more powerful than any systems relying on one processing paradigm     

Evaluation of Linguistic Features for Word Sense Disambiguation with Self-Organized Document Maps

Word sense disambiguation automatically determines the appropriate senses of a word in context. We have previously shown that self-organized document maps have properties similar to a large-scale semantic structure that is useful for word sense disambiguation. This work evaluates the impact of different linguistic features on self-organized document maps for word sense disambiguation. The features evaluated are various qualitative features, e.g. part-of-speech and syntactic labels, and quantitative features, e.g. cut-off levels for word frequency. It is shown that linguistic features help make contextual information explicit. If the training corpus is large even contextually weak features, such as base forms, will act in concert to produce sense distinctions in a statistically significant way. However, the most important features are syntactic dependency relations and base forms annotated with part of speech or syntactic labels. We achieve 62.9% ± 0.73% correct results on the fine grained lexical task of the English SENSEVAL-2 data. On the 96.7% of the test cases which need no back-off to the most frequent sense we achieve 65.7% correct results.     

基于多知识源的词汇消歧一体化处理

词汇消歧是语言分析的基石,本文提出一种基于多知识源的词汇消歧一体化处理机制,该机制充分利用了知识库和文本结构的信息,以句法标签、词频、搭配、上下文语义,语义可选约束,句法线索等知识源为消歧指示器     

Personalised code generation from large schema sets for geospatial mobile applications

XML and XML Schema are used in the geospatial domain for the definition of standards that enhance the interoperability between producers and consumers of spatial data. The size and complexity of these geospatial standards and their associated schemas have been growing with time reaching levels of complexity that make it difficult to build systems based on them in a timely and cost-effective manner. The problem of producing XML processing code based on large schemas has been traditionally solved by using XML data binding generators. Unfortunately, this solution is not always effective when code is generated for resource-constrained devices, such as mobile phones. Large and complex schemas often result in the production of code with a large size and a complicated structure that might not fit the device limitations. In this article we present the instance-based XML data binding approach to produce more compact application-specific XML processing code for geospatial applications targeted to mobile devices. The approach tries to reduce the size and complexity of the generated code by using information about how schemas are used by individual applications. Our experimental results suggest a significant simplification of XML Schema sets to the real needs of client applications accompanied by a substantial reduction of size of the generated code.     

Mapping DTDs to relational schemas with semantic constraints

XML is becoming a prevalent format and standard for data exchange in many applications. With the increase of XML data, there is an urgent need to research some efficient methods to store and manage XML data. As relational databases are the primary choices for this purpose considering their data management power, it is necessary to research the problem of mapping XML schemas to relational schemas. The semantics of XML schemas are crucial to design, query, and store XML documents and functional dependencies are very important representations of semantic information of XML schemas. As DTDs are one of the most frequently used schemas for XML documents in these days, we will use DTDs as schemas of XML documents here. This paper proposes the concept and the formal definition of XML functional dependencies over DTDs. A method to map XML DTDs to relational schemas with constraints such as functional dependencies, domain constraints, choice constraints, reference constraints, and cardinality constraints over DTDs is given, which can preserve the structures of DTDs as well as the semantics implied by the above constraints over DTDs. The concepts and method of mapping DTDs to relational schemas presented in the paper can be extended to the field of XML Schema just with some modifications in related formal definitions.     

一种新的重名消解算法在保险领域中的应用研究*

研究客户重名消解问题。针对以往重名消解方法如文本聚类的方法需考虑大量无用词汇并需要人工设定阈值以及类别数量,而基于信息抽取的人物相关属性相似度方法对于人物信息的抽取具有依赖性,提出了一种改进的重名消解算法。该算法首先对具有相同标志的客户进行属性匹配,合并匹配成功的标志;然后进行链接分析,对客户合作网的结构进行分析,将具有相同标志并与同一个代理人实体合作的客户归为一个客户实体,并把具有相同合作对的信息加以分析合并;最后通过原子团簇分析法进行聚类分析。仿真实验结果表明,所提改进算法对中文字符串的匹配处理进行了优化,执行效率高,适合于以大量数据为特征的保险领域的重名消解。     

基于《知网》的中文信息结构消歧研究

《中文信息结构库》是《知网》的重要组成部分之一,可以作为中文语义分析的规则库,对其进行消歧是实际应用的基础之一。因此,该文首先对中文信息结构进行了形式化描述;接着对其进行优先级划分;然后根据其构成形式提出了四种不同的消歧方法 即词性序列消歧法、图相容匹配消歧法、图相容度计算消歧法、基于实例的语义相似度计算消歧法;最后针对不同优先级的中文信息结构集设计了不同消歧流程。实验结果证明消歧正确率达到了90% 以上。     

基于组合特征的Web人名消歧方法

重名问题在Web人物搜索过程中是很普遍的现象.研究了Web人名消歧相关问题,提取与待消歧人名相关的不同特征集,运用向量空间模型构造人物实体的组合特征,最后通过层次聚类算法将相似度高的文档优先聚类,由此实现人名消歧.在WePS数据集上的实验结果表明,提出的方法具有良好的消歧效果.     

利用海量知识库实现实体标注的一种方法

互联网上聚集了大量的文本、图像等非结构化信息,RDF作为W3C提出的互联网上的资源描述框架,非常适合于描述网络上的非结构化信息,因此形成了大量的RDF知识库,如Freebase、Yago、DBPedia等。RDF知识库中包含丰富的语义信息,可以对来自网页的名字实体进行标注,实现语义扩充。将网页上的名字实体映射到知识库中对应实体上称作实体标注。实体标注包括两个主要部分:实体间的映射和标注去歧义。利用海量RDF知识库的特性,提出了一种有效的实体标注方法。该方法采用简单的图加权及计算解决实体标注的去歧义问题。该方法已在云平台上实现,并通过实验验证了其准确度和可扩展性。     

MDSM: Microarray database schema matching using the Hungarian method

Current microarray databases use different terminologies and structures and thereby limit the sharing of data and collating of results between laboratories. Consequently, an effective integrated microarray data model is required. One important process to develop such an integrated database is schema matching. In this paper, we propose an effective schema matching approach called MDSM, to syntactically and semantically map attributes of different microarray schemas. The contribution from this work will be used later to create microarray global schemas. Since microarray data is complex, we use microarray ontology to improve the measuring accuracy of the similarity between attributes. The similarity relations can be represented as weighted bipartite graphs. We determine the best schema matching by computing the optimal matching in a bipartite graph using the Hungarian optimisation method. Experimental results show that our schema matching approach is effective and flexible to use in different kinds of database models such as; database schema, XML schema, and web site map. Finally, a case study on an existing public microarray schema is carried out using the proposed method.     

基于SVM的离合词词义消歧

离合词词义消歧要解决如何让计算机理解离合词中的歧义词在具体上下文中的含义。针对离合词中歧义词在机器翻译中造成的对照翻译不准确以及在信息检索中无法匹配有效信息等问题,将词义消歧的方法应用于离合词中的歧义词,采用SVM模型建立分类器。为了提高离合词词义消歧的正确率,在提取特征时,结合离合词的特点,不仅提取了歧义词上下文中的局部词、局部词性、局部词及词性3类特征,还提取了“离”形式的歧义词的中间插入部分的特征;将文本特征转换为特征向量时,对布尔权重法进行了改进,依次固定某种类型特征权重,分别改变另外两种类型特征权重的消歧正确率来验证3类特征的消歧效果。实验结果表明,局部词特征、局部词及词性特征对消歧效果的影响高于局部词性特征,且采用不同类型的特征权重与采用相同的权重相比,消歧正确率提高了1.03%~5.69%。     

Double-layered schema integration of heterogeneous XML sources

Schema integration aims to create a mediated schema as a unified representation of existing heterogeneous sources sharing a common application domain. These sources have been increasingly written in XML due to its versatility and expressive power. Unfortunately, these sources often use different elements and structures to express the same concepts and relations, thus causing substantial semantic and structural conflicts. Such a challenge impedes the creation of high-quality mediated schemas and has not been adequately addressed by existing integration methods. In this paper, we propose a novel method, named XINTOR, for automating the integration of heterogeneous schemas. Given a set of XML sources and a set of correspondences between the source schemas, our method aims to create a complete and minimal mediated schema: it completely captures all of the concepts and relations in the sources without duplication, provided that the concepts do not overlap. Our contributions are fourfold. First, we resolve structural conflicts inherent in the source schemas. Second, we introduce a new statistics-based measure, called path cohesion, for selecting concepts and relations to be a part of the mediated schema. The path cohesion is statistically computed based on multiple path quality dimensions such as average path length and path frequency. Third, we resolve semantic conflicts by augmenting the semantics of similar concepts with context-dependent information. Finally, we propose a novel double-layered mediated schema to retain a wider range of concepts and relations than existing mediated schemas, which are at best either complete or minimal, but not both. Performed on both real and synthetic datasets, our experimental results show that XINTOR outperforms existing methods with respect to (i) the mediated-schema quality using precision, recall, F-measure, and schema minimality; and (ii) the execution performance based on execution time and scale-up performance.     

基于主题关系的中文短文本图模型实体消歧

实体消歧作为知识库构建、信息检索等应用的重要支撑技术,在自然语言处理领域有着重要的作用。然而在短文本环境中,对实体的上下文特征进行建模的传统消歧方式很难提取到足够多用以消歧的特征。针对短文本的特点,提出一种基于实体主题关系的中文短文本图模型消歧方法,首先,通过TextRank算法对知识库信息构建的语料库进行主题推断,并使用主题推断的结果作为实体间关系的表示;然后,结合基于BERT的语义匹配模型给出的消歧评分对待消歧文本构建消歧网络图;最终,通过搜索排序得出最后的消歧结果。使用CCKS2020短文本实体链接任务提供的数据集对所提方法进行评测,实验结果表明,该方法对短文本的实体消歧效果优于其他方法,能有效解决在缺乏知识库实体关系情况下的中文短文本实体消歧问题。     

Modeling and manipulating the structure of hierarchical schemas for the web

The Semantic Web is the next step of the current Web where information will become more machine-understandable to support effective data discovery and integration. Hierarchical schemas, either in the form of tree-like structures (e.g., DTDs, XML schemas), or in the form of hierarchies on a category/subcategory basis (e.g., thematic hierarchies of portal catalogs), play an important role in this task. They are used to enrich semantically the available information. Up to now, hierarchical schemas have been treated rather as sets of individual elements, acting as semantic guides for browsing or querying data. Under that view, queries like “find the part of a portal catalog which is not present in another catalog” can be answered only in a procedural way, specifying which nodes to select and how to get them. For this reason, we argue that hierarchical schemas should be treated as full-fledged objects so as to allow for their manipulation. This work proposes models and operators to manipulate the structural information of hierarchies, considering them as first-class citizens. First, we explore the algebraic properties of trees representing hierarchies, and define a lattice algebraic structure on them. Then, turning this structure into a boolean algebra, we present the operators S-union, S-intersection and S-difference to support structural manipulation of hierarchies. These operators have certain algebraic properties to provide clear semantics and assist the transformation, simplification and optimization of sequences of operations using laws similar to those of set theory. Also, we identify the conditions under which this framework is applicable. Finally, we demonstrate an application of our framework for manipulating hierarchical schemas on tree-like hierarchies encoded as RDF/s files.     

一种基于全局领域和短期记忆因子的图模型

词义消歧是一项具有挑战性的自然语言处理难题。作为词义消歧中的一种优秀的半监督消歧算法,遗传蚁群词义消歧算法能快速进行全文词义消歧。该算法采用了一种局部上下文的图模型来表示语义关系,以此进行词义消歧。然而,在消歧过程中却丢失了全局语义信息,出现了消歧结果冲突的问题,导致算法精度降低。因此, 提出了一种基于全局领域和短期记忆因子改进的图模型来表示语义以解决这个问题。该图模型引入了全局领域信息,增强了图对全局语义信息的处理能力。同时根据人的短期记忆原理,在模型中引入了短期记忆因子,增强了语义间的线性关系,避免了消歧结果冲突对词义消歧的影响。大量实验结果表明：与经典词义消歧算法相比,所提的改进图模型提高了词义消歧的精度。