Learning Preferences with Kernel-Based Methods

Learning of preference relations has recently received significant attention in machine learning community. It is closely related to the classification and regression analysis and can be reduced to these tasks. However, preference learning involves prediction of ordering of the data points rather than prediction of a single numerical value as in case of regression or a class label as in case of classification. Therefore, studying preference relations within a separate framework facilitates not only better theoretical understanding of the problem, but also motivates development of the efficient algorithms for the task. Preference learning has many applications in domains such as information retrieval, bioinformatics, natural language processing, etc. For example, algorithms that learn to rank are frequently used in search engines for ordering documents retrieved by the query. Preference learning methods have been also applied to collaborative filtering problems for predicting individual customer choices from the vast amount of user generated feedback. In this thesis we propose several algorithms for learning preference relations. These algorithms stem from well founded and robust class of regularized least-squares methods and have many attractive computational properties. In order to improve the performance of our methods, we introduce several non-linear kernel functions. Thus, contribution of this thesis is twofold: kernel functions for structured data that are used to take advantage of various non-vectorial data representations and the preference learning algorithms that are suitable for different tasks, namely efficient learning of preference relations, learning with large amount of training data, and semi-supervised preference learning. Proposed kernel-based algorithms and kernels are applied to the parse ranking task in natural language processing, document ranking in information retrieval, and remote homology detection in bioinformatics domain. Training of kernel-based ranking algorithms can be infeasible when the size of the training set is large. This problem is addressed by proposing a preference learning algorithm whose computation complexity scales linearly with the number of training data points. We also introduce sparse approximation of the algorithm that can be efficiently trained with large amount of data. For situations when small amount of labeled data but a large amount of unlabeled data is available, we propose a co-regularized preference learning algorithm. To conclude, the methods presented in this thesis address not only the problem of the efficient training of the algorithms but also fast regularization parameter selection, multiple output prediction, and cross-validation. Furthermore, proposed algorithms lead to notably better performance in many preference learning tasks considered.

Service Description Creation Empowerment to Mobile End Users

Increasing usage of Web Services has been result of efforts to automate Web Services discovery and interoperability. The Semantic Web Service descriptions create basis for automatic Web Service information management tasks such as discovery and interoperability. The discussion of opportunities enabled by service descriptions have arisen in recent years. The end user has been considered only as a consumer of services and information sharing occurred from one service provider to public in service distribution. The social networking has changed the nature of services. The end user cannot be seen anymore only as service consumer, because by enabling semantically rich environment and right tools, the end user will be in the future the producer of services. This study investigates the ways to provide for end users the empowerment to create service descriptions on mobile device. Special focus is given to the changed role of the end user in service creation. In addition, the Web Services technologies are presented as well as different Semantic Web Service description approaches are compared. The main focus in the study is to investigate tools and techniques to enable service description creation and semantic information management on mobile device.

Value co-creation in gaming the industry: A qualitative research on virtual discussion forums

This study discusses the procedures of value co-creation that persist in gaming industry. The purpose of this study was to identify the procedures that persist in current video gaming industry which answers the main research problem how value is co-created in video gaming industry followed by three sub questions: (i) What is value co-creation in gaming industry? (ii) Who participates in value co-creation in gaming industry? (iii) What are the procedures that are involved in value co-creation in gaming industry? The theoretical background of the study consists of literature relating to the theory of marketing i.e., notion of value, conventional understanding of value creation, value chain, co-creation approach, co-production approach. The research adopted qualitative research approach. As a platform of relationship researcher used web 2.0 tool interface. Data were collected from the social networks and netnography method was applied for analyzing them. Findings show that customer and company both co-create optimum level of value while they interact with each other and within the customers as well. However mostly the C2C interaction, discussions and dialogues threads that emerged around the main discussion facilitated to co-create value. In this manner, companies require exploiting and further motivating, developing and supporting the interactions between customers participating in value creation. Hierarchy of value co-creation processes is the result derived from the identified challenges of value co-creation approach and discussion forums data analysis. Overall three general sets and seven topics were found that explored the phenomenon of customer to customer (C2C) and business to customer (B2C) interaction/debating for value co-creation through user generated contents. These topics describe how gamer contributes and interacts in co-creating value along with companies. A methodical quest in current research literature acknowledged numerous evolving flows of value in this study. These are general management perspective, new product development and innovation, virtual customer environment, service science and service dominant logic. Overall the topics deliver various realistic and conceptual implications for using and handling gamers in social networks for augmenting customers’ value co-creation process.

Aikakauslehden tuttuuden vaikutus halukkuuteen osallistua verkkopalvelun sisällöntuotantoon

Tässä pro gradu -tutkielmassa selvitettiin, miten aikakauslehden eli emobrändin tuttuus vaikuttaa halukkuuteen osallistua verkkopalvelun eli brändilaajennuksen sisällöntuotantoon joukkoistamalla. Kysymys liittyy muuttuneeseen tilanteeseen median kulutuksessa. Teknologia on kehittynyt, paperisten mediatuotteiden kysyntä vähentynyt ja aikakauslehdet kehittävät digitaalisia palvelujaan. Medialla kuitenkaan ole yksinoikeutta sisällöntuotantoon. Kuka tahansa voi tehdä ja julkaista sisältöä. Monet mediat ovat ottaneet lukijat mukaan sisällöntuotantoon. Aiempien tutkimusten perusteella emobrändin tuttuus vaikuttaa myönteisesti asi-akkaiden halukkuuteen käyttää brändilaajennusta. Se vaikuttaa myös ostopäätökseen ja lisää luottamusta verkkopalvelua kohtaan. Tässä tutkimuksessa tuote oli tuttu niille, jotka ovat lehden tilaajia. Lukijalähtöistä sisältöä ovat mm. blogit, verkkokeskustelut, lukijoiden tarinat, runot, kuvat, kilpailut ja kyselyt. Tutkielman aineisto kerättiin verkkokyselyn avulla ja siihen vastasi 437 vastaajaa. Tulosten perusteella emobrändin tuttuus vaikuttaa myönteisesti tilaajien halukkuuteen käyttää aikakauslehden verkkopalvelua. Tilaajat käyvät palvelussa muita käyttäjiä useammin ja osallistuvat tai haluaisivat osallistua ei-tilaajia innokkaammin sisällöntuotantoon. He ovat myös ei-tilaajia kiinnostuneempia lukijalähtöisestä sisällöstä.

Future of Internet News Portals after the Case of Delfi

The thesis analyzes liability of Internet news portals for third-party defamatory comments. After the case of Delfi AS v. Estonia, decided by the Grand Chamber of the European Court of Human Rights on 16 June 2015, a portal can be held liable for user-generated unlawful comments. The thesis aims at exploring consequences of the case of Delfi for Internet news portals’ business model. The model is described as a mixture of two modes of information production: traditional industrial information economy and new networked information economy. Additionally, the model has a generative comment environment. I name this model “the Delfian model”. The thesis analyzes three possible strategies which portals will likely apply in the nearest future. I will discuss these strategies from two perspectives: first, how each strategy can affect the Delfian model and, second, how changes in the model can, in their turn, affect freedom of expression. The thesis is based on the analysis of case law, legal, and law and economics literature. I follow the law and technology approach in the vein of ideas developed by Lawrence Lessig, Yochai Benkler and Jonathan Zittrain. The Delfian model is researched as an example of a local battle between industrial and networked information economy modes. The thesis concludes that this local battle is lost because the Delfian model has to be replaced with a new walled-garden model. Such a change can seriously endanger freedom of expression.

Analysis of Automotive Cybersecurity - Attack surfaces and users’ privacy concerns

Modern automobiles are no longer just mechanical tools. The electronics and computing services they are shipping with are making them not less than a computer. They are massive kinetic devices with sophisticated computing power. Most of the modern vehicles are made with the added connectivity in mind which may be vulnerable to outside attack. Researchers have shown that it is possible to infiltrate into a vehicle’s internal system remotely and control the physical entities such as steering and brakes. It is quite possible to experience such attacks on a moving vehicle and unable to use the controls. These massive connected computers can be life threatening as they are related to everyday lifestyle. First part of this research studied the attack surfaces in the automotive cybersecurity domain. It also illustrated the attack methods and capabilities of the damages. Online survey has been deployed as data collection tool to learn about the consumers’ usage of such vulnerable automotive services. The second part of the research portrayed the consumers’ privacy in automotive world. It has been found that almost hundred percent of modern vehicles has the capabilities to send vehicle diagnostic data as well as user generated data to their manufacturers, and almost thirty five percent automotive companies are collecting them already. Internet privacy has been studies before in many related domain but no privacy scale were matched for automotive consumers. It created the research gap and motivation for this thesis. A study has been performed to use well established consumers privacy scale – IUIPC to match with the automotive consumers’ privacy situation. Hypotheses were developed based on the IUIPC model for internet consumers’ privacy and they were studied by the finding from the data collection methods. Based on the key findings of the research, all the hypotheses were accepted and hence it is found that automotive consumers’ privacy did follow the IUIPC model under certain conditions. It is also found that a majority of automotive consumers use the services and devices that are vulnerable and prone to cyber-attacks. It is also established that there is a market for automotive cybersecurity services and consumers are willing to pay certain fees to avail that.

La jurisprudence en accès libre à l'ère du contenu généré par les usagers

La collaboration et le contenu généré par les usagers, aussi appelé « Web 2. 0 », sont des phénomènes nouveaux, qui bâtissent sur l'ouverture et le foisonnement d'Internet. Les environnements numériques qui emploient ces moyens mettent à contribution la communauté qui gravite autour d'une présence virtuelle afin d'en enrichir l’expérience. Suivant une approche constructiviste, nous explorons commnent la collaboration peut servir les usagers d'une banque de donnée de jugements en accès libre par Internet, comme le site de l'Institut canadien d'information juridique (www.CanLIT.org). La collaboration s'articule grâce à un gabarit d'analyse que nous nommons «Cadre de diffusion de la collaboration». Il comporte deux classes d'objets, les usagers et les documents, qui interagissent selon quatre relations : les liens documentaires, les échanges entre usagers, l'écriture (de l'usager vers le document) et la consommation (du document vers l'usager). Le Cadre de diffusion de la collaboration met en lumière les modalités de la collaboration comme mécanisme de création de contenu dans un contexte numérique, au profit d'une classe de documents. Suite à une analyse les modalités de la jurisprudence comme système documentaire et d'un exposé illustratif des besoins des usagers de la société civile, le Cadre de diffusion de la collaboration est employé pour explorer les mécanismes à retenir pour enrichir le contenu d'un système diffusant des jugements par Internet.

La diffamation sur Internet : actualiser la responsabilité en droit civil et en common law au Canada

En cette ère que plusieurs surnomment le « Web 2.0 », les usagers se sont emparés avec enthousiasme des fonctions liées aux communications et au partage sur Internet, ce médium devenant ainsi une nouvelle plate-forme pour les enjeux liés à la vie privée et à la réputation. La diffamation constitue justement un des problèmes prédominants constatés en lien avec ce contenu électronique, plus particulièrement lorsqu’il est question de contenu généré par les utilisateurs. Face à cet outil permettant une diffusion et une intéractivité sans précédent, comment devons-nous aborder Internet au regard des règles de droit applicables au Canada en matière de diffamation? L’analyse juridique traditionnelle sied-elle aux nouvelles réalités introduites par ce médium? Le bijuridisme canadien nous impose d’étudier parallèlement les régimes de droit civil et de common law et ce, dans une optique comparative afin de comprendre les concepts et le fonctionnement propres à chacune des approches juridiques cohabitant au pays. Cette analyse nous permettra de mettre en lumière les particularités du médium électronique qui se révèlent pertinentes lorsqu’il est question de diffamation et qui font la spécificité des situations et des acteurs en ligne, distinguant ainsi Internet des modes de communications traditionnels que le droit connaît. Cette approche comparative permet de poser un regard critique sur chacun des régimes de droit en vigueur au Canada, considérant la réalité propre à Internet et au contenu généré par les utilisateurs, mais surtout, vise à promouvoir le développement de méthodes d’analyse véritablement ancrées dans le fonctionnement du médium en cause et susceptibles d’évoluer avec celui-ci.

UEBERflow

In der Dissertation wird der Frage nachgegangen, welche globalen bildungspolitischen Maßnahmen erforderlich sind, um auch bislang exkludierten Menschen den Kompetenzerwerb zu ermöglichen, der benötigt wird, eine positive User Experience in benutzergenerierten, digitalen Lernumgebungen auszubilden, damit sie an der modernen Weltgesellschaft selbstbestimmt teilhaben können. Zu diesem Zweck wurden Castells ‘Netzwerkgesellschaft’ und Csikszentmihalys ‘Theorie der optimalen Erfahrung’ als analytische Grundlagen zur Einordnung der sozialen Netzwerk-Aktivitäten herangezogen. Dies ermöglichte es, unter Rückgriff auf aktuelle Lerntheorien, Kompetenzdebatten, ökonomische Analysen des Bildungssystems und User Experience-Forschungen, einige individuelle und gesamtgesellschaftliche Voraussetzungen abzuleiten, um in der Netzwerkgesellschaft konstruktiv überleben zu können. Mit Blick auf unterschiedliche sozio-kulturelle Bedingungen für persönlichen Flow im ‘space of flows’ liessen sich schließlich differenzierte Flow-Kriterien entwickeln, die als Grundlage für die Operationalisierung im Rahmen einer Real-Time Delphi (RTD)-Studie mit einem internationalen Expertinnen-Panel dienen konnten. Ziel war es, bildungspolitische Ansatzpunkte zu finden, den bislang Exkludierten bis zum Jahre 2020 erste Rahmenbedingungen zu bieten, damit sie potentiell teilhaben können an der Gestaltung der zukünftigen Netzwerkgesellschaft. Das Ergebnis der Expertinnen-Befragung wurde unter Rückgriff auf aktuelle Global und Educational Governance-Studien und das Einflusspotenzial der Zivilgesellschaft auf den Digital Divide reflektiert. Vor diesem Hintergrund konnten abschließend vier bildungspolitische Verlaufsszenarien entworfen werden, die es ermöglichen könnten, bis 2020 die Kluft zu den global Exkludierten wenigstens etwas zu schließen.

Mobiles Lernen

In ihrer Arbeit "Mobiles Lernen. Analyse des Wissenschaftsprozesses der britischen und deutschsprachigen medienpädagogischen und erziehungswissenschaftlichen Mobile-Learning-Diskussion." zeichnet Judith Seipold die Phasen und Entwicklungslinien der bisherigen und naturwüchsigen Diskussion um das Mobile Lernen nach. Dabei eröffnet sie mit ihrer Analyse des vornehmlich britischen Wissenschaftsprozesses des Mobilen Lernens ab Beginn des 21. Jahrhunderts den Blick auf die Struktur der medienpädagogischen und erziehungswissenschaftlichen Mobile-Learning-Diskussion, auf deren Kontexte, Bezugspunkte, Perspektiven und theoretische Schwerpunkte, aber auch Erfolge und Problembereiche in der praktischen Umsetzung des Lernens mit Mobiltechnologien in formalisierten Lernkontexten wie dem des Schulunterrichts. Auf diese Weise liefert die Autorin eine Systematik, die nicht nur die britische Diskussion für die deutschsprachige Medienpädagogik verfügbar macht, sondern auch eine neue und systematisch begründete Reflexionsebene für eine aktuelle medienpädagogische Entwicklung – das Mobile Lernen – bildet. Zunächst widmet sich die Autorin der Analyse des Wissenschaftsprozesses der bisherigen britischen und teils auch der deutschsprachigen Mobile-Learning-Diskussion. Um diesen Prozess in seiner Struktur greifbar zu machen, skizziert sie die Bezugsdisziplinen, aus denen die medienpädagogische und erziehungswissenschaftliche Mobile-Learning-Forschung schöpft, arbeitet die Legitimationsbasis auf, auf die sich Mobiles Lernen stützt, zeichnet die historische Entwicklung der Diskussion nach, die sowohl Alltagsnutzung von Mobiltechnologien als auch den Fachdiskurs einschließt, und erläutert innerhalb ihres Modells der Phasen und Entwicklungslinien Mobilen Lernens Theorien, Konzepte und Modelle, die in der Mobile-Learning-Community als zentral für Analyse und Planung Mobilen Lernens erachtet werden. Konzepte und Modelle, die in der Mobile-Learning-Diskussion die Rolle der Lernenden in das Zentrum der Betrachtungen rücken und eine explizite Handlungsorientierung unterstützen, sind Kernbereich des theoriebasierten Teils der Arbeit. Überlegungen zu einer „Sozio-kulturellen Ökologie Mobilen Lernens“, zu „user-generated contexts“ und zu einer „kulturökologisch informierten Didaktik des Mobilen Lernens“ sind dabei sowohl reflektierende Einordnung aktueller Konzepte und Modelle als auch theoretische und konzeptionelle Basis für die praktische Umsetzung Mobilen Lernens. Um die Rolle der Handlungskompetenzen, kulturellen Praktiken und Strukturen der Lerner bei der Nutzung von Mobiltechnologien für Lernen analytisch zu fassen, finden im Praxis-Kapitel Planungs- und Analyseschemata unter Rückgriff auf Beispiele aus der Praxisforschung und der Implementierung mobilen Lernens in den Schulunterricht Anwendung. Ihre Erörterungen rahmt die Autorin abschließend kritisch und weist zum einen auf bildungspolitische, strukturelle und handlungspraktische Implikationen hin, die sich aus der Mobile-Learning-Diskussion ergeben; zum anderen lenkt sie das Augenmerk auf Gegensätze und Dialektiken des Mobilen Lernens, die sich im Spannungsfeld zwischen alltäglicher mobiler Mediennutzung und der teils theoretisch informierten schulischen Verwendung von Mobiltechnologien zum Lernen entfalten. Sie gilt es, im weiteren Verlauf der Diskussion zu hinterfragen und aufzulösen.

WAIS seminar: Linked Cultural Media

Wednesday 9th April 2014 Speaker(s): Guus Schreiber Time: 09/04/2014 11:00-11:50 Location: B32/3077 File size: 546Mb Abstract In this talk I will discuss linked data for museums, archives and libraries. This area is known for its knowledge-rich and heterogeneous data landscape. The objects in this field range from old manuscripts to recent TV programs. Challenges in this field include common metadata schema's, inter-linking of the omnipresent vocabularies, cross-collection search strategies, user-generated annotations and object-centric versus event-centric views of data. This work can be seen as part of the rapidly evolving field of digital humanities. Speaker Biography Guus Schreiber Guus is a professor of Intelligent Information Systems at the Department of Computer Science at VU University Amsterdam. Guus’ research interests are mainly in knowledge and ontology engineering with a special interest for applications in the field of cultural heritage. He was one of the key developers of the CommonKADS methodology. Guus acts as chair of W3C groups for Semantic Web standards such as RDF, OWL, SKOS and REFa. His research group is involved in a wide range of national and international research projects. He is now project coordinator of the EU Integrated project No Tube concerned with integration of Web and TV data with the help of semantics and was previously Scientific Director of the EU Network of Excellence “Knowledge Web”.

The Agendas of The Stream

The new tech has introduced the traits of the hypermedia, interactivity and convergence in the journalism field. Moreover, it has involved the emergence of new media and changes in the informative contents. In 2006 Al Jazeera in English was created for covering the underreported regions, inasmuch as new media arise, new contents do so, thus, the research examines the agendas of The Stream, the TV show from Al Jazeera that relies heavily on social networks.

Branding in the post-internet era

The internet and its related e-technologies have to a large extent upset the asymmetry of information that for so many years worked in favour of brand managers. Consumers are now empowered to interact with brands and other consumers but also to create their own content on user generated content sites leading to a more participative approach to branding. Internet brands adopt a more relaxed stance on brand management, which involves the consumer in fundamental stages of the brand building process. In this context, the brand manager is no longer a `guardian' of the brand but becomes more of a brand `host'. The question is to what extent can traditional companies follow suit? Are they comfortable to cede control to consumers? Do we need a new theory of branding in an e-space?

Watch-and-Comment as a Paradigm toward Ubiquitous Interactive Video Editing

The literature reports research efforts allowing the editing of interactive TV multimedia documents by end-users. In this article we propose complementary contributions relative to end-user generated interactive video, video tagging, and collaboration. In earlier work we proposed the watch-and-comment (WaC) paradigm as the seamless capture of an individual`s comments so that corresponding annotated interactive videos be automatically generated. As a proof of concept, we implemented a prototype application, the WACTOOL, that supports the capture of digital ink and voice comments over individual frames and segments of the video, producing a declarative document that specifies both: different media stream structure and synchronization. In this article, we extend the WaC paradigm in two ways. First, user-video interactions are associated with edit commands and digital ink operations. Second, focusing on collaboration and distribution issues, we employ annotations as simple containers for context information by using them as tags in order to organize, store and distribute information in a P2P-based multimedia capture platform. We highlight the design principles of the watch-and-comment paradigm, and demonstrate related results including the current version of the WACTOOL and its architecture. We also illustrate how an interactive video produced by the WACTOOL can be rendered in an interactive video environment, the Ginga-NCL player, and include results from a preliminary evaluation.

Volunteered Geographic Information For Water Management: A Prototype Architecture

Driven by Web 2.0 technology and the almost ubiquitous presence of mobile devices, Volunteered Geographic Information (VGI) is knowing an unprecedented growth. These notable technological advancements have opened fruitful perspectives also in the field of water management and protection, raising the demand for a reconsideration of policies which also takes into account the emerging trend of VGI. This research investigates the opportunity of leveraging such technology to involve citizens equipped with common mobile devices (e.g. tablets and smartphones) in a campaign of report of water-related phenomena. The work is carried out in collaboration with ADBPO - Autorità di bacino del fiume Po (Po river basin Authority), i.e. the entity responsible for the environmental planning and protection of the basin of river Po. This is the longest Italian river, spreading over eight among the twenty Italian Regions and characterized by complex environmental issues. To enrich ADBPO official database with user-generated contents, a FOSS (Free and Open Source Software) architecture was designed which allows not only user field-data collection, but also data Web publication through standard protocols. Open Data Kit suite allows users to collect georeferenced multimedia information using mobile devices equipped with location sensors (e.g. the GPS). Users can report a number of environmental emergencies, problems or simple points of interest related to the Po river basin, taking pictures of them and providing other contextual information. Field-registered data is sent to a server and stored into a PostgreSQL database with PostGIS spatial extension. GeoServer provides then data dissemination on the Web, while specific OpenLayers-based viewers were built to optimize data access on both desktop computers and mobile devices. Besides proving the suitability of FOSS in the frame of VGI, the system represents a successful prototype for the exploitation of user local, real-time information aimed at managing and protecting water resources.