Search

Microblogging sites, like Twitter, have emerged as ubiquitous sources of information. Two important tasks related to the automatic extraction and analysis of information in Microblogs are Entity Mention Detection (EMD) and Entity Detection (ED). The state-of-the-art EMD systems aim to model the non-literary nature of microblog text by training upon offline static datasets. They extract a combination of surface-level features -- orthographic, lexical, and semantic -- from individual messages for noisy text modeling and entity extraction. But given the constantly evolving nature of microblog streams, detecting all entity mentions from such varying yet limited context of short messages remains a difficult problem. To this end, we propose a framework named EMD Globalizer, better suited for the execution of EMD learners on microblog streams. It deviates from the processing of isolated microblog messages by existing EMD systems, where learned knowledge from the immediate context of a message is used to suggest entities. After an initial extraction of entity candidates by an EMD system, the proposed framework leverages occurrence mining to find additional candidate mentions that are missed during this first detection. Aggregating the local contextual representations of these mentions, a global embedding is drawn from the collective context of an entity candidate within a stream. The global embeddings are then utilized to separate entities within the candidates from false positives. All mentions of said entities from the stream are produced in the framework's final outputs. Our experiments show that EMD Globalizer can enhance the effectiveness of all existing EMD systems that we tested (on average by 25.61%) with a small additional computational overhead.

We present YumiInt a deep Web integration system for local search engines for Geo-referenced objects. YumiInt consists of two systems: YumiDev and YumiMeta. YumiDev is an off-line integration system that builds the key components (e.g., query translation and entity resolution) of YumiMeta. YumiMeta is the Web application to which users post queries. It translates queries to multiple sources and gets back aggregated lists of results. We present the two systems in this paper.

The contents of Web databases are accessed through queries formulated on complex user interfaces. In many domains of interest (e.g. Auto) users are interested in obtaining information from alternative sources. Thus, they have to access many individual Web databases via query interfaces. We aim to construct automatically a well-designed query interface that integrates a set of interfaces in the same domain. This will permit users to access information uniformly from multiple sources. Earlier research in this area includes matching attributes across multiple query interfaces in the same domain and grouping related attributes. In this paper, we investigate the naming of the attributes in the integrated query interface. We provide a set of properties which are required in order to have consistent labels for the attributes within an integrated interface so that users have no difficulty in understanding it. Based on these properties, we design algorithms to systematically label the attributes. Experimental results on seven domains validate our theoretical study. In the process of naming attributes, a set of logical inference rules among the textual labels is discovered. These inferences are also likely to be applicable to other integration problems sensitive to naming: HTML forms, HTML tables or concept hierarchies in the semantic Web.

Orientadores: Beatriz Daltrini, Weiyi Meng Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica Resumo: Esse trabalho apresenta um algoritmo que transforma esquemas relacionais em esquemas orientados a objetos num sistema de múltiplos bancos de dados heterogêneos. Nesse contexto, transformação de esquemas é um passo necessário antes de se fazer processamento de comandos de consulta. A implementação do algoritmo tem as seguintes características especiais: transforma esquemas entre dois sistemas de bancos de dados comerciais, especificamente de esquemas relacionais Sybase para esquemas orientados a objetos 02; inicia sessões interativas limitadas com o usuário, quando o transformador de esquemas necessita de informações adicionais, embora a maioria das transformações seja automática; os esquemas transformados são típicos esquemas orientados a objetos que podem conter hierarquia de classes, hierarquia de composição, hierarquia de tuplas e conjuntos de atributos; e o número de classes no esquema orientado a objetos transformado é relativamente pequeno, comparado com o número de relações no equivalente esquema relacional Abstract: This work presents an algorithm that transforms relational schemas to object -oriented schemas in a heterogeneous multidatabase system. In this context, schema transformation is a necessary step before query processing. The implementation of the algorithm has the following special features: it transforms schemas between two commercial database systems, specifically from Sybase relational schemas to 02 objectoriented schemas; it initiates limited interactive sessions between the schema transformer and its user when additional input to the schema transformer is necessary, although most of the transformation is automated; the transformed schemas are typical object-oriented schemas that may contain class hierarchies, composition hierarchies, set attributes and subtuple constructs; and the number of classes in the transformed object-oriented schema is relatively small. compared with the number of relations in the equivalent relational schema Mestrado Mestre em Engenharia Elétrica