Improving cross language information retrieval using corpus based query suggestion approach

Users seeking information may not find relevant information pertaining to their information need in a specific language. But information may be available in a language different from their own, but users may not know that language. Thus users may experience difficulty in accessing the information present in different languages. Since the retrieval process depends on the translation of the user query, there are many issues in getting the right translation of the user query. For a pair of languages chosen by a user, resources, like incomplete dictionary, inaccurate machine translation system may exist. These resources may be insufficient to map the query terms in one language to its equivalent terms in another language. Also for a given query, there might exist multiple correct translations. The underlying corpus evidence may suggest a clue to select a probable set of translations that could eventually perform a better information retrieval. In this paper, we present a cross language information retrieval approach to effectively retrieve information present in a language other than the language of the user query using the corpus driven query suggestion approach. The idea is to utilize the corpus based evidence of one language to improve the retrieval and re-ranking of news documents in the other language. We use FIRE corpora - Tamil and English news collections in our experiments and illustrate the effectiveness of the proposed cross language information retrieval approach.

Learning representations for Information Retrieval

La recherche d'informations s'intéresse, entre autres, à répondre à des questions comme: est-ce qu'un document est pertinent à une requête ? Est-ce que deux requêtes ou deux documents sont similaires ? Comment la similarité entre deux requêtes ou documents peut être utilisée pour améliorer l'estimation de la pertinence ? Pour donner réponse à ces questions, il est nécessaire d'associer chaque document et requête à des représentations interprétables par ordinateur. Une fois ces représentations estimées, la similarité peut correspondre, par exemple, à une distance ou une divergence qui opère dans l'espace de représentation. On admet généralement que la qualité d'une représentation a un impact direct sur l'erreur d'estimation par rapport à la vraie pertinence, jugée par un humain. Estimer de bonnes représentations des documents et des requêtes a longtemps été un problème central de la recherche d'informations. Le but de cette thèse est de proposer des nouvelles méthodes pour estimer les représentations des documents et des requêtes, la relation de pertinence entre eux et ainsi modestement avancer l'état de l'art du domaine. Nous présentons quatre articles publiés dans des conférences internationales et un article publié dans un forum d'évaluation. Les deux premiers articles concernent des méthodes qui créent l'espace de représentation selon une connaissance à priori sur les caractéristiques qui sont importantes pour la tâche à accomplir. Ceux-ci nous amènent à présenter un nouveau modèle de recherche d'informations qui diffère des modèles existants sur le plan théorique et de l'efficacité expérimentale. Les deux derniers articles marquent un changement fondamental dans l'approche de construction des représentations. Ils bénéficient notamment de l'intérêt de recherche dont les techniques d'apprentissage profond par réseaux de neurones, ou deep learning, ont fait récemment l'objet. Ces modèles d'apprentissage élicitent automatiquement les caractéristiques importantes pour la tâche demandée à partir d'une quantité importante de données. Nous nous intéressons à la modélisation des relations sémantiques entre documents et requêtes ainsi qu'entre deux ou plusieurs requêtes. Ces derniers articles marquent les premières applications de l'apprentissage de représentations par réseaux de neurones à la recherche d'informations. Les modèles proposés ont aussi produit une performance améliorée sur des collections de test standard. Nos travaux nous mènent à la conclusion générale suivante: la performance en recherche d'informations pourrait drastiquement être améliorée en se basant sur les approches d'apprentissage de représentations.

Learning representations for Information Retrieval

La recherche d'informations s'intéresse, entre autres, à répondre à des questions comme: est-ce qu'un document est pertinent à une requête ? Est-ce que deux requêtes ou deux documents sont similaires ? Comment la similarité entre deux requêtes ou documents peut être utilisée pour améliorer l'estimation de la pertinence ? Pour donner réponse à ces questions, il est nécessaire d'associer chaque document et requête à des représentations interprétables par ordinateur. Une fois ces représentations estimées, la similarité peut correspondre, par exemple, à une distance ou une divergence qui opère dans l'espace de représentation. On admet généralement que la qualité d'une représentation a un impact direct sur l'erreur d'estimation par rapport à la vraie pertinence, jugée par un humain. Estimer de bonnes représentations des documents et des requêtes a longtemps été un problème central de la recherche d'informations. Le but de cette thèse est de proposer des nouvelles méthodes pour estimer les représentations des documents et des requêtes, la relation de pertinence entre eux et ainsi modestement avancer l'état de l'art du domaine. Nous présentons quatre articles publiés dans des conférences internationales et un article publié dans un forum d'évaluation. Les deux premiers articles concernent des méthodes qui créent l'espace de représentation selon une connaissance à priori sur les caractéristiques qui sont importantes pour la tâche à accomplir. Ceux-ci nous amènent à présenter un nouveau modèle de recherche d'informations qui diffère des modèles existants sur le plan théorique et de l'efficacité expérimentale. Les deux derniers articles marquent un changement fondamental dans l'approche de construction des représentations. Ils bénéficient notamment de l'intérêt de recherche dont les techniques d'apprentissage profond par réseaux de neurones, ou deep learning, ont fait récemment l'objet. Ces modèles d'apprentissage élicitent automatiquement les caractéristiques importantes pour la tâche demandée à partir d'une quantité importante de données. Nous nous intéressons à la modélisation des relations sémantiques entre documents et requêtes ainsi qu'entre deux ou plusieurs requêtes. Ces derniers articles marquent les premières applications de l'apprentissage de représentations par réseaux de neurones à la recherche d'informations. Les modèles proposés ont aussi produit une performance améliorée sur des collections de test standard. Nos travaux nous mènent à la conclusion générale suivante: la performance en recherche d'informations pourrait drastiquement être améliorée en se basant sur les approches d'apprentissage de représentations.

The impact of spatial data redundancy on SOLAP query performance

Geographic Data Warehouses (GDW) are one of the main technologies used in decision-making processes and spatial analysis, and the literature proposes several conceptual and logical data models for GDW. However, little effort has been focused on studying how spatial data redundancy affects SOLAP (Spatial On-Line Analytical Processing) query performance over GDW. In this paper, we investigate this issue. Firstly, we compare redundant and non-redundant GDW schemas and conclude that redundancy is related to high performance losses. We also analyze the issue of indexing, aiming at improving SOLAP query performance on a redundant GDW. Comparisons of the SB-index approach, the star-join aided by R-tree and the star-join aided by GiST indicate that the SB-index significantly improves the elapsed time in query processing from 25% up to 99% with regard to SOLAP queries defined over the spatial predicates of intersection, enclosure and containment and applied to roll-up and drill-down operations. We also investigate the impact of the increase in data volume on the performance. The increase did not impair the performance of the SB-index, which highly improved the elapsed time in query processing. Performance tests also show that the SB-index is far more compact than the star-join, requiring only a small fraction of at most 0.20% of the volume. Moreover, we propose a specific enhancement of the SB-index to deal with spatial data redundancy. This enhancement improved performance from 80 to 91% for redundant GDW schemas.

Familial partial epilepsy with variable foci: A new partial epilepsy syndrome with suggestion of linkage to chromosome 2

Familial partial epilepsy with variable foci (FPEVF) joins the recently recognized group of inherited partial epilepsies. We describe an Australian family with 10 individuals with partial seizures over four generations. Detailed electroclinical studies were performed on all affected and 17 clinically unaffected family members. The striking finding was that the clinical features of the seizures and interictal electroencephalographic foci differed among family members and included frontal, temporal, occipital, and centroparietal seizures. Mean age of seizure onset was 13 years (range, 0.75-43 years). Two individuals without seizures had epileptiform abnormalities on electroencephalographic studies. Penetrance of seizures was 62%. A genome-wide search failed to demonstrate definitive linkage, but a suggestion of linkage was found on chromosome 2q with a LOD score of 2.74 at recombination fraction of zero with the marker D2S133. FPEVF differs from the other inherited partial epilepsies where partial seizures in different family members are clinically similar. The inherited nature of this new syndrome may be overlooked because of relatively low penetrance and because of the variability in age at onset and electroclinical features between affected family members.

The effects of information request ambiguity and construct incongruence on query development

This paper examines the effects of information request ambiguity and construct incongruence on end user's ability to develop SQL queries with an interactive relational database query language. In this experiment, ambiguity in information requests adversely affected accuracy and efficiency. Incongruities among the information request, the query syntax, and the data representation adversely affected accuracy, efficiency, and confidence. The results for ambiguity suggest that organizations might elicit better query development if end users were sensitized to the nature of ambiguities that could arise in their business contexts. End users could translate natural language queries into pseudo-SQL that could be examined for precision before the queries were developed. The results for incongruence suggest that better query development might ensue if semantic distances could be reduced by giving users data representations and database views that maximize construct congruence for the kinds of queries in typical domains. (C) 2001 Elsevier Science B.V. All rights reserved.

Cognitive style factors affecting database query performance

As end-user computing becomes more pervasive, an organization's success increasingly depends on the ability of end-users, usually in managerial positions, to extract appropriate data from both internal and external sources. Many of these data sources include or are derived from the organization's accounting information systems. Managerial end-users with different personal characteristics and approaches are likely to compose queries of differing levels of accuracy when searching the data contained within these accounting information systems. This research investigates how cognitive style elements of personality influence managerial end-user performance in database querying tasks. A laboratory experiment was conducted in which participants generated queries to retrieve information from an accounting information system to satisfy typical information requirements. The experiment investigated the influence of personality on the accuracy of queries of varying degrees of complexity. Relying on the Myers–Briggs personality instrument, results show that perceiving individuals (as opposed to judging individuals) who rely on intuition (as opposed to sensing) composed queries more accurately. As expected, query complexity and academic performance also explain the success of data extraction tasks.

Adherence to health-related behaviors: effectiveness of implementation intentions and posthypnotic suggestion in college students

Resumo: Com base no conceito de implementação de intenções (Gollwitzer, 1993, 1999) e na teoria do contexto de resposta de Kirsch & Lynn (1997), o presente trabalho testou a eficácia de uma intervenção combinada de implementação de intenções com hipnose e sugestão pós-hipnótica na promoção da adesão a uma tarefa simples (avaliação do humor) e uma tarefa difícil (actividade física). Os participantes são estudantes universitários de uma universidade na Nova Jérsia, (N=124, Estudo 1, EUA) e em Lisboa (N=323, Estudo 2, Portugal). Em ambos os estudos os participantes foram seleccionados a partir de uma amostra mais vasta baseado num escrutínio da sua sugestibilidade hipnótica avaliada por meio da Escala de Grupo de Sugestibilidade Hipnótica de Waterloo-Stanford (WSGC): Forma C. O Estudo 1 usou um desenho factorial do tipo 2x2x3 (tipo de intenção formada x hipnose x nível de sugestionabilidade) e o Estudo 2 usou um desenho factorial do tipo 2 x 2x 2 x 4 (tipo de tarefa x tipo de intenção formada x hipnose x nível de sugestionabilidade). No Estudo 1 foi pedido aos participantes que corressem todos os dias e durante três semanas durante 5 minutos, que medissem a sua pulsação antes e depois da actividade física e que mandassem um e-mail ao experimentador, fornecendo assim uma medida comportamental e uma medida de auto-relato. Aos participantes no grupo de intenções de meta foi apenas pedido que corressem todos os dias. Aos participantes no grupo de implementação de intenções foi pedido que especificasses com exactidão quando e onde iriam correr e enviar o e-mail. Para além disso, cerca de metade dos participantes foram hipnotizados e receberam uma sugestão pós-hipnótica em que lhes foi sugerido que o pensamento de correr todos os dias lhes viria à mente sem esforço no momento apropriado. A outra metade dos participantes não recebeu qualquer sugestão hipnótica. No Estudo 2 foi seguido o mesmo procedimento, mas a cerca de metade dos participantes foi atribuída uma tarefa fácil (enviar um Adherence to health-related behaviors ix SMS com a avaliação diária do seu estado de humor naquele momento) e à outra metade da amostra foi atribuída a tarefa de exercício físico atrás descrita (tarefa difícil). Os resultados do estudo 1 mostraram uma interacção significativa entre o nível de sugestionabilidade dos participantes e a sugestão pós-hipnótica (p<.01) indicando que a administração da sugestão pós-hipnótica aumentou a adesão nos participantes muito sugestionáveis, mas baixou a adesão nos participantes pouco sugestionáveis. Não se encontraram diferenças entre os grupos que formaram intenções de meta e os que formaram implementação de intenções. No Estudo 2 os resultados indicaram que os participantes aderiram significativamente mais à tarefa fácil do que à tarefa difícil (p<.001). Os resultados não revelaram diferenças significativas entre as condições implementações de intenções, hipnose e as duas estratégias combinadas, indicando que a implementação de intenções não foi eficaz no aumento da adesão às duas tarefas propostas e não beneficiou da combinação com as sugestões pós-hipnóticas. A utilização da hipnose com sugestão pós-hipnótica significativamente reduziu a adesão a ambas as tarefas. Dado que não existiam instrumentos em Português destinados a avaliar a sugestionabilidade hipnótica, traduziu-se e adaptou-se para Português Escala de Grupo de sugestibilidade hipnótica de Waterloo-Stanford (WSGC): Forma C. A amostra Portuguesa (N=625) apresentou resultados semelhantes aos encontrados nas amostras de referência em termos do formato da distribuição dos padrões da pontuação e do índice de dificuldade dos itens. Contudo, a proporção de estudantes portugueses encontrada que pontuaram na zona superior de sugestionabilidade foi significativamente inferior à proporção de participantes na mesma zona encontrada nas amostras de referência. No sentido de lançar alguma luz sobre as razões para este resultado, inquiriu-se alguns dos participantes acerca das suas atitudes face à hipnose utilizando uma versão portuguesa da Escala de Valência de Atitudes e Crenças face à Hipnose e comparou-se com a opinião de Adherence to health-related behaviors xAbstract: On the basis of Gollwitzer’s (1993, 1999) implementation intentions’ concept, and Kirsch & Lynn’s (1997) response set theory, this dissertation tested the effectiveness of a combined intervention of implementation intentions with hypnosis with posthypnotic suggestions in enhancing adherence to a simple (mood report) and a difficult (physical activity) health-related task. Participants were enrolled in a university in New Jersey (N=124, Study 1, USA) and in two universities in Lisbon (N=323, Study 2, Portugal). In both studies participants were selected from a broader sample based on their suggestibility scores using the Waterloo-Stanford Group C (WSGC) scale of hypnotic susceptibility and then randomly assigned to the experimental groups. Study 1 used a 2x2x3 factorial design (instruction x hypnosis x level of suggestibility) and Study 2 used a 2 x 2x 2 x 4 factorial design (task x instructions x hypnosis x level of suggestibility). In Study 1 participants were asked to run in place for 5 minutes each day for a three-week period, to take their pulse rate before and after the activity, and to send a daily email report to the experimenter, thus providing both a self-report and a behavioral measure of adherence. Participants in the goal intention condition were simply asked to run in place and send the e-mail once a day. Those in the implementation intention condition were further asked to specify the exact place and time they would perform the physical activity and send the e-mail. In addition, half of the participants were given a post-hypnotic suggestion indicating that the thought of running in place would come to mind without effort at the appropriate moment. The other half did not receive a posthypnotic suggestion. Study 2 followed the same procedure, but additionally half of the participants were instructed to send a mood report by SMS (easy task) and half were assigned to the physical activity task described above (difficult task). Adherence to health-related behaviors vii Study 1 result’s showed a significant interaction between participant’s suggestibility level and posthypnotic suggestion (p<.01) indicating that posthypnotic suggestion enhanced adherence among highly suggestible participants, but lowered it among low suggestible individuals. No differences between the goal intention and the implementation intentions groups were found. In Study 2, participants adhered significantly more (p<.001) to the easy task than to the difficult task. Results did not revealed significant differences between the implementation intentions, hypnosis and the two conditions combined, indicating that implementation intentions was not enhanced by hypnosis with posthypnotic suggestion, neither was effective as single intervention in enhancing adherence to any of the tasks. Hypnosis with posthypnotic suggestion alone significantly reduced adherence to both tasks in comparison with participants that did not receive hypnosis. Since there were no instruments in Portuguese language to asses hypnotic suggestibility, the Waterloo-Stanford Group C (WSGC) scale of hypnotic susceptibility was translated and adapted to Portuguese and was used in the screening of a sample of college students from Lisbon (N=625). Results showed that the Portuguese sample has distribution shapes and difficulty patterns of hypnotic suggestibility scores similar to the reference samples, with the exception of the proportion of Portuguese students scoring in the high range of hypnotic suggestibility, that was found lower than the in reference samples. In order to shed some light on the reasons for this finding participant’s attitudes toward hypnosis were inquired using a Portuguese translation and adaptation of the Escala de Valencia de Actitudes y Creencias Hacia la Hipnosis, Versión Cliente, and compared with participants with no prior hypnosis experience (N=444). Significant differences were found between the two groups with participants without hypnosis experience scoring higher in factors indicating misconceptions and negative attitudes about hypnosis.

Questing one Brazilian query: reporting 16 cases of Q fever from Minas Gerais, Brazil

Q fever has been considered non-existing in Brazil where reports of clinical cases still cannot be found. This case-series of 16 patients is a result of a systematic search for such illness by means of clinical and serologic criteria. Serologic testing was performed by the indirect microimmunofluorescence technique using phase I/II C. burnetii antigens. Influenza-like syndrome was the most frequent clinical form (eight cases - 50%), followed by pneumonia, FUO (fever of unknown origin), mono-like syndrome (two cases - 12.5% each), lymphadenitis (one case - 6.3%) and spondylodiscitis associated with osteomyelitis (one case - 6.3%). The ages varied from four to 67 years old with a median of 43.5. All but one patient had positive serologic tests for phase II IgG whether or not associated with IgM positivity compatible with acute infection. One patient had both phase I and phase II IgG antibodies compatible with chronic Q fever. Seroconvertion was detected in 10 patients. Despite the known limitations of serologic diagnosis, the cases here reported should encourage Brazilian doctors to include Q fever as an indigenous cause of febrile illness.

A qualitive reasoning approach for improving query results for sketch based queries by topological analysis of spatial aggregation

Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies.

Expressiveness improvements of OutSystems DSL query primitives

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

Efficient query processing in co-processor-accelerated databases

...Diese Dissertation zeigt, wie wir Datenbankmanagementsysteme bauen können, die heterogene Prozessoren effizient und zuverlässig zur Beschleunigung der Anfrageverarbeitung nutzen können. Daher untersuchen wir typische Entwurfsentscheidungen von coprozessorbeschleunigten Datenbankmanagementsystemen und leiten darauf aufbauend eine generische Architektur für solche Systeme ab. Unsere Untersuchungen zeigen, dass eines der wichtigsten Probleme für solche Datenbankmanagementsysteme die Entscheidung ist, welche Operatoren einer Anfrage auf welchem Prozessor ausgeführt werden sollen...