Caio Saraiva Coneglian

Objects Assessment Approach Using Natural Language Processing and Data Quality to Support Emergency Situation Assessment

Situation Awareness (SAW) is a cognitive process that is defined by the perception of relevant elements present in a monitored environment (e.g., people, objects, vehicles, places), the understanding of their meaning (i.e., what they are doing) and the projection of their statuses in the near future. In the domain of emergency management, the data employed to the process of acquisition and maintenance of SAW are provided by several sources, using different formats and different classifications, such as: images from security cameras, reports made to the emergency response center, posts in social networks and several physical sensors, such as: positional, altitude and movement. Data from this sources, if well processed and understood by a specialist, may contribute to the decision-making process, supporting the establishment of emergency response tactics and a better allocation of operational resources. The acquisition of SAW demands the characterization of the ongoing situation. Typically, knowing exactly what is going on demands exhaustive routines of intelligent data assessment. In the emergency management domain, it means to better explore and analyze what the humans say about the events. This paper presents a general architecture that integrates objects and situational assessment for the emergency management domain and a specific process for the objects assessment using natural language processing (NLP) and semantic practices, to better identify relevant elements that may be useful for the situation assessment routines, such as information fusion. Known approaches are limited due to the absence of data quality analysis as part of the process, undesirable when decision makers need to rely on emergency information. Preliminary results of a case study of an intelligent object assessment of a robbery situation reported in Brazilian Portuguese demonstrate the advantages and practical particularities of our solution.

Towards semantic fusion using information quality and the assessment of objects and situations to improve emergency situation awareness

Information Fusion is the integration of synergic information to support cognition and high-level processing. Emergency management systems may take advantage of such integration and better support human operators in the development of Situational Awareness (SAW) for decision-making. The critical and dynamic nature of real emergency scenarios impose challenges to reveal, integrate and derive useful information for decision processes. The problem increases when humans are the main source of data, leading to information quality issues, such as imprecision, inconsistency and uncertainty. Current syntactical-only fusion approaches are limited regarding the assessment of situational meaning and human language nuances. Semantic models help to describe and to apply relationships among entities that may be useful for a net centric fusion and Situation Assessment (SA) routines. The objective of this paper is to present advances towards a new semantic fusion approach supported by information quality inferences and semantic web concepts to improve the SA about emergency situations and hence supporting SAW. For such, a new architecture is presented to integrate objects and situation assessment approaches by syntactical and semantic means. A previous fusion approach based on a syntactic integration with quality indexes is used to illustrate the improvements on information fusion results with the semantic models.

Multi-criteria Fusion of Heterogeneous Information for Improving Situation Awareness on Emergency Management Systems

Information Fusion is the synergic integration of data from different sources for the support to decision-making. The emergency management systems predominance of such application has driven the development to new and better sensors, new methods, for data processing and architectures that promote access, composition, refinement and information handling, with the active participation of specialists as data providers and specialists of the systems. In this scenario of data fusion, uncertainty of diverse natures can be aggregated to both data and information at different levels of the process, creating distorted information to the specialist. As a result the situation awareness and cognitive process can be affected leading to poor quality support to decision-making as a generalization of information quality, uncertainty need to be reduced to improve awareness about the situation of interest. The objective of our work is the mitigation of uncertainty propagated by other quality attributes such as information completeness, so specialists can be able to convey an improved understanding. For such, a new fusion framework fed by multi-criteria parameterization, including information quality measures and its semantics, is depicted as an engine to build more accurate information from diverse sensed possibilities. A case study with a situation assessment application is in course to validate the effectiveness of the generated solution. Preliminary and promising results are discussed as a more valuable tool to support decision-making.

Materialização da Web Semântica: um modelo de construção dinâmica de consultas baseados em mapeamento de ontologias

Aproximar a linguagem computacional da humana e um desafio no processo de Recuperacao da Informacao. Nessa perspectiva, a Web Semântica tem permitido novos meios de tratamento dos dados contidos na Web, afim de permitir melhor compreensao do significado destas informacoes. Contudo, alguns dos conceitos que envolvem a Web Semântica, como as ontologias, apresentam alta complexidade, dificultando o seu uso. A dificuldade de compreensao, a estrutura que estas possuem, e a diversidade como elas sao construidas, tornam o manuseio de ontologias para a descoberta do significado dos termos, algo nao efetivo, impossibilitando oferecer um panorama concreto do sentido dos termos. Assim, essa pesquisa tem como objetivo propor um modelo que verifique o contexto e o significado de um termo dentro de uma ontologia. Para tal, foi realizado pesquisa bibliografica, construcao do modelo e a implementacao de um prototipo, como prova de conceito, para a verificacao dos resultados. O modelo e a implementacao apresentam como resultado o tratamento generico de ontologias, obtendo o significado e o contexto que um termo possui. Baseados nos resultados identifica-se que a utilizacao de ontologias permite que sistemas computacionais possam apresentar uma perspectiva ampliada do contexto dos dados, atendendo mais eficazmente as necessidades informacionais dos usuarios.

Audiovisuais e Linked data: um estudo das bases DBpedia e LMDB

A As proponent of the Semantic Web and Linked data principles, the Linking Open Data initiative offers an enormous proportion of audiovisual data, which can assist in the search and retrieval of more accurate information. In this sense, the objective of this work is to explore the possible relations between audiovisual databases and Linking Open Data, to present the potential of this initiative for users who seek detailed sources of information about the audiovisual resources. A qualitative research was used, with an exploratory and applied nature, based on the scientific literature of the Linked Data, Semantic Web and audiovisual, and subsequently the DBpedia and LMDB with the use of the SPARQL protocol. The study considers that datasets available in the Linking Open Data can assist the link between information on audiovisual resources, as well as, it may be a source for the construction of more dynamic catalogues, reducing rework during the process of description of informational resources.

Agente de extração e identificação de estruturas semânticas em ambientes informacionais digitais

No cenario atual da Internet com a producao massiva de informacoes, as areas de Extracao e Tratamento da Informacao estao sendo desafiadas pelo volume, variedade e velocidade de dados semiestruturados e nao estruturados de natureza complexa que devem ser encontrados e julgados quanto ao seu valor e veracidade, que tambem oferece as organizacoes excelentes oportunidades de terem um aprofundamento no conhecimento mais preciso de seus negocios. Neste contexto, ao processo de inovacao tornou-se o foco de muitas empresas e esta sendo explorado como meio de melhorar a competitividade e posicionamento de empresas em novos mercados. Este trabalho tem como objetivo estabelecer um mecanismo computacional de extracao e identificacao de estruturas semânticas de fontes informacionais especificas, onde o espaco informacional sera o site de noticias da FAPESP (Fundacao de Amparo a Pesquisa do Estado de Sao Paulo). Alem disso, obtem-se resultados que vao desde a criacao de extrator semântico especifico e geral ate o modelo de RDF para persistencia de metadados e dados em ambientes informacionais digitais.

Conceitos e tecnologias da Web semântica no contexto da colaboração acadêmico-científica: um estudo da plataforma Vivo

Univ Sao Paulo, Fac Filosofia Ciencias & Letras, Dept Educ Informacao & Comunicacao, Av Bandeirantes 3900, BR-14040901 Ribeirao Preto, SP, Brazil

Big Data: fatores potencialmente discriminatórios em análise de dados

As mudancas tecnologicas vividas a partir da virada do seculo causaram uma revolucao na sociedade, chamada de Big Data, em que as analises de dados para determinar padroes e comportamentos puderam utilizar grandes quantidades de dados. Verifica-se que algumas analises, no contexto do Big Data, estao sendo conduzidas a gerar resultados discriminatorios. O estudo tem como objetivo identificar fatores que, potencialmente, possam gerar discriminacao durante o processo de analise de dados. Para tal, a metodologia utilizada foi de natureza qualitativa, exploratoria e bibliografica, enumerando em um quadro os casos de discriminacao. Como resultado, identificam-se fatores possivelmente discriminatorios, alem de ser feita uma explanacao desses fatores. Por meio da pesquisa, verifica-se uma necessidade de existir reflexoes profundas dos resultados que sao obtidos a partir de analises de dados, ficando clara a necessidade da Ciencia da Informacao retratar tais questoes, a fim de apontar os caminhos a serem tomados.

Ontological Semantic Agent in the Context of Big Data: A Tool Applied to Information Retrieval in Scientific Research

The large increase in the creation and dissemination of data on the Internet can offer information of high value-added to organizations. This information can be provided by heterogeneous databases that may not be considered relevant by most systems, e.g., social media data, blogs and more. If organizations would use such sources, they could build a new management vision known as Competitive Intelligence. In the context of architectures of Information Retrieval, this research aims on implementing a semantic extraction agent for the Web environment, allowing information finding, storage, processing and retrieval, such as those from the Big Data context produced by several informational sources on the Internet, serving as a basis for the implementation of information environments for decision support. Using this method, it will be possible to verify that the agent and ontology proposal addresses this part and can play the role of a semantic level of the architecture.

Semantic Agent in the Context of Big Data - Usage in Ontological Information Retrieval in Scientific Research

The evolution of information technology caused an expansion in the amount of data available on the internet. Moreover, such developments demanded that new tools were created to allow processing at high velocity, trying various informational sources. In this context, in flocking to the three V (Velocity, Variety and Volume), emerged the phenomenon called Big Data. From the emergence of this phenomenon, the need to generate new architectures that allow that users, enjoy the high volume of data spread throughout the Web. One way to improve the processes carried out, insert the question of semantic information processing, in which the use of domain ontologies can expand as computational agents interpret the meaning of the data. Thus, this paper aims to present a proposal for architecture that places the elements of Big Data and semantic, seeking to insert a model that is adapted to the current computing needs. As proof of concept performed the implementation of the architecture, exploring the question of scientific research, where a user is assisted to find relevant information in academic databases. Through the implementation, it was found that the use ontologies in a Big Data architecture, significantly improves the recovery of information performed by computational agents.