Diversified query expansion

La diversification des résultats de recherche (DRR) vise à sélectionner divers documents à partir des résultats de recherche afin de couvrir autant d’intentions que possible. Dans les approches existantes, on suppose que les résultats initiaux sont suffisamment diversifiés et couvrent bien les aspects de la requête. Or, on observe souvent que les résultats initiaux n’arrivent pas à couvrir certains aspects. Dans cette thèse, nous proposons une nouvelle approche de DRR qui consiste à diversifier l’expansion de requête (DER) afin d’avoir une meilleure couverture des aspects. Les termes d’expansion sont sélectionnés à partir d’une ou de plusieurs ressource(s) suivant le principe de pertinence marginale maximale. Dans notre première contribution, nous proposons une méthode pour DER au niveau des termes où la similarité entre les termes est mesurée superficiellement à l’aide des ressources. Quand plusieurs ressources sont utilisées pour DER, elles ont été uniformément combinées dans la littérature, ce qui permet d’ignorer la contribution individuelle de chaque ressource par rapport à la requête. Dans la seconde contribution de cette thèse, nous proposons une nouvelle méthode de pondération de ressources selon la requête. Notre méthode utilise un ensemble de caractéristiques qui sont intégrées à un modèle de régression linéaire, et génère à partir de chaque ressource un nombre de termes d’expansion proportionnellement au poids de cette ressource. Les méthodes proposées pour DER se concentrent sur l’élimination de la redondance entre les termes d’expansion sans se soucier si les termes sélectionnés couvrent effectivement les différents aspects de la requête. Pour pallier à cet inconvénient, nous introduisons dans la troisième contribution de cette thèse une nouvelle méthode pour DER au niveau des aspects. Notre méthode est entraînée de façon supervisée selon le principe que les termes reliés doivent correspondre au même aspect. Cette méthode permet de sélectionner des termes d’expansion à un niveau sémantique latent afin de couvrir autant que possible différents aspects de la requête. De plus, cette méthode autorise l’intégration de plusieurs ressources afin de suggérer des termes d’expansion, et supporte l’intégration de plusieurs contraintes telles que la contrainte de dispersion. Nous évaluons nos méthodes à l’aide des données de ClueWeb09B et de trois collections de requêtes de TRECWeb track et montrons l’utilité de nos approches par rapport aux méthodes existantes.

Formaci??n en lenguas extranjeras para la sociedad de la informaci??n. 'Foreign Language Teaching in the Information Age'

Resumen tomado del autor

Verb retrieval in fluent aphasia: a clinical study

Background: Problems with lexical retrieval are common across all types of aphasia but certain word classes are thought to be more vulnerable in some aphasia types. Traditionally, verb retrieval problems have been considered characteristic of non-fluent aphasias but there is growing evidence that verb retrieval problems are also found in fluent aphasia. As verbs are retrieved from the mental lexicon with syntactic as well as phonological and semantic information, it is speculated that an improvement in verb retrieval should enhance communicative abilities in this population as in others. We report on an investigation into the effectiveness of verb treatment for three individuals with fluent aphasia. Methods & Procedures: Multiple pre-treatment baselines were established over 3 months in order to monitor language change before treatment. The three participants then received twice-weekly verb treatment over approximately 4 months. All pre-treatment assessments were administered immediately after treatment and 3 months post-treatment. Outcome & Results: Scores fluctuated in the pre-treatment period. Following treatment, there was a significant improvement in verb retrieval for two of the three participants on the treated items. The increase in scores for the third participant was statistically nonsignificant but post-treatment scores moved from below the normal range to within the normal range. All participants were significantly quicker in the verb retrieval task following treatment. There was an increase in well-formed sentences in the sentence construction test and in some samples of connected speech. Conclusions: Repeated systematic treatment can produce a significant improvement in verb retrieval of practised items and generalise to unpractised items for some participants. An increase in well-formed sentences is seen for some speakers. The theoretical and clinical implications of the results are discussed.

Language in calculation: A core mechanism?

Although there is evidence that exact calculation recruits left hemisphere perisylvian language systems, recent work has shown that exact calculation can be retained despite severe damage to these networks. In this study, we sought to identify a “core” network for calculation and hence to determine the extent to which left hemisphere language areas are part of this network. We examined performance on addition and subtraction problems in two modalities: one using conventional two-digit problems that can be easily encoded into language; the other using novel shape representations. With regard to numerical problems, our results revealed increased left fronto-temporal activity in addition, and increased parietal activity in subtraction, potentially reflecting retrieval of linguistically encoded information during addition. The shape problems elicited activations of occipital, parietal and dorsal temporal regions, reflecting visual reasoning processes. A core activation common to both calculation types involved the superior parietal lobule bilaterally, right temporal sub-gyral area, and left lateralized activations in inferior parietal (BA 40), frontal (BA 6/8/32) and occipital (BA 18) regions. The large bilateral parietal activation could be attributed to visuo-spatial processing in calculation. The inferior parietal region, and particularly the left angular gyrus, was part of the core calculation network. However, given its activation in both shape and number tasks, its role is unlikely to reflect linguistic processing per se. A possibility is that it serves to integrate right hemisphere visuo-spatial and left hemisphere linguistic and executive processing in calculation.

Processing reflexives in a second language: the timing of structural and discourse-level information

We report the results from two eye-movement monitoring experiments examining the processing of reflexive pronouns by proficient German-speaking learners of second language (L2) English. Our results show that the nonnative speakers initially tried to link English argument reflexives to a discourse-prominent but structurally inaccessible antecedent, thereby violating binding condition A. Our native speaker controls, in contrast, showed evidence of applying condition A immediately during processing. Together, our findings show that L2 learners’ initial focusing on a structurally inaccessible antecedent cannot be due to first language influence and is also independent of whether the inaccessible antecedent c-commands the reflexive. This suggests that unlike native speakers, nonnative speakers of English initially attempt to interpret reflexives through discourse-based coreference assignment rather than syntactic binding.

From XML to relational view updates: applying old solutions to solve a new problem

XML has become an important medium for data exchange, and is frequently used as an interface to - i.e. a view of - a relational database. Although lots of work have been done on querying relational databases through XML views, the problem of updating relational databases through XML views has not received much attention. In this work, we give the rst steps towards solving this problem. Using query trees to capture the notions of selection, projection, nesting, grouping, and heterogeneous sets found throughout most XML query languages, we show how XML views expressed using query trees can be mapped to a set of corresponding relational views. Thus, we transform the problem of updating relational databases through XML views into a classical problem of updating relational databases through relational views. We then show how updates on the XML view are mapped to updates on the corresponding relational views. Existing work on updating relational views can then be leveraged to determine whether or not the relational views are updatable with respect to the relational updates, and if so, to translate the updates to the underlying relational database. Since query trees are a formal characterization of view de nition queries, they are not well suited for end-users. We then investigate how a subset of XQuery can be used as a top level language, and show how query trees can be used as an intermediate representation of view de nitions expressed in this subset.

Homonymy in natural language processes: A representation using Pustejovsky's Qualia Structure and ontological information

This paper presents a proposal for the semantic treatment of ambiguous homographic forms in Brazilian Portuguese, and to offer linguistic strategies for its computational implementation in Systems of Natural Language Processing (SNLP). Pustejovsky's Generative Lexicon was used as a theoretical model. From this model, the Qualia Structure - QS (and the Formal, Telic, Agentive and Constitutive roles) was selected as one of the linguistic and semantic expedients for the achievement of disambiguation of homonym forms. So that analyzed and treated data could be manipulated, we elaborated a Lexical Knowledge Base (LKB) where lexical items are correlated and interconnected by different kinds of semantic relations in the QS and ontological information.

XML na ciência da informação: uma análise do MARC21

Pós-graduação em Ciência da Informação - FFC

Wikis semânticos: da Web para a Web Semântica

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

A scalable re-ranking method for content-based image retrieval

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)