This thesis explores the use of multimodality to communicate retrieved results of Internet search engines. The investigation aimed to investigate suitable multimodal metaphors which would increase the level of usability of Internet search engine interfaces and enhance users` experience in the search activity. The study consisted of three experiments based on questionnaires and Internet search activities with a set of text-based and multimodal interfaces. These interfaces were implemented in two browsing platforms, named AVBRO and AVBRO II. In the first experiment, the efficiency of specific multimodal metaphors to communicate additional information of retrieved results was investigated. The experiment also sought to obtain users` views of these metaphors with a questionnaire. An experimental multimodal interface of the AVBRO platform, which communicated additional information with a combination of three 2D graphs and musical stimuli, was used as a basis for the experiment, together with the Google search engine. The results obtained led to the planning of a second experiment. The aim of this experiment was to obtain and compare the level of usability of four different experimental multimodal interfaces and one traditional text-based interface, all implemented in the AVBRO II platform. Effectiveness, efficiency and users` satisfaction were used as criteria to evaluate the usability of these interfaces. In the third and final experiment the usability analysis of a traditional text-based interface and the two most suitable experimental multimodal interfaces of the AVBRO II platform was further investigated. Learnability, errors rate, efficiency, memorability and users` satisfaction were used as criteria to evaluate the usability of these interfaces. The analysis of the results obtained from these experiments provided the basis for a set of design guidelines for the development of usable interfaces based on a multimodal approach.

<p>This thesis discusses the possibilities to visualise the presentation layer of a search engine on a mobile device in an alternative way. Previous work in the area has shown that the use of text-based-lists can be problematic when accessed on a device with a limited display. In the scope of the thesis and in order to tackle the current problems when displaying the results a literature review was carried out. The findings of the review formed the basis for a requirement definition on which a mock-up was developed. The mock-up was then evaluated and tested during a usability test where a number of users got to experience the alternative presentation layer that uses a visualisation technique called tree- map. The results from the test show that the mock-up could be seen as a alternative to the current presentation of results. The mock-up also shows that a future implementation could also include the use of categories and sorting of information in order to provide content with a meaning.</p>

Au cours des dernières années les progrès technologiques permettant de collecter, stocker et traiter d'importantes quantités de données pour un faible coût, ont soulevés de sérieux problèmes concernant la vie privée. La protection de la vie privée concerne de nombreux domaines, en particulier les sites internet fréquemment utilisés comme les moteurs de recherche (ex. : Google, Bing, Yahoo!). Ces services permettent aux utilisateurs de retrouver efficacement du contenu sur Internet en exploitant leurs données personnelles. Dans ce contexte, développer des solutions pour permettre aux utilisateurs d'utiliser ces moteurs de recherche tout en protégeant leurs vies privées est devenu primordial. Dans cette thèse, nous introduirons SimAttack, une attaque contre les solutions protégeant la vie privée de l'utilisateur dans ses interactions avec les moteurs de recherche. Cette attaque vise à retrouver les requêtes initialement envoyées par l'utilisateur. Nous avons montré avec cette attaque que trois mécanismes représentatifs de l’état de l’art ne sont pas satisfaisants pour protéger la vie privée des utilisateurs. Par conséquent, nous avons développé PEAS, un nouveau mécanisme de protection qui améliore la protection de la vie privée de l'utilisateur. Cette solution repose sur deux types de protection : cacher l'identité de l'utilisateur (par une succession de deux serveurs) et masquer sa requête (en la combinant avec des fausses requêtes). Afin de générer des fausses requêtes réalistes, PEAS se base sur les précédentes requêtes envoyées par les utilisateurs du système. Pour finir, nous présenterons des mécanismes permettant d'identifier la sensibilité des requêtes. Notre objectif est d'adapter les mécanismes de protection existants pour protéger uniquement les requêtes sensibles, et ainsi économiser des ressources (ex. : CPU, mémoire vive). Nous avons développé deux modules pour identifier les requêtes sensibles. En déployant ces modules sur des mécanismes de protection existants, nous avons établi qu'ils permettent d'améliorer considérablement leurs performances. / During the last few years, the technological progress in collecting, storing and processing a large quantity of data for a reasonable cost has raised serious privacy issues. Privacy concerns many areas, but is especially important in frequently used services like search engines (e.g., Google, Bing, Yahoo!). These services allow users to retrieve relevant content on the Internet by exploiting their personal data. In this context, developing solutions to enable users to use these services in a privacy-preserving way is becoming increasingly important. In this thesis, we introduce SimAttack an attack against existing protection mechanism to query search engines in a privacy-preserving way. This attack aims at retrieving the original user query. We show with this attack that three representative state-of-the-art solutions do not protect the user privacy in a satisfactory manner. We therefore develop PEAS a new protection mechanism that better protects the user privacy. This solution leverages two types of protection: hiding the user identity (with a succession of two nodes) and masking users' queries (by combining them with several fake queries). To generate realistic fake queries, PEAS exploits previous queries sent by the users in the system. Finally, we present mechanisms to identify sensitive queries. Our goal is to adapt existing protection mechanisms to protect sensitive queries only, and thus save user resources (e.g., CPU, RAM). We design two modules to identify sensitive queries. By deploying these modules on real protection mechanisms, we establish empirically that they dramatically improve the performance of the protection mechanisms.

O modelo de recuperação de informação mais amplamente utilizado no contexto de acervos digitais é o Vector Space Model. Algoritmos implementados para este modelo que aproveitam informações sobre relevância obtidas dos usuários (chamados feedbacks) na tentativa de melhorar os resultados da busca. Porém, estes algoritmos de feedback de relevância não possuem uma estratégia global e permanente, as informações obtidas desses feedbacks são descartadas para cada nova sessão de usuário (são perenes) ou não modificam os documentos como um todo (são alterações locais). Este trabalho apresenta um método de feedbacks de relevância denominado orientado a termos, permitindo que as modificações realizadas por influência dos feedbacks dos usuários sejam globais e permanentes. Foram realizados experimentos utilizando o dataset ClueWeb09 que dão evidências de que este método melhora a qualidade dos resultados da busca em relação ao modelo tradicional Vector Space Model. / The Vector Space Model is the most widely used information retrieval model within digital libraries\' systems. Algorithms developed to be used with this model use relevance information obtained from users (called feedbacks) to improve the search results. However, the relevance feedback algorithms developed are not global nor permanent, the feedbacks are discarded in users new sessions and do not affect every document. This paper presents a method that uses of relevance feedback named terms oriented. In this method, users\' feedbacks lead to modifications in the terms\' vectors representations. These modifications are global and permanent, influencing further searches. An experiment was conducted using the ClueWeb09 dataset, giving evidence that this method improves the quality of search results when compared with Vector Space Model.

A localização de páginas relevantes na Internet em atividades como clipping de notícias, detecção de uso indevido de marcas ou em serviços anti-phishing torna-se cada vez mais complexa devido a vários fatores como a quantidade cada vez maior de páginas na Web e a grande quantidade de páginas irrelevantes retornadas por mecanismos de busca. Em muitos casos as técnicas tradicionais utilizadas em mecanismos de busca na Internet, isto é, localização de termos em páginas e ordenação por relevância, não são suficientes para resolver o problema de localização de páginas específicas em atividades como as citadas anteriormente. A contextualização das páginas, ou seja, a classificação de páginas segundo um contexto definido pelo usuário baseando-se nas necessidades de uma atividade específica deve permitir uma busca mais eficiente por páginas na Internet. Neste trabalho é estudada a utilização de métodos de mineração na Web para a composição de métodos de contextualização de páginas, que permitam definir contextos mais sofisticados como seu assunto ou alguma forma de relacionamento. A contextualização de páginas deve permitir a solução de vários problemas na busca de páginas na Internet pela composição de métodos, que permitam a localização de páginas através de um conjunto de suas características, diferentemente de mecanismos de busca tradicionais que apenas localizam páginas que possuam um ou mais termos especificados. / Internet services as news clipping service, anti-phising, anti-plagiarism service and other that require intensive searching in Internet have a difficult work, because of huge number of existing pages. Search Engines try driver this problem, but search engines methods retrieve a lot of irrelevant pages, some times thousands of pages and more powerful methods are necessary to drive this problem. Page content, subject, hyperlinks or location can be used to define page context and create a more powerful method that can retrieve more relevant pages, improving precision. Classification of page context is defined as classification of a page by a set of its feature. This report presents a study about Web Mining, Search Engines and application of web mining technologies to classify page context. Page context classification applied to search engines must solve the problem of irrelevant pages flood by allowing search engines retrieve pages of a context.

Esta tese reflete sobre um processo específico: a transmídia e seu potencial de criar novas possibilidades interativas e artísticas apropriando-se dos ambientes digitais. A disseminação tecnológica no cotidiano das pessoas nas últimas décadas alavanca uma proliferação e diversidade de mídias que propiciam um ambiente fértil para desenvolvimento de processos transmidiáticos. Apesar de a transmídia não ser um fenômeno novo, as novas possibilidades que o ambiente digital hiperconectado e rico em novas mídias e interfaces traz são inúmeras, não apenas nas áreas das artes, mas também em quaisquer áreas que envolvam comunicação, informação e interação. Novos modos de interação trazem consigo novas possibilidades, tanto quanto novos desafios. O foco deste trabalho é o contexto atual transmidiático e as oportunidades e desafios emergentes de seu uso. O escopo compreende também a apresentação e análise de algumas obras de arte que se beneficiaram do uso da transmídia como processo fundamental para seu desenvolvimento, razão pela qual não seriam possíveis em outro cenário que não o transmidiático atual. Foram desenvolvidos pela autora quatro trabalhos de arte como experimentação transmídia que refletem os resultados obtidos. / This thesis reflects about a particular process: the transmedia and its potential to create new interactive and artistic possibilities that appropriate from the digital environments. The technological dissemination in people\'s daily lives in the past decades leverages a media proliferation and diversification that propitiate a fertile environment for the development of transmedia processes. Although transmedia is not a new phenomenon, the new possibilities that the actual digital environment provide -- hiper-connected and rich in new media and interfaces -- are countless, not only in the field of Arts, but also in any other areas related to communication, information and interaction. New interactive modes bring with themselves new possibilities as much as new challenges. The focus of this work is the emergent transmedia context and the opportunities and challenges that rise from its use. The scope here also involves the presentation and analysis of some artworks that take advantage from the use of transmedia as a fundamental process to their development, not being able to exist in any other scenario than the present one. It was developed by the author four artworks as transmedia experimentation that reflect the achieved results.

[pt] O crescimento explosivo e a acessibilidade generalizada da World Wide Web (WWW) levaram ao aumento da atividade de pesquisa na área da recuperação de informação para páginas Web. A WWW é um rico e imenso ambiente em que as páginas se assemelham a uma comunidade grande de elementos conectada através de hiperlinks em razão da semelhança entre o conteúdo das páginas, a popularidade da página, a autoridade sobre o assunto e assim por diante, sabendo-se que, em verdade, quando um autor de uma página a vincula à outra, está concebendo-a como importante para si. Por isso, a estrutura de hiperlink da WWW é conhecida por melhorar significativamente o desempenho das pesquisas para além do uso de estatísticas de distribuição simples de texto. Nesse sentido, a abordagem Hyperlink Induced Topic Search (HITS) introduz duas categorias básicas de páginas Web, hubs e autoridades, que revelam algumas informações semânticas ocultas a partir da estrutura de hiperlink. Em 2005, fizemos uma primeira extensão do HITS, denominada de Extended Hyperlink Induced Topic Search (XHITS), que inseriu duas novas categorias de páginas Web, quais sejam, novidades e portais. Na presente tese, revisamos o XHITS, transformando-o em uma generalização do HITS, ampliando o modelo de duas categorias para várias e apresentando um algoritmo eficiente de aprendizagem de máquina para calibrar o modelo proposto valendo-se de múltiplas categorias latentes. As descobertas aqui expostas indicam que a nova abordagem de aprendizagem fornece um modelo XHITS mais preciso. É importante registrar, por fim, que os experimentos realizados com a coleção ClueWeb09 25TB de páginas da WWW, baixadas em 2009, mostram que o XHITS pode melhorar significativamente a eficácia da pesquisa Web e produzir resultados comparáveis aos do TREC 2009/2010 Web Track, colocando-o na sexta posição, conforme os resultados publicados. / [en] The rapid growth and generalized accessibility of the World Wide Web (WWW) have led to an increase in research in the field of the information retrieval for Web pages. The WWW is an immense and prodigious environment in which Web pages resemble a huge community of elements. These elements are connected via hyperlinks on the basis of similarity between the content of the pages, the popularity of a given page, the extent to which the information provided is authoritative in relation to a given field etc. In fact, when the author of a Web page links it to another, s/he is acknowledging the importance of the linked page to his/her information. As such the hyperlink structure of the WWW significantly improves research performance beyond the use of simple text distribution statistics. To this effect, the HITS approach introduces two basic categories of Web pages, hubs and authorities which uncover certain hidden semantic information using the hyperlink structure. In 2005, we made a first extension of HITS, called Extended Hyperlink Induced Topic Search (XHITS), which inserted two new categories of Web pages, which are novelties and portals. In this thesis, we revised the XHITS, transforming it into a generalization of HITS, broadening the model from two categories to various and presenting an efficient machine learning algorithm to calibrate the proposed model using multiple latent categories. The findings we set out here indicate that the new learning approach provides a more precise XHITS model. It is important to note, in closing, that experiments with the ClueWeb09 25TB collection of Web pages, downloaded in 2009, demonstrated that the XHITS is capable of significantly improving Web research efficiency and producing results comparable to those of the TREC 2009/2010 Web Track.

Hu Ke. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (leaves 56-58). / Abstracts in English and Chinese.

Qin An. / Thesis (Ph.D.)--Chinese University of Hong Kong, 1998. / Includes bibliographical references (p. 135-142). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Mode of access: World Wide Web. / Abstracts in English and Chinese.

Wang, Wei. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (p. 106-122). / Abstracts in English and Chinese. / Abstract --- p.ii / Acknowledgement --- p.vi / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Overview --- p.1 / Chapter 1.2 --- Thesis Contributions --- p.5 / Chapter 1.3 --- Thesis Organization --- p.8 / Chapter 2 --- Background and Literature Review --- p.9 / Chapter 2.1 --- Label Ranking in Machine Learning --- p.11 / Chapter 2.1.1 --- Label Ranking --- p.11 / Chapter 2.1.2 --- Semi-Supervised Learning --- p.12 / Chapter 2.1.3 --- The Development of Label Ranking --- p.14 / Chapter 2.2 --- Question Retrieval in Community Question Answering --- p.16 / Chapter 2.2.1 --- Question Retrieval --- p.16 / Chapter 2.2.2 --- Basic Question Retrieval Models --- p.18 / Chapter 2.2.3 --- The Development of Question Retrieval Models --- p.21 / Chapter 2.3 --- Ranking through CTR by Building Click Models --- p.24 / Chapter 2.3.1 --- Click Model's Importance --- p.24 / Chapter 2.3.2 --- A Simple Example of Click Model --- p.25 / Chapter 2.3.3 --- The Development of Click Models --- p.27 / Chapter 3 --- Semi-Supervised Label Ranking --- p.30 / Chapter 3.1 --- Motivation: The Limitations of Supervised Label Ranking --- p.30 / Chapter 3.2 --- Label Ranking and Semi-Supervised Learning Framework --- p.32 / Chapter 3.2.1 --- Label Ranking and Semi-Supervised Learning Setup --- p.32 / Chapter 3.2.2 --- Information Gain Decision Tree for Label Ranking --- p.37 / Chapter 3.2.3 --- Instance Based Label Ranking --- p.39 / Chapter 3.2.4 --- Mallows Model Decision Tree for Label Ranking --- p.40 / Chapter 3.3 --- Experiments --- p.40 / Chapter 3.3.1 --- Dataset Description --- p.41 / Chapter 3.3.2 --- Experimental Results --- p.42 / Chapter 3.3.3 --- Discussion --- p.42 / Chapter 3.4 --- Summary --- p.44 / Chapter 4 --- An Application of Label Ranking --- p.45 / Chapter 4.1 --- Motivation: The Limitations of Traditional Question Retrieval --- p.45 / Chapter 4.2 --- Intention Detection Using Label Ranking --- p.47 / Chapter 4.2.1 --- Question Intention Detection --- p.48 / Chapter 4.2.2 --- Label Ranking Algorithms --- p.50 / Chapter 4.2.3 --- Some Other Learning Algorithms --- p.53 / Chapter 4.3 --- Improved Question Retrieval Using Label Ranking --- p.54 / Chapter 4.3.1 --- Question Retrieval Models --- p.55 / Chapter 4.3.2 --- Improved Question Retrieval Model --- p.55 / Chapter 4.4 --- Experimental Setup --- p.56 / Chapter 4.4.1 --- Experiment Objective --- p.56 / Chapter 4.4.2 --- Experiment Design --- p.56 / Chapter 4.4.3 --- DataSet Description --- p.57 / Chapter 4.4.4 --- Question Feature --- p.59 / Chapter 4.5 --- Experiment Result and Comments --- p.60 / Chapter 4.5.1 --- Question Classification --- p.60 / Chapter 4.5.2 --- Classification Enhanced Question Retrieval --- p.63 / Chapter 4.6 --- Summary --- p.69 / Chapter 5 --- Ranking by CTR in Click Models --- p.71 / Chapter 5.1 --- Motivation: The Relational Influence's Importance in Click Models --- p.71 / Chapter 5.2 --- Click Models in Sponsored Search --- p.75 / Chapter 5.2.1 --- A Brief Review on Click Models --- p.76 / Chapter 5.3 --- Collaborating Influence Identification from Data Analysis --- p.77 / Chapter 5.3.1 --- Quantity Analysis --- p.77 / Chapter 5.3.2 --- Psychology Interpretation --- p.82 / Chapter 5.3.3 --- Applications Being Influenced --- p.82 / Chapter 5.4 --- Incorporating Collaborating Influence into CCM . --- p.83 / Chapter 5.4.1 --- Dependency Analysis of CCM --- p.83 / Chapter 5.4.2 --- Extended CCM --- p.84 / Chapter 5.4.3 --- Algorithms --- p.85 / Chapter 5.5 --- Incorporating Collaborating Influence into TCM . --- p.87 / Chapter 5.5.1 --- TCM --- p.87 / Chapter 5.5.2 --- Extended TCM --- p.88 / Chapter 5.5.3 --- Algorithms --- p.88 / Chapter 5.6 --- Experiment --- p.90 / Chapter 5.6.1 --- Dataset Description --- p.90 / Chapter 5.6.2 --- Experimental Setup --- p.91 / Chapter 5.6.3 --- Evaluation Metrics --- p.91 / Chapter 5.6.4 --- Baselines --- p.92 / Chapter 5.6.5 --- Performance on RMS --- p.92 / Chapter 5.6.6 --- Performance on Click Perplexity --- p.93 / Chapter 5.6.7 --- Performance on Log-Likelihood --- p.93 / Chapter 5.6.8 --- Significance Discussion --- p.98 / Chapter 5.6.9 --- Sensitivity Analysis --- p.98 / Chapter 5.7 --- Summary --- p.102 / Chapter 6 --- Conclusion and Future Work --- p.103 / Chapter 6.1 --- Conclusion --- p.103 / Chapter 6.2 --- Future Work --- p.105 / Bibliography --- p.106