Modeling and querying graphical representations of statistical data

Although pictorial renditions of statistical data are ubiquitous, few techniques and standards exist to exchange, search and query these graphical representations. We present several improvements to human–graph interaction including (i) a new approach to manage statistical graph knowledge by semantic annotation of graphs that bridges the gap between Web 2.0 social tagging and formal, logic-based approaches, (ii) knowledge management and discovery across a non-trivial graph knowledge base and (iii) sophisticated question answering that requires background knowledge.     

An empirical study of representing adjectives over knowledge bases: Approach,lexicon and application

Adjectives are common in natural language, and their usage and semantics have been studied broadly. In recent years, with the rapid growth of knowledge bases (KBs), many knowledge-based question answering (KBQA) systems are developed to answer users’ natural language questions over KBs. A fundamental task of such systems is to transform natural language questions into structural queries, e.g., SPARQL queries. Thus, such systems require knowledge about how natural language expressions are represented in KBs, including adjectives. In this paper, we specifically address the problem of representing adjectives over KBs. We propose a novel approach, called Adj2SP, to represent adjectives as SPARQL query patterns. Adj2SP contains a statistic-based approach and a neural network-based approach, both of them can effectively reduce the search space for adjective representations and overcome the lexical gap between input adjectives and their target representations. Two adjective representation datasets are built for evaluation, with adjectives used in QALD and Yahoo! Answers, as well as their representations over DBpedia. Experimental results show that Adj2SP can generate representations of high quality and significantly outperform several alternative approaches in F1-score. Furthermore, we publish Lark, a lexicon for adjective representations over KBs. Current KBQA systems show an improvement of over 24% in F1-score by integrating Adj2SP.     

Fast graph similarity search via hashing and its application on image retrieval

Similarity search in graph databases has been widely investigated. It is worthwhile to develop a fast algorithm to support similarity search in large-scale graph databases. In this paper, we investigate a k-NN (k-Nearest Neighbor) similarity search problem by locality sensitive hashing (LSH). We propose an innovative fast graph search algorithm named LSH-GSS, which first transforms complex graphs into vectorial representations based on prototypes in the database and later accelerates a query in Euclidean space by employing LSH. Because images can be represented as attributed graphs, we propose an approach to transform attributed graphs into n-dimensional vectors and apply LSH-GSS to execute further image retrieval. Experiments on three real graph datasets and two image datasets show that our methods are highly accurate and efficient.     

Extending SPARQL with regular expression patterns (for querying RDF)

RDF is a knowledge representation language dedicated to the annotation of resources within the framework of the semantic web. Among the query languages for RDF, SPARQL allows querying RDF through graph patterns, i.e., RDF graphs involving variables. Other languages, inspired by the work in databases, use regular expressions for searching paths in RDF graphs. Each approach can express queries that are out of reach of the other one. Hence, we aim at combining these two approaches. For that purpose, we define a language, called PRDF (for “Path RDF”) which extends RDF such that the arcs of a graph can be labeled by regular expression patterns. We provide PRDF with a semantics extending that of RDF, and propose a correct and complete algorithm which, by computing a particular graph homomorphism, decides the consequence between an RDF graph and a PRDF graph. We then define the PSPARQL query language, extending SPARQL with PRDF graph patterns and complying with RDF model theoretic semantics. PRDF thus offers both graph patterns and path expressions. We show that this extension does not increase the computational complexity of SPARQL and, based on the proposed algorithm, we have implemented a correct and complete PSPARQL query engine.     

一种基于树分解的图上点区间编码方法及应用

根据图上节点所在位置与邻居节点特征,可以使用不同策略为每个图上节点进行区间编码,基于区间编码,许多在大型图上的应用如知识图谱查询、智能问答等的处理可以加速或得到准确性上的提升。针对此种情况,提出一种基于树分解算法的图上点区间编码方法,并在大型知识图谱上通过智能问答歧义消除的应用验证该方法的有效性。实验结果表明,该方法能够有效地表达出图上节点的位置特征,并帮助智能问答中的实体消除歧义。     

Health-guided recipe recommendation over knowledge graphs

While the availability of large-scale online recipe collections presents opportunities for health consumers to access a wide variety of recipes, it can be challenging for them to discover relevant recipes. Whereas most recommender systems are designed to offer selections consistent with users’ past behavior, it remains an open problem to offer selections that can help users’ transition from one type of behavior to another, intentionally. In this paper, we introduce health-guided recipe recommendation as a way to incrementally shift users towards healthier recipe options while respecting the preferences reflected in their past choices. Introducing a knowledge graph (KG) into recommender systems as side information has attracted great interest, but its use in recipe recommendation has not been studied. To fill this gap, we consider the task of recipe recommendation over knowledge graphs. In particular, we jointly learn recipe representations via graph neural networks over two graphs extracted from a large-scale Food KG, which capture different semantic relationships, namely, user preferences and recipe healthiness, respectively. To integrate the nutritional aspects into recipe representations and the recommendation task, instead of simple fusion, we utilize a knowledge transfer scheme to enable the transfer of useful semantic information across the preferences and healthiness aspects. Experimental results on two large real-world recipe datasets showcase our model’s ability to recommend tasty as well as healthy recipes to users.     

Instance-sequence reasoning for video question answering

Video question answering (Video QA) involves a thorough understanding of video content and question language, as well as the grounding of the textual semantic to the visual content of videos. Thus, to answer the questions more accurately, not only the semantic entity should be associated with certain visual instance in video frames, but also the action or event in the question should be localized to a corresponding temporal slot. It turns out to be a more challenging task that requires the ability of conducting reasoning with correlations between instances along temporal frames. In this paper, we propose an instance-sequence reasoning network for video question answering with instance grounding and temporal localization. In our model, both visual instances and textual representations are firstly embedded into graph nodes, which benefits the integration of intra- and inter-modality. Then, we propose graph causal convolution (GCC) on graph-structured sequence with a large receptive field to capture more causal connections, which is vital for visual grounding and instance-sequence reasoning. Finally, we evaluate our model on TVQA+ dataset, which contains the groundtruth of instance grounding and temporal localization, three other Video QA datasets and three multimodal language processing datasets. Extensive experiments demonstrate the effectiveness and generalization of the proposed method. Specifically, our method outperforms the state-of-the-art methods on these benchmarks.     

Mining Semantic Correlation of Heterogeneous Multimedia Data for Cross-Media Retrieval

Although multimedia objects such as images, audios and texts are of different modalities, there are a great amount of semantic correlations among them. In this paper, we propose a method of transductive learning to mine the semantic correlations among media objects of different modalities so that to achieve the cross-media retrieval. Cross-media retrieval is a new kind of searching technology by which the query examples and the returned results can be of different modalities, e.g., to query images by an example of audio. First, according to the media objects features and their co-existence information, we construct a uniform cross-media correlation graph, in which media objects of different modalities are represented uniformly. To perform the cross-media retrieval, a positive score is assigned to the query example; the score spreads along the graph and media objects of target modality or MMDs with the highest scores are returned. To boost the retrieval performance, we also propose different approaches of long-term and short-term relevance feedback to mine the information contained in the positive and negative examples.     

图依赖研究与应用综述

图依赖是用于解决图数据的数据一致性问题的数据质量规则。基于图依赖提升数据一致性的过程通常分为图依赖定义与形式化、图依赖自动挖掘、基于图依赖的数据一致性提升三步。介绍了针对数据一致性的图依赖理论,并根据拓展类型将图依赖分为基于结构约束拓展、基于语义约束拓展和基于外部约束拓展的图依赖;综述并对比了从图数据中自动挖掘图依赖及其拓展的算法;分析了应用图依赖提高数据一致性的研究现状;总结了当前研究中仍存在的问题,并依据问题展望了图依赖在数据质量领域的应用前景。     

基于概念图的中文FAQ问答系统

提出一种利用概念图计算问句相似度的方法,并在此基础上实现基于概念图的中文FAQ问答系统,在该系统中采用概念图的形式表示用户问句及在FAQ库中找到的候选问句集中的问句,通过改进的概念图语义相似度计算问句相似度,在候选问句集中找到相似的问句并将答案返回给用户。该系统能够自动更新和维护FAQ库。实验结果表明,与基于关键词的句子相似度相比,基于语义的句子相似度提高了问题匹配的准确率。     

Domain structure-based transfer learning for cross-domain word representation

Cross-domain word representation aims to learn high-quality semantic representations in an under-resourced domain by leveraging information in a resourceful domain. However, most existing methods mainly transfer the semantics of common words across domains, ignoring the semantic relations among domain-specific words. In this paper, we propose a domain structure-based transfer learning method to learn cross-domain representations by leveraging the relations among domain-specific words. To accomplish this, we first construct a semantic graph to capture the latent domain structure using domain-specific co-occurrence information. Then, in the domain adaptation process, beyond domain alignment, we employ Laplacian Eigenmaps to ensure the domain structure is consistently distributed in the learned embedding space. As such, the learned cross-domain word representations not only capture shared semantics across domains, but also maintain the latent domain structure. We performed extensive experiments on two tasks, namely sentiment analysis and query expansion. The experiment results show the effectiveness of our method for tasks in under-resourced domains.     

基于锁增广分段图的多线程程序死锁检测

死锁是并行程序常见的缺陷之一,动态死锁分析方法根据程序运行轨迹构建锁图、分段图等模型来检测死锁.然而,锁图及其现有的各种变型无法区分同一循环中锁授权语句的多次执行,扩展锁图中记录的锁集无法捕捉线程曾经持有而又随后释放的锁信息,分段图无法刻画锁的获取和释放操作与线程启动操作耦合而导致的段间依赖关系,上述问题导致了多种死锁的误报.为解决上述问题,本文对已有的锁图和分段图模型进行改进,在锁图基础上扩充语句的执行时序信息,在分段图的基础上扩充锁的获取和释放信息,对段进行更细粒度地划分以建模锁对象导致的段间依赖关系;最终,在上述锁增广分段图与时序增广锁图的基础上,提出一种新的死锁检测方法.所提方法能有效消除前述各种误报,从而提高死锁检测的准确率.文中开发相应的原型系统,并结合多个程序实例对所提方法的有效性进行评估验证.     

双特征空间的实体排序学习

随着大规模知识图谱的出现以及企业高效管理领域知识图谱的需求,知识图谱中的自组织实体检索成为研究热点。给定知识图谱以及用户查询,实体检索的目标在于从给定的知识图谱中返回实体的排序列表。从匹配的角度来看,传统的实体检索模型大都将用户查询和实体统一映射到词的特征空间。这样做具有明显的缺点,例如,将同属于一个实体的两个词视为独立的。为此,该文提出将用户查询和实体同时映射到实体与词两个特征空间方法,称为双特征空间的排序学习。首先将实体抽象成若干个域。之后从词空间和实体空间两个维度分别抽取排序特征,最终应用于排序学习算法中。实验结果表明,在标准数据集上,双特征空间的实体排序学习模型性能显著优于当前先进的实体检索模型。     

Adding inclusion dependencies to an object-oriented data model with uniqueness constraints

We study an object-oriented data model that allows to express both uniqueness constraints and inclusion dependencies as semantic constraints. The data model is based on a subset of F-logic. Uniqueness constraints comprise path functional dependencies which generalise functional dependencies and reflect the navigational power of object-oriented query languages. As inclusion dependencies, we consider explicit class inclusion constraints, besides inclusions required by class hierarchies, and onto constraints that enforce reachability of objects. For these classes of semantic constraints we present an axiomatisation and prove its inference rules to be correct and complete with respect to general logical implication, leaving the decision problem open. The completeness proof combines the known construction for path functional dependencies alone with a possibly infinite model generation process to enforce onto constraints. The results prepare the grounds for normal forms in object-oriented data models and subsequently for computer aided object-oriented database design, following the decomposition approach for the relational data model. Beyond the application for schema design, the achievements could also be exploited for related tasks like semantic query optimisation and mediated data integration within a variety of graph based data models. Received: 11 October 2000 / 27 January 2003     

基于跨模态自蒸馏的零样本草图检索

零样本草图检索将未见类的草图作为查询样本,用于检索未见类的图像。因此,这个任务同时面临两个挑战：草图和图像之间的模态差异以及可见类和未见类的不一致性。过去的方法通过将草图和图像投射到一个公共空间来消除模态差异,还通过利用语义嵌入（如词向量和词相似度）来弥合可见类和未见类的语义不一致。在本文中,我们提出了跨模态自蒸馏方法,从知识蒸馏的角度研究可泛化的特征,无需语义嵌入参与训练。具体而言,我们首先通过传统的知识蒸馏将预训练的图像识别网络的知识迁移到学生网络。然后,通过草图和图像的跨模态相关性,跨模态自蒸馏将上述知识间接地迁移到草图模态的识别上,提升草图特征的判别性和泛化性。为了进一步提升知识在草图模态内的集成和传播,我们进一步地提出草图自蒸馏。通过为数据学习辨别性的且泛化的特征,学生网络消除了模态差异和语义不一致性。我们在三个基准数据集,即Sketchy、TU-Berlin和QuickDraw,进行了广泛的实验,证明了我们提出的跨模态自蒸馏方法与当前方法相比较的优越性。     

A survey on question answering technology from an information retrieval perspective

This article provides a comprehensive and comparative overview of question answering technology. It presents the question answering task from an information retrieval perspective and emphasises the importance of retrieval models, i.e., representations of queries and information documents, and retrieval functions which are used for estimating the relevance between a query and an answer candidate. The survey suggests a general question answering architecture that steadily increases the complexity of the representation level of questions and information objects. On the one hand, natural language queries are reduced to keyword-based searches, on the other hand, knowledge bases are queried with structured or logical queries obtained from the natural language questions, and answers are obtained through reasoning. We discuss different levels of processing yielding bag-of-words-based and more complex representations integrating part-of-speech tags, classification of the expected answer type, semantic roles, discourse analysis, translation into a SQL-like language and logical representations.     

Natural language technology and query expansion: issues, state-of-the-art and perspectives

The availability of an abundance of knowledge sources has spurred a large amount of effort in the development and enhancement of Information Retrieval techniques. Users’ information needs are expressed in natural language and successful retrieval is very much dependent on the effective communication of the intended purpose. Natural language queries consist of multiple linguistic features which serve to represent the intended search goal. Linguistic characteristics that cause semantic ambiguity and misinterpretation of queries as well as additional factors such as the lack of familiarity with the search environment affect the users’ ability to accurately represent their information needs, coined by the concept “intention gap”. The latter directly affects the relevance of the returned search results which may not be to the users’ satisfaction and therefore is a major issue impacting the effectiveness of information retrieval systems. Central to our discussion is the identification of the significant constituents that characterize the query intent and their enrichment through the addition of meaningful terms, phrases or even latent representations, either manually or automatically to capture their intended meaning. Specifically, we discuss techniques to achieve the enrichment and in particular those utilizing the information gathered from statistical processing of term dependencies within a document corpus or from external knowledge sources such as ontologies. We lay down the anatomy of a generic linguistic based query expansion framework and propose its module-based decomposition, covering topical issues from query processing, information retrieval, computational linguistics and ontology engineering. For each of the modules we review state-of-the-art solutions in the literature categorized and analyzed under the light of the techniques used.     

基于改进多嵌入空间的实时语义数据流推理

将语义数据流处理引擎与知识图谱嵌入表示学习相结合,可以有效提高实时数据流推理查询性能,但是现有的知识表示学习模型更多关注静态知识图谱嵌入,忽略了知识图谱的动态特性,导致难以应用于实时动态语义数据流推理任务。为了使知识表示学习模型适应知识图谱的在线更新并能够应用于语义数据流引擎,建立一种基于改进多嵌入空间的动态知识图谱嵌入模型PUKALE。针对传递闭包等复杂推理场景,提出3种嵌入空间生成算法。为了在进行增量更新时更合理地选择嵌入空间,设计2种嵌入空间选择算法。基于上述算法实现PUKALE模型,并将其嵌入数据流推理引擎CSPARQL-engine中,以实现实时语义数据流推理查询。实验结果表明,与传统的CSPARQL和KALE推理相比,PUKALE模型的推理查询时间分别约降低85%和93%,其在支持动态图谱嵌入的同时能够提升实时语义数据流推理准确率。     

Querying multimedia presentations based on content

Considers the problem of querying multimedia presentations based on content information. Multimedia presentations are modeled as presentation graphs, which are directed acyclic graphs that visually specify the presentations. We present a graph data model for the specification of multimedia presentations and discuss query languages as effective tools to query and manipulate multimedia presentation graphs with respect to content information. To query the information flow throughout a multimedia presentation, as well as in each individual multimedia stream, we use revised versions of temporal operators Next, Connected and Until, together with path formulas. These constructs allow us to specify and query paths along a presentation graph. We present an icon-based graphical query language, GVISUAL, that provides iconic representations for these constructs and a user-friendly graphical interface for query specification. We also present an OQL-like language, GOQL (Graph OQL), with similar constructs, that allows textual and more traditional specifications of graph queries. Finally, we introduce GCalculus (Graph Calculus), a calculus-based language that establishes the formal grounds for the use of temporal operators in path formulas and for querying presentation graphs with respect to content information. We also discuss GCalculus/S (GCalculus with Sets) which avoids highly complex query expressions by eliminating the universal path quantifier, the negation operator and the universal quantifier. GCalculus/S represents the formal basis for GVISUAL, i.e. GVISUAL uses the constructs of GCalculus/S directly     

基于汉语框架网的旅游信息问答系统设计

该文借助汉语框架网(Chinese FrameNet,简称CFN)在语义表达方面的独特优势,探讨用本体描述语言建立面向特定领域的汉语框架语义知识库,并且以旅游交通领域中问答系统设计为例分析方法的有效性。方法中首先利用TREC分类与本体分类相结合的方式为查询问句分类,然后提出基于CFN的问句分析策略,通过CFN语义分析得到问句中三元组语义谓词、语义主体和语义客体,在问句分析的基础上从旅游本体知识库中对答案进行抽取并对答案处理,同时用本体编辑工具Protégé编码,实验证实方法是有效的。