Search

The goal of an information retrieval system is to provide the information that is relevant to the user's query. In some cases the information relevant to the user request may not exist in the user's native language. Situations may also arise where the user is able to read documents in languages different from the native one, but might have difficulty in formulating queries in them. The main intention behind Multilingual Information Retrieval is to find the relevant information available irrespective of the language used in the query.

Cross language information retrieval (CLIR) is a retrieval process in which the user fires queries in one language to retrieve information from another (different) language. The amount of information generated in different languages and disseminated via the World Wide Web and social media platforms is growing exponentially. It is important to ensure that online information resources and services are efficiently and equitably accessible to all users, regardless of their linguistic and cultural backgrounds. The need for exploring and building more specialized information systems that enable speakers of an Odia language to discover valuable information beyond linguistic and cultural barriers. This study is aimed to build an experimental CLIR system between one of the under-resourced languages (i.e. Odia), and one of the most commonly used online languages, (i.e. English). However, we have not yet considered any prototype approach to develop English-Odia CLIR system. In this research we studied on the various IR variants and techniques used in CLIR system. Further, based on available literature, a number of challenges and issues in CLIR have been identified and discussed. Firstly, we have focused on building basic linguistic resources and IR tools required for designing an English-Odia CLIR using Apache Lucene engine. Secondly, some of the major linguistic resources that we developed during the course of this study include a Unicode compatible Odia Keyboard, normalised Odia corpus, a generic computational model of Odia morphology, a bilingual dictionary and an Odia stemmer from scratch. While our EnglishOdia dictionary has been adopted and used for query translation.

Translingual Text Mining (TTM) is an innovative technology of natural language processing for building multilingual parallel corpora, processing machine translation, contextual knowledge acquisition, information extraction, query profiling, language modeling, contextual word sensing, creating feature test sets and for variety of other purposes. The Keynote Lecture will discuss opportunities and challenges of this computational technology. In particular, the focus will be made on identification of language pair phenomena and their applications to building holistic language model which is a novel tool for processing machine translation, supporting professional translations, evaluation of translingual systems efficiency and also for improving the process of teaching foreign languages regardless of the level of students' proficiency. Some components incorporated to machine translation systems rely specifically on language-pair phenomena like for example language recognizer, or word deliminer. Nowadays, declarative programming is becoming widely used just for describing language pair phenomena by graphical formalism. Translation irreversibility represents a unique language and system independent asymmetrical translation innovative technology, which will be presented during the Keynote lecture.

All the languages belonging to the same language family have a certain number of the common characteristics called language pair phenomena, which can be found quite useful for processing them for multilingual purposes like translation across the cognate languages, building dictionaries, thesauri, transcript collections, or for multilingual text retrieval of digital documents. In addition, it is estimated that more than 30% of English vocabulary has been inherited from Latin, which has dominated medical terminology in particular. We use this fact by exploring word sense disambiguation (WSD) in multilingual environment. Specifically in the medical domain, language pair phenomena can be limited to synonymy of the cognate technical terms. Our approach is investigated based on Boolean and Free Text Search modes on the comparison basis. For measuring the efficiency of our methodology we use the classical Salton model of tf-idf term weighting schemes, however extended by Karen Sparck Jones. Our results are very promising since they indicate that similarity between the synonymous words being English medical terms and their target language equivalents enables significant limitation of the target word senses even those outside the language family like e.g. for the English and Polish language pair phenomena. Such a limitation of the number of target word senses results in better disambiguation and is more context-driven. Also, consequently it translates onto the higher precision in multilingual medical information retrieval.

In recent years, technological advancements in Natural Language Processing (NLP) such as machine translation have made it possible for users to access information in multiple languages, even those in which they may not be proficient. The current study investigated the information searching behavior of bi/multilingual academic users, and examined their practices and perceptions regarding searching for information on the Internet and on electronic databases. Bi/multilingual students were recruited from a Canadian university and a community college both located in London, Ontario. A total of 250 ( N = 250) students completed a web survey through a link that was embedded in an invitation e-mail. Results showed that though advancements in NLP technology have alleviated some of the linguistic related challenges that some bi/multilingual academic users face while searching for information online, language barriers do still exist for some especially at the query formulation stage . The study found that an increase in Multilingual Information Access (MLIA) tools on electronic databases coupled with appropriate information literacy instruction could be helpful in further alleviating language barriers.

En este trabajo se evalúan diferentes modelos de indexación, así como la aplicación de técnicas de Procesamiento del Lenguaje Natural al texto asociado a las imágenes en un problema de Recuperación de Imágenes Multilingüe. Los resultados muestran que traducir el texto asociado y usar Entidades Nombradas, junto con sus categorías, permite mejorar el proceso de recuperación, mientras que la mejora obtenida con el uso de sintagmas nominales no compensa el coste computacional que conlleva. In this paper we evaluate several indexing models and the application of different Natural Language Processing techniques to textual descriptions associated with images in an Image Retrieval task. The results show that the use of a translation-based model and to take into account the Named Entities with their categories, improves the retrieval process. However, the improvement obtained with the use of noun phrases is not worth the computational cost involved. Proyectos de investigación MA2VICMR (S2009/TIC-1542), HOLOPEDIA (TIN2010-21128-C02) y PROYECTO UNED (2012V/PUNED/0004).

Περιέχει το πλήρες κείμενο Purpose: This study aims to identify, collect and critical review the research literature on Multilingual Digital Libraries in English language from 1997 to 2012. Design/methodology/approach: The present literature review has followed the rules of systematic review. In particular, the identified relevant papers were categorized based on their expressed aim on two core themes, that of system-centered and user-centered studies. The assigned papers were further analyzed and six sub-themes emerged for the system-centered studies and four for the user-centered studies. Additional categorization was also provided according to type of publication. Findings: The literature concerning Multilingual Digital Libraries is vast and mainly focuses on two aspects the “System” and the “Users”. The majority of papers tried to meet the challenges raised for enabling multilingual information retrieval in Digital Libraries. Unfortunately, these efforts undertaken by a small number of researchers or research groups apparently working in isolation and therefore resulting in the development of numerous different tools and techniques. Relatively few studies have focused on the user and aimed to explore users’ behavior and expectations when interacting with Multilingual Digital Libraries. As a result, further research is needed to reach to some tangible and usable findings. Originality/value: This literature review captures the diversity of the research conducted regarding multilingual information access and retrieval in Digital Libraries. It organizes the vast literature in comprehensive themes and sub-themes enabling easy access to specific information. Limitations: This study reviews only papers in English due to language restrictions from 1997 to 2012.

Peters, Braschler and Clough present a comprehensive description of the technologies involved in designing and developing systems for Multilingual Information Retrieval (MLIR). They provide readers with broad coverage of the various issues involved in creating systems to make accessible digitally stored materials regardless of the language(s) they are written in. Details on Cross-Language Information Retrieval (CLIR) are also covered that help readers to understand how to develop retrieval systems that cross language boundaries. Their work is divided into six chapters and accompanies the reader step-by-step through the various stages involved in building, using and evaluating MLIR systems. The book concludes with some examples of recent applications that utilise MLIR technologies. Some of the techniques described have recently started to appear in commercial search systems, while others have the potential to be part of future incarnations. The book is intended for graduate students, scholars, and practitioners with a basic understanding of classical text retrieval methods. It offers guidelines and information on all aspects that need to be taken into consideration when building MLIR systems, while avoiding too many 'hands-on details' that could rapidly become obsolete. Thus it bridges the gap between the material covered by most of the classical IR textbooks and the novel requirements related to the acquisition and dissemination of information in whatever language it is stored.

The current lack of recent and long-term query logs makes the verifiability and repeatability of log analysis experiments very limited. A first attempt in this direction has been made within the Cross-Language Evaluation Forum in 2009 in a track named LogCLEF which aims to stimulate research on user behaviour in multilingual environments and promote standard evaluation collections of log data. We report on similarities and differences of the most recent activities for LogCLEF.

Actualmente existen algunas iniciativas dentro de la tarea de Búsqueda de Respuestas orientadas al desarrollo de sistemas capacitados para encontrar respuestas sobre colecciones multilingües. Esto es, el usuario puede formular su consulta en diversos idiomas, y puede obtener la respuesta a partir de colecciones documentales escritas igualmente en diversos idiomas. Por un lado, para una pregunta dada, buscar en más documentos conlleva más documentos relevantes y por lo tanto mayor probabilidad de encontrar una respuesta satisfactoria. Por otro lado, la multilingualidad requiere necesariamente de un proceso de traducción automática, bien de la consulta, bien de los pasajes relevantes o bien de ambos, con la consiguiente introducción de ruido ¿Hasta qué punto ese ruido distorsiona los resultados? ¿Aportan todas las colecciones monolingües al rendimiento global del sistema multilingüe o por el contrario depende de cómo de bien representado esté ese idioma dentro del sistema multilingüe? En este trabajo hemos encontrado que un sistema de búsquedas de respuestas multilingüe puede superar a uno monolingüe, y que la aportación de cada idioma es desigual, pero en todos los casos es significativa. Nowadays there are some works in the Question Answering (QA) task that have a multilingual orientation. The user can make questions in several languages, he will obtain the answers in the same language and these answers could be in collections in several languages. On the one hand, a multilingual system requires an automatic translation module to translate the questions, the relevant documents or passages or both, and this translation introduces some noise in the system. How important is the noise? Are all the collections significant to the performance of the QA system? In this work we conclude that a multilingual QA system improves a monolingual one, and the performance of each language is not similar, but it is significant. Este trabajo ha sido financiado por el Ministerio de Ciencia y Tecnología a través del proyecto TIMOM (TIN2006-15265-C06-03).

Actualmente existen algunas iniciativas dentro de la tarea de Búsqueda de Respuestas orientadas al desarrollo de sistemas capacitados para encontrar respuestas sobre colecciones multilingües. Esto es, el usuario puede formular su consulta en diversos idiomas, y puede obtener la respuesta a partir de colecciones documentales escritas igualmente en diversos idiomas. Por un lado, para una pregunta dada, buscar en más documentos conlleva más documentos relevantes y por lo tanto mayor probabilidad de encontrar una respuesta satisfactoria. Por otro lado, la multilingualidad requiere necesariamente de un proceso de traducción automática, bien de la consulta, bien de los pasajes relevantes o bien de ambos, con la consiguiente introducción de ruido ¿Hasta qué punto ese ruido distorsiona los resultados? ¿Aportan todas las colecciones monolingües al rendimiento global del sistema multilingüe o por el contrario depende de cómo de bien representado esté ese idioma dentro del sistema multilingüe? En este trabajo hemos encontrado que un sistema de búsquedas de respuestas multilingüe puede superar a uno monolingüe, y que la aportación de cada idioma es desigual, pero en todos los casos es significativa. Nowadays there are some works in the Question Answering (QA) task that have a multilingual orientation. The user can make questions in several languages, he will obtain the answers in the same language and these answers could be in collections in several languages. On the one hand, a multilingual system requires an automatic translation module to translate the questions, the relevant documents or passages or both, and this translation introduces some noise in the system. How important is the noise? Are all the collections significant to the performance of the QA system? In this work we conclude that a multilingual QA system improves a monolingual one, and the performance of each language is not similar, but it is significant. Este trabajo ha sido financiado por el Ministerio de Ciencia y Tecnología a través del proyecto TIMOM (TIN2006-15265-C06-03).

Los Diccionarios Analógico Conceptuales que contiene la Plataforma Intercultural Multilenguaje Zirano (Zirano Cúpula) se basan en la Semántica Conceptual Multilenguaje, donde el átomo inicial es el concepto, con ello desaparecen algunas limitaciones de los sistemas que parten del inglés, como WordNet o Universal Networking Language. La doble representación en el signo lingüístico permite que cada signo actúe como una variable conceptual: representamos con él todos los significantes que puede llegar a tener (a veces tantos como lenguas hay en el mundo). Esto permite búsquedas inclusivas en corpora desambiguados. Así, por ejemplo, en nuestro Diccionario Analógico Conceptual Castellano, ‘noche’ no sólo está asociado con ‘nocturno’ o ‘anochecer’ sino también con ‘aquelarre’, ‘cena’, ‘estrigiforme’, ‘asmático’, ‘vivaquear’, etc., conceptos/palabras que contienen la variable conceptual ‘noche’. The Analogical Conceptual Dictionaries present in the ‘Plataforma Intercultural Multilenguaje Zirano (Zirano Cúpula)’ are based on Multilanguage Conceptual Semantics, where the initial atom is the concept. Therefore several limitations of the systems based on English (like WordNet or Universal Networking Language) disappear. The double representation in the linguistic sign makes possible that a sign acts as a conceptual variable. That sign represents all the signifiers that it can have (sometimes, as many as languages in the world). This makes possible inclusive searches in disambiguated corpora. So, in our Spanish Analogical Conceptual Dictionary ‘night’ is associated not only with ‘nocturne’ or ‘night fall’ but also with ‘aquelarre’, ‘dinner’, ‘estrigiforme’, ‘quiropter’, ‘asthmatic’, ‘to bivouac’, etc., concepts/words which contain the conceptual variable ‘night’.

The rise of globalization, including technological innovations and the dismantling of trade barriers, has spurred the steady acceleration of global trade and, in barely a decade, has transformed the competitive environment of enterprises. The area of activity has been expanded by the emergence of new markets with very attractive potential. So are the BRIC (Brazil, Russia, India and China). Among the four of them, all impressive by their size, population and economic potential they represent, China is the least accessible and the more closed to our understanding because of a linguistic system radically different from the Indo-European languages on the one hand and of the fact of a culture and a thought system at odds with those of Western countries. Yet for a company of international size, which wants to extend its influence or simply to maintain its market position, including its own market, it is now essential to be present on the Chinese market. How does a western company operate on a market that appears at first as inherently complex and enigmatic because of its otherness? During six years of observation in China, we have found out the pitfalls in access to information about the Chinese market. As on many markets, our companies are subject to some unimaginable destabilization. The inability to “read” China and understand the issues that take place in spite of sustained efforts, the tactical errors that arise from a misjudgement of the market or a biased understanding of the game players led us to consider a methodology that could provide French companies an approach to China as a market. The methodologies of Business Intelligence (BI) came out to be the most suitable for several reasons: the goal of BI is to find out the right action to realise, the specificity of the context in which the organization is evolving is taken into consideration and the analysis is done just in time. If a cultural approach is made of human interactions and subtleties, a market approach is now possible by the automatic processing of information and its modelling. In any process of economic intelligence accompanying the establishment of a foreign operation, a large part of the strategic information comes from analysis of the game players operating in the same sector of activity. Such an automation of knowledge creation is, in addition to the human approach on the field, a real high value added to help the understanding of the interactions between the players. It provides a set of knowledge, which taking into account more large entities, are more comprehensive. They are more elusive anywhere else. Because has highly developed technologies linked to the knowledge economy, it is now possible to explore the scientific and technological sources of information science in China. We are also convinced that Chinese sources of information will take a more and more crucial importance in any global watch. It is therefore an urgent need for organizations to get solutions that not only allow the access to this information but also are able to handle the masses of information from these sources. The aim of this thesis is mainly to adapt the tools and the methods invented by French university research to the analysis of Chinese information in order to create useful knowledge. Matheo software will provide some bibliometrical treatments that will give a global vision of the Chinese strategy. Tetralogy software, a tool dedicated to data-mining, will be tailored to the linguistic environment and to the structure of the Chinese scientific databases. In addition, we participate in the development of a method for the information retrieval (MEVA) which integrates the data of recent discoveries in cognitive science. We develop this method to the research of the relevant and appropriate information among the Chinese datas. As this thesis is conduced under a contract university /enterprise with Limagrain, an application of our approach will be implemented in the field of agricultural biotechnology and in particular around issues of research on techniques for hybridization of wheat. The analysis of this sector, which is an area of fundamental research, experimental and applied is a very current topic as it gives rise to the acquisition of patents and to the marketing of commercial products. Is China really, as we suppose to, a new territory Global Scientific Research of the 21st century? Can the methods of BI be adapted to the Chinese market? After providing some answers to these questions in the first two parts of our study, the third part will describe the global context of agricultural biotechnologies and its issues in terms of economic and financial power but also geopolitical. Then we will focus on the problematic of research on hybridization wheat. Then we will see in the fourth and last part how to implement a search for information on the Chinese market and the major interest in terms of added value of information analysis in China, Le décloisonnement des économies et l’accélération mondiale des échanges commerciaux ont, en une décennie à peine, transformés l’environnement concurrentiel des entreprises. La zone d’activités s’est élargie en ouvrant des nouveaux marchés à potentiels très attrayants. Ainsi en est-il des BRIC (Brésil, Russie, Inde et Chine). De ces quatre pays, impressionnants par la superficie, la population et le potentiel économique qu’ils représentent, la Chine est le moins accessible et le plus hermétique à notre compréhension de par un système linguistique distinct des langues indo-européennes d’une part et du fait d’une culture et d’un système de pensée aux antipodes de ceux de l’occident d’autre part. Pourtant, pour une entreprise de taille internationale, qui souhaite étendre son influence ou simplement conserver sa position sur son propre marché, il est aujourd’hui absolument indispensable d’être présent sur le marché chinois. Comment une entreprise occidentale aborde-t-elle un marché qui de par son altérité, apparaît tout d’abord comme complexe et foncièrement énigmatique ? Six années d’observation en Chine, nous ont permis de constater les écueils dans l’accès à l’information concernant le marché chinois. Comme sur de nombreux marchés extérieurs, nos entreprises sont soumises à des déstabilisations parfois inimaginables. L’incapacité à « lire » la Chine et à comprendre les enjeux qui s’y déroulent malgré des effets soutenus, les erreurs tactiques qui découlent d’une mauvaise appréciation du marché ou d’une compréhension biaisée des jeux d’acteurs nous ont incités à réfléchir à une méthodologie de décryptage plus fine de l’environnement d’affaire qui puisse offrir aux entreprises françaises une approche de la Chine en tant que marché. Les méthodes de l’Intelligence Economique (IE) se sont alors imposées comme étant les plus propices pour plusieurs raisons : le but de l’IE est de trouver l’action juste à mener, la spécificité du contexte dans lequel évolue l’organisation est prise en compte et l’analyse se fait en temps réel. Si une approche culturelle est faite d’interactions humaines et de subtilités, une approche « marché » est dorénavant possible par le traitement automatique de l’information et de la modélisation qui s’en suit. En effet, dans toute démarche d’Intelligence Economique accompagnant l’implantation d’une activité à l’étranger, une grande part de l’information à portée stratégique vient de l’analyse du jeu des acteurs opérants dans le même secteur d’activité. Une telle automatisation de la création de connaissance constitue, en sus de l’approche humaine « sur le terrain », une réelle valeur ajoutée pour la compréhension des interactions entre les acteurs car elle apporte un ensemble de connaissances qui, prenant en compte des entités plus larges, revêtent un caractère global, insaisissable par ailleurs. La Chine ayant fortement développé les technologies liées à l’économie de la connaissance, il est dorénavant possible d’explorer les sources d’information scientifiques et techniques chinoises. Nous sommes en outre convaincus que l’information chinoise prendra au fil du temps une importance de plus en plus cruciale. Il devient donc urgent pour les organisations de se doter de dispositifs permettant non seulement d’accéder à cette information mais également d’être en mesure de traiter les masses d’informations issues de ces sources. Notre travail consiste principalement à adapter les outils et méthodes issues de la recherche française à l’analyse de l’information chinoise en vue de la création de connaissances élaborées. L’outil MATHEO, apportera par des traitements bibliométriques une vision mondiale de la stratégie chinoise. TETRALOGIE, outil dédié au data-mining, sera adapté à l’environnement linguistique et structurel des bases de données scientifiques chinoises. En outre, nous participons au développement d’un outil d’information retreival (MEVA) qui intègre les données récentes des sciences cognitives et oeuvrons à son application dans la recherche de l’information chinoise, pertinente et adéquate. Cette thèse étant réalisée dans le cadre d’un contrat CIFRE avec le Groupe Limagrain, une application contextualisée de notre démarche sera mise en œuvre dans le domaine des biotechnologies agricoles et plus particulièrement autour des enjeux actuels de la recherche sur les techniques d’hybridation du blé. L’analyse de ce secteur de pointe, qui est à la fois une domaine de recherche fondamentale, expérimentale et appliquée donne actuellement lieu à des prises de brevets et à la mise sur le marché de produits commerciaux et représente donc une thématique très actuelle. La Chine est-elle réellement, comme nous le supposons, un nouveau territoire mondial de la recherche scientifique du 21e siècle ? Les méthodes de l’IE peuvent-elles s’adapter au marché chinois ? Après avoir fourni les éléments de réponses à ces questions dans es deux premières parties de notre étude, nous poserons en troisième partie, le contexte des biotechnologies agricoles et les enjeux mondiaux en terme de puissance économico-financière mais également géopolitique de la recherche sur l’hybridation du blé. Puis nous verrons en dernière partie comment mettre en œuvre une recherche d’information sur le marché chinois ainsi que l’intérêt majeur en terme de valeur ajoutée que représente l’analyse de l’information chinoise

With the availability of external resources to documents, the Information Retrieval Systems evolve. These resources provide not only information on the terms and concepts for a more precise indexing, but also the semantic relations between these terms or concepts. Our thesis work lies in the use of external resources in information retrieval. We study firstly the indexing conceptual in comparison with term-based indexing. The problem arrise when the documents and the query don't share the same concepts, but the concepts of the documents are semantically related with the concepts of the query. We propose to take into account these semantic relationships between concepts by a information retrieval model which is based on a Bayesian network of concepts and their semantic relationships. In addition, we propose the use of knowledge of the problem from external ressource to improve the performance of retrieval. The validation of the proposed model is achieved by experiments in the medical domaine information retrieval, with the use of meta thesaurus UMLS as external resource. The application for a system of information retrieval multi modality (text and images) was also performed.; Dans les systèmes de recherche d'information, une indexation à base de termes et une correspondance à base d'intersection introduisent le problème de la disparité à cause des variations linguistiques. Avec l'objectif de résoudre ce problème, notre travail de thèse se positionne dans l'utilisation des ressources externes dans la recherche d'information. Ces ressources offrent non seulement les concepts pour une indexation plus précise et indépendante de langue, mais aussi une base de relations sémantiques entre ces concepts. Nous étudions en premier une indexation par concepts extraits à partir d'une ressource externe. Nous proposons ensuite de prendre en compte ces relations sémantiques entre les concepts dans la correspondance par un modèle de recherche d'information basé sur un réseau Bayésien des concepts et leurs relations sémantiques. Ainsi, nous étudions les extensions de l'indexation conceptuelle à des documents et requête structurés et multi-médias. Les fonctions de reclassement et de combinaison ont été proposées afin d'améliorer la performance de la recherche dans ces contextes. La validation des propositions est effectuée par des expérimentations dans la recherche d'information multilingue médicale, avec l'utilisation du méta thésaurus UMLS comme ressource externe.

With the availability of external resources to documents, the Information Retrieval Systems evolve. These resources provide not only information on the terms and concepts for a more precise indexing, but also the semantic relations between these terms or concepts. Our thesis work lies in the use of external resources in information retrieval. We study firstly the indexing conceptual in comparison with term-based indexing. The problem arrise when the documents and the query don't share the same concepts, but the concepts of the documents are semantically related with the concepts of the query. We propose to take into account these semantic relationships between concepts by a information retrieval model which is based on a Bayesian network of concepts and their semantic relationships. In addition, we propose the use of knowledge of the problem from external ressource to improve the performance of retrieval. The validation of the proposed model is achieved by experiments in the medical domaine information retrieval, with the use of meta thesaurus UMLS as external resource. The application for a system of information retrieval multi modality (text and images) was also performed.; Dans les systèmes de recherche d'information, une indexation à base de termes et une correspondance à base d'intersection introduisent le problème de la disparité à cause des variations linguistiques. Avec l'objectif de résoudre ce problème, notre travail de thèse se positionne dans l'utilisation des ressources externes dans la recherche d'information. Ces ressources offrent non seulement les concepts pour une indexation plus précise et indépendante de langue, mais aussi une base de relations sémantiques entre ces concepts. Nous étudions en premier une indexation par concepts extraits à partir d'une ressource externe. Nous proposons ensuite de prendre en compte ces relations sémantiques entre les concepts dans la correspondance par un modèle de recherche d'information basé sur un réseau Bayésien des concepts et leurs relations sémantiques. Ainsi, nous étudions les extensions de l'indexation conceptuelle à des documents et requête structurés et multi-médias. Les fonctions de reclassement et de combinaison ont été proposées afin d'améliorer la performance de la recherche dans ces contextes. La validation des propositions est effectuée par des expérimentations dans la recherche d'information multilingue médicale, avec l'utilisation du méta thésaurus UMLS comme ressource externe.

Περιέχει το πλήρες κείμενο This study aims to explore users' image seeking behaviour when searching for a known, non-annotated image in Flickling provided by iCLEF2008 track. The task assigned to users was to search for the three first images given after first login. Users did not know in advance in which of the six languages (English, German, Dutch, Spanish, French, Italian) the images were described, forcing them to search across languages. The main focus of our study was threefold: a) to identify the reasons that determined users' choice over a specific interface, b ) to examine whether users were thinking about languages when searching for images and to what extent and c) to examine if used, how helpful the translations proved to be for finding the images. This study used four different, both quantitative and qualitative methods (questionnaires, retrospective thinking aloud, observation and interviews) to meet its research questions. Results show that two out of ten users were using only the monolingual interface because they did not feel confident with languages and the rest were switching between interfaces for a variety of reasons in which languages played a small part. Only four out of ten users were actually thinking about languages when searching for the images, while the rest were more preoccupied with finding the images and completing the task successfully. As a consequence, only four users paid attention to translations and only judged the translations in languages known to them. Overall, the translations were not considered to be helpful due to their inconsistency in coverage and their tendency to lead to irrelevant results.

En este artículo se evalúa el rendimiento del algoritmo de fusión documental 2stepRSV como parte de un sistema de Búsqueda de Respuestas multilingüe. En estos sistemas el usuario puede formular su consulta en diversos idiomas, y puede obtener la respuesta a partir de colecciones documentales escritas igualmente en diversos idiomas. En concreto, el sistema propuesto denominado BRUJA soporta los idiomas inglés, español y francés tanto en la formulación de la consulta como en la base documental sobre la que se realiza la búsqueda. Se ha evaluado el rendimiento de 2stepRSV atendiendo a dos criterios: por un lado se ha comparado 2stepRSV con dos de los algoritmos de fusión documental más conocidos (Round-Robin y Raw-Scoring), y por otro se ha comparado el sistema multilingüe con el rendimiento que se habría obtenido en caso de que el idioma de la consulta y el de la respuesta fuera el mismo (búsqueda de respuestas monolingüe). En ambos casos 2stepRSV obtiene mejoras consistentes, entre un 15 % y un 35 %, respecto al rendimiento global de BRUJA. This paper presents the evaluation of the performance of the document fusion algorithm 2stepRSV as a part of a multilingual Question Answering system. The user makes questions in several languages and obtains answers extracted from document collections in different languages. The proposed system, called BRUJA, supports English, Spanish and French languages, for questions and document collections. We have evaluated the performance of the 2stepRSV approach according to the following motivations: on the one hand, we have compared the 2stepRSV algorithm with two algorithms related to the documents merging Round-Robin and Raw-Scoring); on the other hand, we have compared the results obtained with this multilingual system vs a monolingual one. In both cases, the 2stepRSV algorithm obtains consistent improvements, between 15 % and 35 %, with regard to the overall system performance. Este trabajo ha sido financiado por el Ministerio de Ciencia y Tecnología a través del proyecto TIMOM (TIN2006-15265-C06-03).

Περιέχει το πλήρες κείμενο The volume of information available on the Web in various languages and the growing number of the non-English native speaking users generate a need for developing and providing Cross Language Information Retrieval (CLIR) systems. The efficient development of such systems depends greatly on gaining a better insight into users' interactions and search behaviours in multilingual environments. In this context, the decision about which method/s to employ is considered essential. This study reports on the reasons why a mixed methods research was adopted; the decision for implementing questionnaire, observation, retrospective thinking aloud and interviews; the way in which they were combined and the values and limitations posed by such an implementation.

Περιέχει το πλήρες κείμενο This study aims to explore users' image seeking behaviour when searching for known, non-annotated images across languages in FlickLing. In particular, the investigation reported in this paper aims to demonstrate the value in focusing on user's trust and con dence in the exploration of seeking behaviour to reveal users' perceptions of the tasks involved when searching across languages. The task assigned to our 24 users was to search for three speci c images described in Dutch, Ger- man and Spanish consecutively. In this context, four di erent methods, both qualitative and quantitative (questionnaires, observation, retrospec- tive thinking aloud and interviews) were employed. Preliminary nding regarding users' switching between the two modes, users' thinking of languages and their answers to the two questionnaires on trust and con- dence are only reported here. It appears from our preliminary ndings that there is variation in our users perceptions of searching for images across languages and their approach to using translations and that this occurs regardless of the amount or type of help or guidance given.

This book constitutes the thoroughly refereed post-proceedings of the 8th Workshop of the Cross-Language Evaluation Forum, CLEF 2007, held in Budapest, Hungary, September 2007. The revised papers presented together with an introduction were carefully reviewed and selected for inclusion in the book. The papers are organized in topical sections on Multilingual Textual Document Retrieval, Domain-Specific Information Retrieval, Multiple Language Question Answering, Cross-language Retrieval in Image Collections, Cross-Language Speech retrieval, Multingual Web Retrieval, and Cross-Language Geographical retrieval.