Academic literature on the topic 'Instance matching'

Establishing correct links among the coreference ontology instances is critical to the success of Linked Open Data (LOD) cloud. However, because of the high level heterogeneity and large scale instance set, matching the coreference instances in LOD cloud is an error prone and time consuming task. To this end, in this work, we present an asymmetrical profile-based similarity measure for instance matching task, construct new optimal models for schema-level and instance-level matching problems, and propose a compact hybrid evolutionary algorithm based ontology matching approach to solve the large scale instance matching problem in LOD cloud. Finally, the experimental results of comprising our approach with the states of the art systems on the instance matching track of OAEI 2015 and real-world datasets show the effectiveness of our approach.

An instance of a size-n stable marriage problem involves n men and n women, each individually ranking all members of opposite sex in order of preference as a potential marriage partner. A complete matching, a set of n marriages, is called stable if no unmatched man and woman prefer each other to their partners in the matching. It is known that, for every instance of marriage partner preferences, there exists at least one stable matching, and that there are instances with exponentially many stable matchings. Our focus is on a random instance chosen uniformly from among all (n!)2n possible instances. The second author had proved that the expected number of stable marriages is of order nlnn, while its likely value is of order n1/2−o(1) at least. In this paper the second moment of that number is shown to be of order (nlnn)2. The combination of the two moment estimates implies that the fraction of problem instances with roughly cnlnn solutions is at least 0.84. Whether this fraction is asymptotic to 1 remains an open question.

Instance matching is a key task in knowledge graph fusion, and it is critical to improving the efficiency of instance matching, given the increasing scale of knowledge graphs. Blocking algorithms selecting candidate instance pairs for comparison is one of the effective methods to achieve the goal. In this paper, we propose a novel blocking algorithm named MultiObJ, which constructs indexes for instances based on the Ordered Joint of Multiple Objects’ features to limit the number of candidate instance pairs. Based on MultiObJ, we further propose a distributed framework named Follow-the-Regular-Leader Instance Matching (FTRLIM), which matches instances between large-scale knowledge graphs with approximately linear time complexity. FTRLIM has participated in OAEI 2019 and achieved the best matching quality with significantly efficiency. In this research, we construct three data collections based on a real-world large-scale knowledge graph. Experiment results on the constructed data collections and two real-world datasets indicate that MultiObJ and FTRLIM outperform other state-of-the-art methods.

The National Resident Matching program seeks a stable matching of medical students to teaching hospitals. With couples, stable matchings need not exist. Nevertheless, for any student preferences, we show that each instance of a matching problem has a “nearby” instance with a stable matching. The nearby instance is obtained by perturbing the capacities of the hospitals. In this perturbation, aggregate capacity is never reduced and can increase by at most four. The capacity of each hospital never changes by more than two. (JEL C78, D47, I11, J41, J44)

Constraint satisfaction problems are of special interest for the artificial intelligence and operations research community due to their many applications. Although heuristics involved in solving these problems have largely been studied in the past, little is known about the relation between instances and the respective performance of the heuristics used to solve them. This paper focuses on both the exploration of the instance space to identify relations between instances and good performing heuristics and how to use such relations to improve the search. Firstly, the document describes a methodology to explore the instance space of constraint satisfaction problems and evaluate the corresponding performance of six variable ordering heuristics for such instances in order to find regions on the instance space where some heuristics outperform the others. Analyzing such regions favors the understanding of how these heuristics work and contribute to their improvement. Secondly, we use the information gathered from the first stage to predict the most suitable heuristic to use according to the features of the instance currently being solved. This approach proved to be competitive when compared against the heuristics applied in isolation on both randomly generated and structured instances of constraint satisfaction problems.

The goal of linked data instance matching is to detect all instances that co-refer to the same objects in two linked data repositories, the source and the target. Since the amount of linked data is rapidly growing, it is important to automate this task. However, the difference between the schemata of source and target repositories remains a challenging barrier. This barrier reduces the portability, accuracy, and scalability of many proposed approaches. The authors present automatic schema-independent interlinking (ASL), which is a schema-independent system that performs instance matching on repositories with different schemata, without prior knowledge about the schemata. The key improvements of ASL compared to previous systems are the detection of useful attribute pairs for comparing instances, an attribute-driven token-based blocking scheme, and an effective modification of existing string similarities. To verify the performance of ASL, the authors conducted experiments on a large dataset containing 246 subsets with different schemata. The results show that ASL obtains high accuracy and significantly improves the quality of discovered coreferences against recently proposed complex systems.

CONSELHO NACIONAL DE DESENVOLVIMENTO CIENTÍFICO E TECNOLÓGICO
Um mediador é um componente de software que auxilia o acesso a fontes de dados. Com o advento da Web, a construção de mediadores impõe desafios importantes, tais como a capacidade de fornecer acesso integrado a fontes de dados independentes e dinâmicas e a habilidade de resolver a heterogeneidade semântica entre os esquemas destas fontes. Para lidar com esses desafios, o alinhamento de esquemas é uma questão fundamental. Nesta tese são propostas abordagens de alinhamento de esquemas de classificação (tesauros) e esquemas conceituais, utilizando instâncias como evidências para os mapeamentos. As abordagens propostas são classificadas em dois tipos: adaptativa e a priori, referindo-se, respectivamente, à descoberta dos mapeamentos de forma incremental ou à definição dos mapeamentos antes da implantação do mediador. Por fim, são apresentados experimentos para validação e teste das abordagens propostas.
A mediator is a software component that helps accessing data sources. With the advent of the Web, the design of mediators imposes important challenges, such as the ability of providing integrated access to independent and dynamic data sources and the ability of resolving the semantic heterogeneity between different data source schemas. To deal with these challenges, schema matching is a fundamental issue. In this thesis, matching approaches for classification schemas (thesauri) and conceptual schemas are proposed, using instances as evidences for the mappings. The proposed approaches are classified as adaptative and a priori, referring to, respectively, the discovery of the mappings in an incremental way or the definition of the mappings before the deployment of the mediator. Finally, experiments to validate and test the proposed approaches are presented.

Context: Determining attribute correspondence is the most important, time consuming and knowledge intensive part during databases integration. It is also used in other data manipulation applications such as data warehousing, data design, semantic web and e-commerce. Objectives: In this thesis the aim is to investigate how to find the types of correspondence between the attributes of heterogeneous datasets when schema design information of the data sets is unknown. Methods: A literature review was conducted to extract the knowledge related to the approaches that are used to find the correspondence between the attributes of heterogeneous datasets. Extracted knowledge from the literature review is used in developing an instance based approach for finding types of correspondence between the attributes of heterogeneous datasets when schema design information is unknown. To validate the proposed approach an experiment was conducted in the real environment using the data provided by the Telecom Industry (Ericsson) Karlskrona. Evaluation of the results was carried using the well known and mostly used measures from information retrieval field precision, recall and F-measure. Results: To find the types of correspondence between the attributes of heterogeneous datasets, good results depend on the ability of the algorithm to avoid the unmatched pairs of rows during the Row Similarity Phase. An evaluation of proposed approach is performed via experiments. We found 96.7% (average of three experiments) F-measure. Conclusions: The analysis showed that the proposed approach was feasible to be used and it provided users a mean to find the corresponding attributes and the types of correspondence between corresponding attributes, based on the information extracted from the similar pairs of rows from the heterogeneous data sets where their similarity based on the same common primary keys values.

En raison du grand nombre de graphes de connaissances et, surtout, de leurs interconnexions encore plus nombreuses à l'aide de la propriété owl:sameas, il est devenu de plus en plus évident que cette propriété est souvent mal utilisée. En effet, les entités liées par la propriété owl:sameas doivent être identiques dans tous les contextes possibles et imaginables. Dans les faits, ceci n'est pas toujours le cas et induit une détérioration de la qualité des données. L'identité doit être considérée comme étant dépendante d'un contexte. Nous avons donc proposé une étude à large échelle sur la présence de la sémantique dans les graphes de connaissances, puisque certaines caractéristiques sémantiques permettent justement de déduire des liens d'identités. Cette étude nous a amenés naturellement à construire une ontologie permettant de donner la teneur en sémantique d'un graphe de connaissances. Nous avons aussi proposé une approche de liage de données fondée à la fois sur la logique permise par les définitions sémantiques, et à la fois sur la prédominance de certaines propriétés pour caractériser la relation d'identité entre deux entités. Nous nous sommes aussi intéressés à la complétude et avons proposé une approche permettant de générer un schéma conceptuel afin de mesurer la complétude d'une entité. Pour finir, à l'aide des travaux précédents, nous avons proposé une approche fondée sur les plongements de phrases permettant de calculer les propriétés pouvant être propagées dans un contexte précis. Ceci permet l'expansion de requêtes SPARQL et, in fine, d'augmenter la complétude des résultats de la requête
Due to a large number of knowledge graphs and, more importantly, their even more numerous interconnections using the owl:sameas property, it has become increasingly evident that this property is often misused. Indeed, the entities linked by the owl:sameas property must be identical in all possible and imaginable contexts. This is not always the case and leads to a deterioration of data quality. Identity must be considered as context-dependent. We have, therefore, proposed a large-scale study on the presence of semantics in knowledge graphs since specific semantic characteristics allow us to deduce identity links. This study naturally led us to build an ontology allowing us to describe the semantic content of a knowledge graph. We also proposed a interlinking approach based both on the logic allowed by semantic definitions, and on the predominance of certain properties to characterize the identity relationship between two entities. We looked at completeness and proposed an approach to generate a conceptual schema to measure the completeness of an entity. Finally, using our previous work, we proposed an approach based on sentence embedding to compute the properties that can be propagated in a specific context. Hence, the propagation framework allows the expansion of SPARQL queries and, ultimately, to increase the completeness of query results

Vi ville med denna uppsats undersöka om instant messaging matchar de kommunikationsbehov som IT-anställda har. För att ta reda på detta använde vi oss av ett kvalitativt angreppssätt i form av intervjuer. Dessa intervjuer skapade även ytterligare kunskap vilken vi benämnt barriärer. En lista på IM-funktioner hämtades från Wikipedia och reviderades senare av oss. Kommunikationsbehoven identifierades under analysen av intervjuerna. Matchningen genererade en tabell som kan användas som karta över vilka IM-funktioner som tillgodoser några vanliga kommunikationsbehov som finns på datortäta arbetsplatser. Barriärerna är till för att visa vad som försvårar tillgodoseendet av behoven. Vidare erhöll vi även kunskap om att de av oss identifierade kommunikationsbehov som i dagsläget inte tillgodoses av IM inte heller behöver göra det. Vi kom också fram till att IM skulle kunna utvecklas till att vara en brygga mellan annan kommunicerande programvara och kollegor.

Dans cette thèse, nous adoptons une approche formelle pour définir et découvrir des mappings d'inclusion probabilistes entre deux taxonomies avec une sémantique claire, dans l'optique d'échange collaboratif de documents. Nous comparons deux façons de modéliser des mappings probabilistes tout en étant compatible avec les contraintes logiques déclarées dans chaque taxonomie selon une propriété de monotonie, puis nous montrons que ces modèles sont complémentaires pour distinguer les mappings pertinents. Nous fournissons un moyen d'estimer les probabilités d'un mapping par une technique bayésienne basée sur les statistiques des extensions des classes impliquées dans le mapping. Si les ensembles d'instances sont disjoints, on utilise des classifieurs pour les fusionner. Nous présentons ensuite un algorithme de type "générer et tester" qui utilise les deux modèles de mappings pour découvrir les plus probables entre deux taxonomies. Nous menons une analyse expérimentale fouillée de ProbaMap. Nous présentons un générateur de données synthétiques qui produit une entrée contrôlée pour une analyse quantitative et qualitative sur un large spectre de situations. Nous présentons aussi deux séries de résultats d'expériences sur des données réelles : l'alignement du jeu de donnée "Directory" d'OAEI, et une comparaison pour l'alignement de Web Directories sur lesquels ProbaMap obtient de meilleurs résultats que SBI (IJCAI 2003). Les perspectives pour ces travaux consistent à concevoir un système de réponse à des requêtes probabilistes en réutilisant des mappings probabilites, et la conversion des coefficients retournés par les méthodes de matching existantes en probabilités
In this thesis, we investigate a principled approach for defining and discovering probabilistic inclusion mappings between two taxonomies, with a clear semantic, in a purpose of collaborative exchange of documents. Firstly, we compare two ways of modeling probabilistic mappings which are compatible with the logical constraints declared in each taxonomy according to a monotony property, then we show that they are complementary for distinguishing relevant mappings. We provide a way to estimate the probabilities associated to a mapping by a Bayesian estimation technique based on classes extensions involved in the mapping, and using classifiers in order to merge the instances of both taxonomies when they are disjoint. Then we describe a generate and test algorithm called ProbaMap which minimizes the number of calls to the probability estimator for determining those mappings whose probability exceeds a chosen threshold. A thorough experimental analysis of ProbaMap is conducted. We introduce a generator that produce controlled data that allows to analyse the quality and the complexity of ProbaMap in a large and generic panel of situations. We present also two series of results for experiments conducted on real-world data: an alignment of the Directory dataset of the Ontology Alignment Evaluation Initiative (OAEI), and a comparative experiment on Web directories, on which ProbaMap outperforms the state-of-the-art contribution SBI (IJCAI'03). The perspectives of this work are the reuse of probabilistic mappings for a probabilistic query answering setting and a way to convert similarities coefficients of existing matching methods into probabilities

Tese de mestrado em Informática, apresentada à Universidade de Lisboa, através da Faculdade de Ciências, 2013
A Web Semântica pretende fornecer formatos comuns para caracterizar semanticamente os dados publicados na Web, melhorando a interoperabilidade e integração de dados. A iniciativa Linked Data visa ligar dados relacionados que não foram previamente ligados. As ontologias têm um papel fundamental nisso, pois, fornecem vocabulários controlados, para caracterizar semanticamente os dados de uma forma inequívoca. Conforme definido por Gruber, uma ontologia é uma especificação de uma conceituação, que se destina a modelar um domínio em particular. A especificação de uma ontologia é composto por dois tipos de declarações: TBox (classes) e ABox (exemplares). TBox são classes que são interpretadas como um conjunto de indivíduos no domínio; ABox são exemplares que são interpretados como indivíduos particulares de um domínio. Al´em disso, uma ontologia também é composta por: Relacionamentos ou relações entre classes e/ou exemplares; Tipos de dados são partes particulares do domínio que especificam valores; Valores de dados são valores simples. Apesar de uma ontologia se destinar a modelar um domínio em particular, existem muitas ontologias de diferentes fontes a modelar o mesmo domínio, isto é, existe um problema de sobreposição. O problema de sobreposição consiste em ontologias distintas que representam as mesmas entidades de uma forma diferente. É, portanto, necessário criar processos capazes de encontrar as sobreposições e fundi-las. Emparelhamento de ontologias é geralmente aplicado para alinhar duas TBox de duas ontologias diferentes, ou seja, para encontrar relações ou correspondências entre as classes ontológicas. Há um caso particular de emparelhamento de ontologias, o Emparelhamento de Exemplares. O objetivo do emparelhamento de exemplares é alinhar dois ABox de duas ontologias diferentes, ou seja, encontrar as correspondências entre exemplares de diferentes ontologias. O Emparelhamento de Exemplares adota o princípio de que, quanto maior for a semelhança entre duas descriçõees de exemplares de duas ontologias distintas, maior é a probabilidade de estes exemplares representarem a mesma entidade de um determinado domínio. Por exemplo, no domínio político, vamos considerar o actual Presidente da Comissão Europeia, Durão Barroso e assumir que na Ontologia 1 tem um exemplar com o descritor: “José Manuel Durão Barroso”, e Ontologia 2 tem um exemplar com o descritor: “José Durão Barroso”. Portanto, é necessário implementar técnicas de emparelhamento de exemplares, para descobrir se estes dois exemplares destas duas ontologias diferentes correspondem à mesma pessoa/entidade, isto é, se eles emparelham. Os objectivos desta dissertação eram: Desenvolvimento de algoritmos de emparelhamento de exemplares que visou o desenvolvimento de algoritmos para o emparelhamento de ontologias ao nível dos seus exemplares, de forma a resolver problemas de emparelhamento de exemplares. O desenvolvimento de algoritmos foi baseado em técnicas de emparelhamento de exemplares já propostas por outros; Alinhamento de exemplares do mundo real Que visou a aplicação dos algoritmos desenvolvidos, para gerar emparelhamentos de alta qualidade em exemplares do mundo real, e avaliar a sua qualidade em termos de Precisão, Sensibilidade, Medida-F, Exatidão e Exatidão Unilateral; Desenvolvimento de um emparelhador de exemplares Web Que visou o desenvolvimento de uma ferramenta capaz de realizar emparelhamento de exemplares através da Web, incorporando os algoritmos desenvolvidos por mim. Os resultados alcançados por esta dissertação foram a produção de alinhamentos de exemplares, entre as ontologias POWER-DBpediaPT, POWER-Verbetes e POWERPOWER. Estas três ontologias contêm exemplares que representam entidades políticas. E também entre as ontologias provenientes do OAEI 2012. O OAEI (Ontology Alignment Evaluation Initiative), é um concurso internacional, realizado todos os anos, que entre vários tipos de competições, tem uma dedicada à avaliação de ferramentas e de técnicas de emparelhamento de exemplares. Para avaliar a qualidade dos alinhamentos produzidos foram implementadas as seguintes métricas: Precisão; Sensibilidade; Medida-F; Exatidão; e Exatidão Unilateral. Esta dissertação também produziu um emparelhador de exemplares disponível através da Web, que implementa as métricas mencionadas para avaliar os alinhamentos produzidos por ele. POWER (Politics Ontology for Web Entity Retrieval) é uma ontologia que modela o domínio da política portuguesa, que foi desenvolvida e fornecida pela grupo REACTION. Os seus exemplares foram alinhados com os das ontologias DBpediaPT e Verbetes. A DBpediaPT é uma ontologia que contém exemplares que representam entidades da DBpedia versão 3.8. Cada entidade é referida na versão Portuguesa da Wikipédia. Esta ontologia foi construída a partir de uma lista, fornecida pelo grupo REACTION. Verbetes é uma ontologia, cujos os exemplares representam entidades que têm pelo menos cinco ocorrências nas notícias agregadas pelo serviço SAPO Verbetes. Para avaliar o alinhamento POWER-DBpediaPT foi usada a métrica Exatidão Unilateral. Usando o algoritmo de emparelhamento FirstLastNamePlusJaccard, alcançou-se 97.29% de Exatidão Unilateral para o POWER, e 87.25% de Exatidão Unilateral para o DBpediaPT. Usando o algoritmo de emparelhamento Stratified 10-fold Cross-Validation, alcançou-se 99.11% de Exatidão Unilateral para o POWER, e 95.97% de Exatidão Unilateral para o DBpediaPT. Estes foram os melhores resultados conseguidos para este alinhamento. No caso do alinhamento POWER-Verbetes não foram calculadas métricas mas, fez-se uma avaliação manual pela minha parte e pela parte do grupo REACTION, e foi positiva. Além disso, porque o POWER contém exemplares duplicados, ou seja, dois ou mais exemplares a representarem a mesma entidade, foi efectuado o alinhamento POWER-POWER de forma a encontrar os exemplares duplicados. No caso do POWER, estas situações não podiam acontecer. O alinhamento foi entregue ao grupo REACTION, para eles poderem melhorar a sua ontologia. Estes dois alinhamentos, POWER-Verbetes e POWER-POWER, foram realizados pelo algoritmo de emparelhamento MachineLearning. Foram também realizados alinhamentos de exemplares entre as ontologias fornecidas pelo OAEI 2012. Estas ontologias encontram-se divididas em dois grupos: o Sandbox que contém onze ontologias; e o IIMB que contém oitenta ontologias. Os alinhamentos produzidos foram realizados dentro de cada grupo. Neste caso, os algoritmos de emparelhamento utilizados foram FirstLastNamePlusJaccard e o Stratified 10-fold Cross- Validation. Na maioria dos alinhamentos produzidos a Medida-F foi maior no segundo algoritmo do que no primeiro. Todas as ontologias cujos exemplares foram alinhados, e os seus respectivos alinhamentos e métricas, estão disponíveis através da ligação: http://lasige.di.fc.ul.pt/webtools/instancematcher/dissertation_work.zip. O emparelhador de exemplares Web, foi outra realização desta dissertação, e está disponível através da ligação: http://lasige.di.fc.ul.pt/webtools/instancematcher/. Este disponibiliza aos utilizadores dois algoritmos de emparelhamento: o FirstLastNamePlusJaccard; e o MachineLearning. Além disso, também permite que o utilizador escolha que tipo de alinhamentos quer. Um-para-um (em Inglês: one-to-one) ou muitos-para-muitos (em Inglês: many-to-many). No primeiro caso, cada exemplar só pode estar presente uma vez no alinhamento, isto é, não pode haver mais do que um emparelhamento por exemplar; no segundo caso, cada exemplar pode estar presente várias vezes no alinhamento, ou seja, pode haver mais do que um emparelhamento por exemplar. Os alinhamentos POWER-DBpediaPT e POWER-Verbetes foram um-para-um. E os alinhamentos OAEI 2012 e POWER-POWER foram muitos-para-muitos. Há ainda a opção Limiar (em Inglês: Threshold) que permite ao utilizador indicar qual é o valor mínimo dos alinhamentos devolvidos pelo emparelhador de exemplares Web. Em cada alinhamento de exemplares é atribuído um valor [0,1] pelos algoritmos de emparelhamento, que determina o grau de confiabilidade/certeza do alinhamento estabelecido. No alinhamento também se podem encontrar exemplares que emparelham para nada, ou seja, para NULL. Estes, são os exemplares para os quais o algoritmo de emparelhamento escolhido, não encontrou nenhum exemplar correspondente. Para que o emparelhador de exemplares Web devolva métricas que atestem a qualidade do alinhamento produzido, o utilizador tem que introduzir o alinhamento de referência (em Inglês: Reference Alignment). Este é um documento, que se assume, que contenha todos os emparelhamentos correctos entre os exemplares de duas ontologias. As métricas são calculadas aquando da comparação do alinhamento produzido com o alinhamento de referência. Existem ainda as opções POWER 2010 e OAEI 2012, que permitem indicar ao emparelhador de exemplares Web, que os exemplares a emparelhar são do POWER e do OAEI 2012. É também necessário que o utilizador insira os identificadores dos descritores dos exemplares, para que o emparelhador obtenha a informação necessária para poder efectuar os alinhamentos. Cada identificador tem que começar pelo prefixo http.
An ontology is an object-based conceptualization of some particular domain. An ontology provides a shared controlled vocabulary to semantically characterize the data of the modelled domain. But it often happens that independently created ontologies model the same domain in different ways. This constitutes a problem because there may be entities being represented differently, therefore creating ambiguity and interoperability problems when linking related data characterized by two ontologies. So it is necessary to develop processes capable of matching the data. The matching can be made at the class level or at the instance level. The goal of the instance matching is to find the correspondences between instances from different ontologies, called instance alignments. The objective of this dissertation was the development of instance matching algorithms for generating instance alignments of real world instances. And the creation of an instance matcher Web tool, where the algorithms developed by me were incorporated. The outcome of this dissertation was the generation of instance alignments between POWER-DBpediaPT, POWER-Verbetes and POWER-POWER. All these three ontologies have instances representing political entities. Furthermore, it was generated instance alignments between ontologies from the OAEI 2012. OAEI (Ontology Alignment Evaluation Initiative), is an international contest, that has a track focus on evaluation of instance matching tools and techniques. To assess the quality of the instance alignments produced, it was implemented the metrics of Precision, Recall, F-measure, Accuracy and Unilateral Accuracy. Another outcome of this dissertation is the instance matcher tool, available through the Web. The tool implements two instance matchers. The FirstLastNamePlusJaccard which is based on element-level matching techniques, that uses the descriptors of the instances to correspond them. And the MachineLearning matcher that uses machine learning approaches to find those correspondences. This Web tool also assesses the instance alignments that it produces, because it implements the already mentioned metrics.

Ontologies are heavily used in life sciences so that there is increasing value to match different ontologies in order to determine related conceptual categories. We propose a simple yet powerful methodology for instance-based ontology matching which utilizes the associations between molecular-biological objects and ontologies. The approach can build on many existing ontology associations for instance objects like sequences and proteins and thus makes heavy use of available domain knowledge. Furthermore, the approach is flexible and extensible since each instance source with associations to the ontologies of interest can contribute to the ontology mapping. We study several approaches to determine the instance-based similarity of ontology categories. We perform an extensive experimental evaluation to use protein associations for different species to match between subontologies of the Gene Ontology and OMIM. We also provide a comparison with metadata-based ontology matching.

Ontologies describe the semantics of data and provide a uniform framework of understanding between different parties. The main common reference to an ontology definition describes them as knowledge bodies, which bring a formal representation of a shared conceptualization of a domain - the objects, concepts and other entities that are assumed to exist in a certain area of interest together with the relationships holding among them. However, in open and evolving systems with decentralized nature (as, for example, the Semantic Web), it is unlikely for different parties to adopt the same ontology. The problem of ontology matching evolves from the need to align ontologies, which cover the same or similar domains of knowledge. The task is to reducing ontology heterogeneity, which can occur in different forms, not in isolation from one another. Syntactically heterogeneous ontologies are expressed in different formal languages. Terminological heterogeneity stands for variations in names when referring to the same entities and concepts. Conceptual heterogeneity refers to differences in coverage, granularity or scope when modeling the same domain of interest. Finally, prgamatic heterogeneity is about mismatches in how entities are interpreted by people in a given context. The work presented in this thesis is a contribution to the problem of reducing the terminological and conceptual heterogeneity of hierarchical ontologies (defined as ontologies, which contain a hierarchical body), populated with text documents. We make use of both intensional (structural) and extensional (instance-based) aspects of the input ontologies and combine them in order to establish correspondences between their elements. In addition, the proposed procedures yield assertions on the granularity and the extensional richness of one ontology compared to another, which is helpful at assisting a process of ontology merging. Although we put an emphasis on the application of instance-based techniques, we show that combining them with intensional approaches leads to more efficient (both conceptually and computationally) similarity judgments. The thesis is oriented towards both researchers and practitioners in the domain of ontology matching and knowledge sharing. The proposed solutions can be applied successfully to the problem of matching web-directories and facilitating the exchange of knowledge on the web-scale.

This chapter articulates a knowledge first theory of doxastic justification—a belief is justified just in case it is relevantly similar to a possible instance of knowledge. In the terminology of the chapter, justification is “potential knowledge”. Relevant similarity is a matter of a matching of basic evidence and cognitive processing. This needn't be assumed to be a matter of the intrinsic; on the externalist approach to basic evidence given in Chapter 3, for one's belief to be justified is for there to be a possible knower who shares the same basic evidence—including factive perceptual states—and cognitive processing. The Appendix to Chapter 4 considers how the view extends to justified beliefs in necessarily false contents.

Tests of the importance of precautionary saving follow several research strategies. One aims to find a variable (or set of variables) that can approximate the variance of the growth rate of consumption. A second strategy seeks to estimate a reduced form for the level of consumption and wealth with proxies for income risk. A third approach simulates the path of consumption and wealth in models with precautionary saving, matching simulations with the observed distribution of wealth and consumption. Other studies provide indirect evidence for or against the precautionary saving hypothesis. Finally, some papers test the null hypothesis of the precautionary saving model (or more generally, self-insurance), in which risks can only be insured via private savings, against specific alternatives in which researchers make the source of market incompleteness explicit (positing, for instance, that it is due to private information).

This article attempts to locate MT (machine translation) in the contemporary context, while identifying its recent trends and themes. The 1990s were marked by the growing clout of empirical approaches, using increasingly available amounts of raw data in the form of parallel corpora. An appropriate instance is statistical MT, depending on bilingual corpus, but wherein, the translation depends on statistical modeling of the word order of target language and of source-target word equivalences. Example-based MT involves matching inputs against actual databases and identifying close matches. Successful MT, especially when attempting translation into languages marking gender of pronouns, those that have zero-anaphora contents, interpretation of anaphora assumes paramountcy. Regarding spoken-language MTs, coupling speech-to-text front-end and text-to-speech, the other way, is inadequate other than rigidly formal languages. More versatile operations such as dealing with dialogues, would involve greater complexities.

This article describes Bayesian modelling for matching and alignment of biomolecules. One particular task where statistical modelling and inference can be useful in scientific understanding of protein structure is that of matching and alignment of two or more proteins. In this regard, statistical shape analysis potentially has something to offer in solving biomolecule matching and alignment problems. The article discusses the use of Bayesian methods for shape analysis to assist with understanding the three-dimensional structure of protein molecules, with a focus on the problem of matching instances of the same structure in the CoMFA (Comparative Molecular Field Analysis) database of steroid molecules. It introduces a Bayesian hierarchical model for pairwise matching and for alignment of multiple configurations before concluding with an overview of some advantages of the Bayesian approach to problems in protein bioinformatics, along with modelling and computation issues, alternative approaches, and directions for future research.

This chapter proposes and discusses how a “purpose-driven” approach can help business reinvent and sustain itself and serve society. It outlines and explains a “tesseract model” which has purpose at its core and four vertices corresponding to economic, human, social, and environmental values. To be effective, a company’s purpose has to be authentic, ambitious, and achievable. A company’s purpose can furnish meaning and status, and act as a matching/sorting device for attracting compatible employees, investors, and customers. Based on the authors’ first-hand experience, this chapter describes two concrete instances of firms that adopted this purpose-driven approach. In an organization that adopts a societally oriented purpose, employees become emotionally engaged, energized, and differentiating in ways that can sustain performance and progress.

Partial Label Learning (PLL) aims to learn from training data where each instance is associated with a set of candidate labels, among which only one is correct. In this paper, we formulate the task of PLL problem as an ``instance-label'' matching selection problem, and propose a DeepGNN-based graph matching PLL approach to solve it. Specifically, we first construct all instances and labels as graph nodes into two different graphs respectively, and then integrate them into a unified matching graph by connecting each instance to its candidate labels. Afterwards, the graph attention mechanism is adopted to aggregate and update all nodes state on the instance graph to form structural representations for each instance. Finally, each candidate label is embedded into its corresponding instance and derives a matching affinity score for each instance-label correspondence with a progressive cross-entropy loss. Extensive experiments on various data sets have demonstrated the superiority of our proposed method.

Multi-Label Image Classification (MLIC) aims to predict a set of labels that present in an image. The key to deal with such problem is to mine the associations between image contents and labels, and further obtain the correct assignments between images and their labels. In this paper, we treat each image as a bag of instances, and reformulate the task of MLIC as a instance-label matching selection problem. To model such problem, we propose a novel deep learning framework named Graph Matching based Multi-Label Image Classification (GM-MLIC), where Graph Matching (GM) scheme is introduced owing to its excellent capability of excavating the instance and label relationship. Specifically, we first construct an instance spatial graph and a label semantic graph respectively, and then incorporate them into a constructed assignment graph by connecting each instance to all labels. Subsequently, the graph network block is adopted to aggregate and update all nodes and edges state on the assignment graph to form structured representations for each instance and label. Our network finally derives a prediction score for each instance-label correspondence and optimizes such correspondence with a weighted cross-entropy loss. Extensive experiments conducted on various datasets demonstrate the superiority of our proposed method.