Mémoire (M.Ling.)--Université du Québec à Chicoutimi, 2001. / Bibliogr.: pp. 124-127. Document électronique également accessible en format PDF. CaQCU

Thesis (Ph. D.)--Miami University, Dept. of Psychology, 2009. / Title from second page of PDF document. Includes bibliographical references (p. 26-31).

Thesis (Ph. D.)--Drexel University, 1994. / eContent provider-neutral record in process. Description based on print version record.

Thesis (Ph. D.)--University of Chicago, Dept. of Linguistics, August 1999. / Includes bibliographical references. Also available on the Internet.

Thesis (Ph. D.)--University of Wisconsin--Madison, 1970. / Typescript. Vita. Description based on print version record. Includes bibliographical references.

Thesis (Ph.D.)--Ohio State University, 2005. / Title from first page of PDF file. Document formatted into pages; contains xii, 110 p.; also includes graphics. Includes bibliographical references (p. 103-110). Available online via OhioLINK's ETD Center

Ζούμε μία εποχή τεχνολογικών εξελίξεων και τεχνολογικών αλμάτων με το Διαδίκτυο να γίνεται ένας από τους βασικότερους εκφραστές των νέων τεχνολογικών τάσεων. Ωστόσο, ο τρόπος λειτουργίας του και δόμησής του παρουσιάζει εξαιρετικά μεγάλη ανομοιογένεια με αποτέλεσμα οι χρήστες να βρίσκονται συχνά μπροστά από αδιέξοδο στην προσπάθεια αναζήτησης πληροφορίας. Άλλωστε η ύπαρξη εκατομμυρίων domains οδηγεί σε δυσκολίες κατά την αναζήτηση πληροφορίας. Η έρευνα που πραγματοποιείται επικεντρώνεται στους δικτυακούς τόπους που αποτελούν πηγές ενημέρωσης και πιο συγκεκριμένα στα ειδησεογραφικά πρακτορεία ειδήσεων, αλλά και στα blogs. Μία απλή αναζήτηση αποκάλυψε περισσότερους από 40 δικτυακούς τόπους από μεγάλα ειδησεογραφικά πρακτορεία στην Αμερική. Αυτό σημαίνει πως στην προσπάθεια αναζήτησης μίας είδησης και δη, όλων των πτυχών της, κάποιος θα πρέπει να επισκεφθεί αν όχι όλους, τους περισσότερους από αυτούς τους δικτυακούς τόπους για να εντοπίσει στοιχεία για το θέμα που τον ενδιαφέρει. Σε αυτό το «πρόβλημα» ή έστω σε αυτή την επίπονη διαδικασία, έχει γίνει προσπάθεια να δοθούν λύσεις μέσα από τη χρήση των καναλιών επικοινωνίας RSS και μέσα από προσωποποιημένους δικτυακούς τόπους που διαθέτουν τα μεγάλα ειδησεογραφικά πρακτορεία ή ακόμα και από τους μηχανισμούς αναζήτησης που αυτοί διαθέτουν. Σε κάθε περίπτωση όμως, υπάρχουν σημαντικά μειονεκτήματα που συχνά οδηγούν και πάλι το χρήστη σε αδιέξοδο. Τα κανάλια επικοινωνίας δε φιλτράρουν πληροφορίες, τροφοδοτώντας τους RSS readers των χρηστών με πληθώρα πληροφοριών που δεν αφορούν τους χρήστες ή ακόμα είναι ενοχλητικές για αυτούς. Για παράδειγμα η προσθήκη δύο (2) μόνον καναλιών από Ελληνικά μεγάλα ειδησεογραφικά portals μας οδήγησε στη λήψη περισσότερων από 1000 ειδήσεων καθημερινά. Από την άλλη, η χρήση των microsites που έχουν οι δικτυακοί τόποι επιβάλει στους χρήστες την επίσκεψη σε όλους τους δικτυακούς τόπους που τους ενδιαφέρουν. Όσον αφορά στη χρήση των μηχανών αναζήτησης, ακόμα και οι πιο μεγάλες από αυτές συχνά επιστρέφουν εκατομμύρια αποτελέσματα στα ερωτήματα των χρηστών ή πληροφορίες που δεν είναι επικαιροποιημένες. Τέλος, επειδή οι δικτυακοί τόποι των ειδησεογραφικών πρακτορείων δεν έχουν κατασκευαστεί για να προσφέρουν εκτενείς υπηρεσίες αναζήτησης ειδήσεων, είναι συχνό το φαινόμενο είτε να μην προσφέρουν καθόλου υπηρεσία αναζήτησης, είτε η υπηρεσία που προσφέρουν να μη μπορεί να απαντήσει με δομημένα αποτελέσματα και αντί να βοηθά τους χρήστες να εντοπίσουν την πληροφορία που αναζητούν, να τους αποπροσανατολίζει. / We live an era of technology advances and huge technological steps where the Internet becomes a basic place of demonstration of the technology trends. Nevertheless, the way of operation and construction of the WWW is extremely uneven and this results in dead-ends when the users are trying to locate information. Besides the existence of billions of domains leads to difficulties in difficulties in recording all this information. The research that we are doing, is focused on websites that are sources of information and specifically news portals and informational blogs. A simple search on the Internet led to more than 40 large scale press agencies in America. This means that when trying to search for information and more specifically a news article in all its existences somebody has to visit all the websites. This problem, or at least this tedious task is of major concern of the research community. Many solutions were proposed in order to overcome the aforementioned issues with usage of RSS feeds or personalized microsites, or even analytical search applications. In any occasion there are many disadvantages that lead the user to a dead-end again. The RSS feeds do not filter information and they feed the user’s RSS readers with large amounts of information that most of it is not of the user’s concern. For example, a simple addition of 2 rss feeds from large Greek portals led to receipt of more that 1000 news articles within a day! On the other side, the usage of microsites that many websites support is a solution if and only if the user visits every single website and of course have and maintain an account to each one of them. The search engines are an alternative but lately, due to the expansion of the WWW, the results to simple queries are often million or the first results retrieved are outdated. Finally, the websites of the major news agencies are not directly constructed to offer extensive searching facilities and thus they usually offer search results through support of a large well-known search engine (eg. Google). According to the aforementioned the research that we are conducting is furthermore focused on the study of techniques and mechanisms that try to give a solution to the everyday issue of being informed about news and having a spherical opinion about an issue. The idea is simple and lies on the problem of the Internet: instead of letting the user do all the search of the news and information that meet their needs we collect all the informationand present them directly to the user, presenting only the information that meet their profile. This sounds pretty simple and logical, but the implementation we have to think of a number of prerequisites. The constraints are: the users of the Internet speak different languages and they want to see the news in their mother language and the users want access to the information from everywhere. This implies that we need a mechanism that would collect news articles from many – if not all – news agencies worldwide so that everybody can be informed. The news articles that we collect should be furthermore analyzed before presented to the users. In parallel we need to apply text pre-processing techniques, categorization and automatic summarization so that the news articles can be presented back to the user in a personalized manner. Finally, the mechanism is able to construct and maintain a user profile and present only articles that meet the profile of the user and not all the articles collected by the system. As it is obvious this is not a simple procedure. Substantially it a multilevel modular mechanism that implements and uses advanced algorithm on every level in order to achieve the required result. We are referring to eight different mechanisms that lead to the desired result. The systems are: 1. Retrieve news and articles from the Internet –advaRSS system 2. HTML page analysis and useful text extraction – CUTER system. 3. Preprocess and Natural Language Processing in order to extract keywords. 4. Categorization subsystem in order to construct ontologies that assigns texts to categories 5. Article Grouping mechanism (web application level) 6. Automatic Text Summarization 7. Web based User Personalization Mechanism 8. Application based User Personalization Mechanism The subsystems and system architecture is presented in figure 1: The procedure of fetching articles and news from the WWW is a procedure that includes algorithms that fetch data of the large database that is called internet. In this research we have included algorithms for instant retrieval of articles and the mechanism has furthermore mechanism for fetching HTML pages that include news articles. As a next step and provided that we own HTML pages with articles we have procedures for efficient useful text extraction. The HTML pages include the body of the article and information that are disrelated to the article like advertisements. Our mechanism introduces algorithms and systems for extraction of the original body of the text out of the aforementioned pages and omitting any irrelevant information. As a furthermore procedure of the same mechanism we try and extract multimedia related to the article. The aforementioned mechanism are communicating directly with the Internet.

Cette recherche propose une analyse de la construction de l’idée du Français prototypique – la francité – et de ses effets sociaux sur la catégorisation des individus. Elle vise à prendre le Français majoritaire comme objet d’étude au même titre que les groupes minoritaires, et par-là, à mettre en évidence la logique de production de figures d’« altérité intérieure » (ceux qui ne sont pas ou pas tout à fait Français). L’étude s’appuie sur les résultats de deux enquêtes : analyse du discours politique, et enquête ethnographique sur une pratique institutionnelle d'intégration. D'une part, l'analyse révèle la transition du Français majoritaire référentiel qui ne nécessite aucune définition spécifique, au Français ethnicisé et dénommé (le « français de souche »). D'autre part, elle montre que les critères définitoires du Français convergent vers la dimension civique, avec pour effet de différencier des populations dans la société française en fonction de l'adhésion à ou du respect des « valeurs républicaines ». C’est cette conception actuelle du Français qui imprègne fortement le Contrat d’accueil et d’intégration. L’enquête sur ce dispositif nous dévoile les catégorisations de « Nous » et « Eux » autour de la francité mises en œuvre par les agents, et les diverses tactiques de résistance et de gestion des impressions par les signataires en réponse à ces catégorisations. / This thesis analyses the construction of the idea of prototypical French – Frenchness – and its social effects on the categorization of individuals. The aim of this research is to take the French majority as an object of study in the same way as minority groups, and to highlight the logics of the production of figures of « internal otherness » (those who are viewed as not or not totally French). The study relies on the data from two types of investigations: political discourse analysis, and ethnographic research on a specific institutional practice of integration. On the one hand, the analysis shows the transition from a figure of the French majority as a reference requiring no specific definition, to an ethnicized and specified French figure (“French-lineage French”). On the other hand, we observe that the criteria which define Frenchness converge into its civic dimension, with the effect of differentiating the populations in French society according to their commitment to or respect of “republican values”. This current conception of Frenchness forms the basis of the “welcome and integration contract”. Fieldwork on this program reveals the categorizations of “Us” and “Them” around the Frenchness implemented by the officials, and the various tactics of resistance and impression management employed by the signatories as a response to these categorizations.

This study investigates the viability of using Human Intelligence Tasks (HIT) in password categorization. To achieve this, this study constructs and performs a HIT experiment on the online crowdsourcing platform Amazon Mechanical Turks. The study performs the experiment on the site Amazon Mechanical Turks, and gathers data in the form of answers from the workers. A a mixed quantitative and qualitative analysis of the data is performed to investigate on the workers ability to derive the categories of passwords from different categories and difficulties. The study results indicate that HIT workers seem to be unable to reliable categorize more complex passwords correctly, compared to more common and simple passwords. With this result, the study concludes that the quality and reliability of HIT password categorization is lower than would be required to make HIT a valid method for password categorization. The study ends with a discussion on how and why this may be the case and briefly discuss on how the HIT task might be changed in future development to increase its viability.

Sidman's (2000) theory of stimulus equivalence states that all positive elements in a reinforcement contingency enter an equivalence class. The theory also states that if an element from an equivalence class conflicts with a programmed reinforcement contingency, the conflicting element will drop out of the equivalence class. Minster et al. (2006) found evidence suggesting that a conflicting element does not drop out of an equivalence class. In an effort to explain maintained accuracy on programmed reinforcement contingencies, the authors seem to suggest that participants will behave in accordance with a particular partitioning of the equivalence class which continues to include the conflicting element. This hypothesis seems to explain their data well, but their particular procedures are not a good test of the notion of "dropping out" due to the pre-establishment of equivalence classes before the conflicting member entered the class. The current experiment first developed unpartitioned equivalence classes and only later exposed participants to reinforcement contingencies that conflicted with pre-established equivalence classes. The results are consistent with the notion that a partition developed such that the conflicting element had dropped out of certain subclasses of the original equivalence class. The notion of a partitioning of an equivalence class seems to provide a fuller description of the phenomenon Sidman (1994, 2000) described as "dropping out" of an equivalence class.