In this paper, we discuss the computer-aided processing of handwritten tabular records of historical weather data. The observationes meteorologicae, which are housed by the Regensburg University Library, are one of the oldest collections of weather data in Europe. Starting in 1771, meteorological data was consistently documented in a standardized form over almost 60 years by several writers. The tabular structure, as well as the unconstrained textual layout of comments and the use of historical characters, propose various challenges in layout and text recognition. We present a customized strategy to digitize tabular and handwritten data by combining various state-of-the-art methods for OCR processing to fit the collection. Since the recognition of historical documents still poses major challenges, we provide lessons learned from experimental testing during the first project stages. Our results show that deep learning methods can be used for text recognition and layout detection. However, they are less efficient for the recognition of tabular structures. Furthermore, a tailored approach had to be developed for the historical meteorological characters during the manual creation of ground truth data. The customized system achieved an accuracy rate of 82% for the text recognition of the heterogeneous handwriting and 87% accuracy for layout recognition of the tables.

Ce mémoire présente une nouvelle approche pour la modélisation des processus d'affaires. L'approche est basée sur la technique de la cartographie des processus d'affaires et UMM (UN/CEFACTModeling Methodology). La cartographie des processus d'affaires est utilisée pour modéliser les processus d'affaires à un haut niveau d'abstraction montrant des éléments tels que leurs entrées, leurs sorties, et leurs interactions. Les détails de ces processus d'affaires sont obtenus à un bas niveau d'abstraction et sont modélisés en utilisant UMM. UMM décrit les interactions et les échanges d'informations entre les partenaires d'affaires. Le modèle obtenu fournit une diversité d'informations à différents niveaux d'abstraction qui répondent à différents besoins de conception de processus, tels que les besoins d'affaires et les besoins fonctionnels. L'approche est appliquée à la modélisation d'un processus d'affaires complexe au gouvernement du Québec : la Gestion Intégrée des Documents (GID). Les résultats montrent que l'approche proposée identifie et documente les processus d'affaires existants tout en offrant une vue globale sur la structure et la dynamique du processus d'affaires.

This study is about teachers in socially oriented subjects and their attitudes and experiences regarding working with controversial issues in teaching. Previous research has shown that there are shortcomings and major challenges in handling controversial subjects in the classroom. Therefore, I chose to investigate this by interviewing eight subject teachers in socially oriented subjects. The purpose of the study was to examine different teachers' work with controversial issues and their experiences of opportunities and challenges respectively in teaching controversial subjects. After conducting a qualitative interview study with eight teachers, the results were consistent with previous research. The results of the study indicate a lack of time and challenges associated with working on controversial issues, as they are difficult to handle and can lead to conflicts and disruptions in the classroom. Consequently, some teachers choose not to teach about controversial issues. However, on the other hand, the results of the study show that there were teachers who encouraged working with controversial issues as they considered it important to socialize students into society by addressing controversial topics. They believed that by doing so, students develop into democratic citizens with political tolerance. Another interesting finding from the interview study is that certain schools as a whole choose not to work with controversial issues to avoid negative consequences. This is interesting because other schools in the study, for example, those with UN profiles, focus on challenging students with controversial issues to develop them into democratic citizens of society. One can therefore conclude that working with controversial issues is a task that must involve the entire school, not just the teacher, as it requires efforts from management, colleagues, teachers, and students. Thus, in schools where only the teacher is responsible for addressing controversial issues without support from management or colleagues, it may lead to a reduction in teaching on controversial subjects.

La conception des documents publics joue un rôle important dans la manière dont l'utilisateur perçoit et comprend les informations importantes qu'ils véhiculent. De plus, la conception de tels documents peut avoir un impact important sur l’utilisateur en ce qui concernesa confiance et son processus décisionnel. Les principes de design d’informationpeuvent apporter de précieuses perspectives sur comment les documents publics devraient être conçus pour qu’ils réalisent leurs objectifs. Cependant, le domaine du design d’informationdes documents publics souffre d’un manque important dans l’application de ces principes. D'autre part, les principes perceptuels delaGestalt peuvent expliquer comment les humains organisent des éléments visuels en groupes et comment ils reconnaissent des patrons. Depuis leur introduction au début du XXesiècle, les principes delaGestalt ont servi à améliorer le processus visuo-perceptuel dans différents champs d'application (telles que la peinture, la sculpture, la création graphique et le design d’information). Dans ce mémoire, nous étudionsl’application des principes de la Gestalt pour améliorer la compréhension de l’utilisateur des documents publics. Pour ce faire, en premier lieu, une méthode d’évaluation heuristique est adaptée pour identifier les problèmes d'utilisabilité dans les documents publics. Deuxièmement, une méthode d’analyse de documents, reposant sur des mesures des éléments visuels dans les documents, est proposée pour étudier l’application des principes de Gestalt, ainsi que les principes de conception graphique, dans les documents publics. Un échantillon de deux documents publics est utilisé pour démontrer et tester la méthodologie proposée. Sur la base des failles de conception identifiées par les deux méthodes susmentionnées, une approche minimale est proposée pour appliquer les principes de la Gestalt pour résoudre les problèmes identifiés. Les résultats obtenus montrent que l’application des principes de la Gestalt peut améliorer l'accessibilité et la compréhensibilité de l'information présentéedans les documents publics. / The design of public documents playsan important role in user perception and understanding of the important information they contain. Moreover, the design of such documents can have an important impact on user self-confidence and decision making. The information design principles pertainingto how to visually structure and verbally construct documents to be easily well perceived and understood can provide valuable insights on how public documents should be designed and presented. Yet, in designing public documents there is a lack in applyingthe information design methods and frameworks. On the other hand, the Gestalt principles of perception can explain how humans organize visual elements into groups and how humans perceive and recognize patterns. Since their introduction in the early twentieth century, the Gestalt principles were examined and studied various times to improve the visual perceptual process in many application areas such as: painting, sculpture, graphic design and information design. Accordingly, this work studies the application of the Gestalt principles to improve the user understanding of public documents. To this end, first an adapted heuristic evaluation method is proposed to identify the usability problems of public documents. Secondly, based on precise measurements of thevisual elements within the documents, a document analytics method is proposed to examine the application of the Gestalt principles as well as the design principles in public documents. A sample comprising two public documents is used to test and demonstrate the proposed methodology. Based on the design flaws detected by the aforementioned two methods, a minimal approach is adopted to apply the Gestalt principles to solve the identified problems. The results obtained show that the integration of the Gestaltprinciples in the design process can improve the accessibility of information in public documents and resolve many of the usability problems in these documents.

In the era of total digitization of documents, navigating vast and heterogeneous data landscapes presents significant challenges for effective information retrieval, both for humans and digital agents. Traditional methods of knowledge organization often struggle to keep pace with evolving user demands, resulting in suboptimal outcomes such as information overload and disorganized data. This thesis presents a case study on a pipeline that leverages principles from cognitive science, graph theory, and semantic computing to generate semantically organized knowledge graphs. By evaluating a combination of different models, methodologies, and algorithms, the pipeline aims to enhance the organization and retrieval of digital documents. The proposed approach focuses on representing documents as vector embeddings, clustering similar documents, and constructing a connected and scalable knowledge graph. This graph not only captures semantic relationships between documents but also ensures efficient traversal and exploration. The practical application of the system is demonstrated in the context of digital libraries and academic research, showcasing its potential to improve information management and discovery. The effectiveness of the pipeline is validated through extensive experiments using contemporary open-source tools.

A holistic study of abandoned house sites on the island of Unst was conducted to address the extent to which perceptions of historic settlement on Shetland are supportable. These perceptions cast long lived nucleated settlement as the normative traditional form of historic settlement, and dispersed settlements as short-lived exceptions to this norm. Historic settlement, in these perceptions are argued to be static, which is not borne out in archaeological evidence. Issues associated with historic Shetland settlement models were identified to parallel traditional views of Scottish highland rural settlement, which cast the highland society as ahistoric and unchanging. Historical, archaeological and geographic evidence for settlement on Unst were used to assess the geographical distribution of historic settlement on the island. Two detailed case studies integrated archaeological prospection techniques with the historical, archaeological and landscape contexts to form new narratives for the field remains around two abandoned house sites. Assessment of the historical settlement of Unst highlighted a much greater degree of variation between the different evidence strands for the perceptions to truly represent the island’s historical settlement. Similarly, findings from the case studies highlighted a much greater degree of alterations to the field systems and enclosures associated with the settlements than would be anticipated. Alternative narratives with several phases were hypothesised for field remains of each case study. / Hunter Archaeological and Historical Trust; Society of Antiquaries of Scotland; School of Archaeological and Forensic Sciences

En cette ère numérique, l’utilisation des documents papier s’avère peu pratique et inefficace, ce qui motive les sociétés à évoluer vers l’utilisation des documents électroniques (ou e-docs). Ce désir d’innover vers une opération sans papier peut améliorer l’efficacité et la qualité des services d’administrations publiques ou privées de manière à accélérer leurs activités et en même temps mieux satisfaire les besoins des clients. Cependant, cette pratique a créé des nouveaux besoins, tels que la signature numérique réelle de documents. Dans ce mémoire, nous avons proposé un nouveau schéma de signature numérique utilisant FIDO2, qui se trouve être une nouvelle norme d’authentification sécurisée en ligne basée sur la signature numérique. Le fait que FIDO2 soit un standard libre permet aux développeurs de logiciel et de matériel d’implémenter plus facilement leurs propres produits. Cela nous a inspiré à l’utiliser pour une fin de signature numérique, l’idée étant de remplacer le défi envoyé par le serveur avec le hash de e-docs et de l’envoyer à l’appareil du signataire afin que ce dernier signe avec sa clé privée. Comme dans le cas de l’infrastructure à clé publique, chaque utilisateur possédait une paire de clés, c’est-à-dire une clé privée et une clé publique. Un signataire doit confirmer son identification biométrique (empreinte digitale, reconnaissance faciale, voix, etc.) ou son code PIN pour accéder à la clé privée stockée localement sur son appareil et signer un document. Au cours de notre recherche, nous avons effectué plusieurs tests avec différents équipements (PC, USB FIDO, Smartphone) ainsi que différentes OS (Android, iOS, Windows). Les résultats de nos tests nous montrent que nous pouvons utiliser FIDO2 pour signer un document électronique. Cette nouvelle approche proposée peut être utilisée pour une signature face à face (en locale) ou à distance (en ligne). Le prototype développé pour la mise en œuvre de notre approche a été validé auprès d’usagers types (membres clients et conseillers) dans une entreprise. / In this digital era, the use of paper documents is impractical and inefficient, which motivates companies to move towards the use of electronic documents (or e-docs). This desire to innovate towards a paperless operation can improve the efficiency and quality of public or private administration services so as to speed up their activities and at the same time better meet customer needs. However, this practice has created new needs, such as the actual digital signature of documents. In this thesis, we have proposed a new digital signature scheme using FIDO2, which happens to be a new standard for secure online authentication based on digital signatures. The fact that FIDO2 is a free standard makes it easier for software and hardware developers to implement their own products. This inspired us to use it a digital signature purpose, the idea being, to replace the challenge sent by the server with the hash of e-docs and send it to the signer’s device so that the latter signs with his private key. As with public key infrastructure, each user had a key pair, that is, a private key and a public key. A signatory must confirm their biometric identification (fingerprint, facial recognition, voice, etc.) or PIN code to access the private key stored locally on their device and sign a document. During our research, we carried out several tests with different equipment (PC, USB FIDO, Smartphone) as well as different OS (Android, iOS, Windows). The results of our tests show us that we can use FIDO2 to sign an electronic document. This proposed new approach can be used for a face-to-face (local) or remote (online) signature. The prototype developed for the implementation of our approach has been validated with typical users (member-clients and advisers) in a company.

Thèse en cotutelle : Université Laval, Québec, Canada et Université Paris-Saclay, Cachan, France. / Le projet de recherche s’inscrit dans le contexte du mouvement d’open legal data, c’est-à-dire des données judiciaires ouvertes. En effet, que ce soit en France ou au Canada, les données judiciaires font l’objet d’une législation en faveur de l’open data. Les données judiciaires sont mises à la disposition des citoyens, de manière variable selon les systèmes juridiques. La question qui est au cœur du problème est l’affrontement de deux valeurs fondamentales : le droit du public à la transparence de l’administration de la justice, qui justifie que les données judiciaires soient consultables, et le droit de l’individu à la protection de sa vie privée. / This research project is set within the broader context of the open data movement, namely that of open judicial data. This type of data has been subject to legislation in favour of open data both in France and in Canada. Each legal system has a different approach as to how judicial data is made available to the population. The underlying issue is the interplay between two fundamental rights: the collective right to an open and transparent justice system, which in turn justifies the openness of judicial data, and the individual right to privacy.

Titre de l'écran-titre (visionné le 24 mai 2023) / Comme dans de nombreuses autres démocraties, il existe au Canada un droit d'accès à l'information judiciaire. Il s'agit d'un élément fondamental de tout processus judiciaire. Ce droit a deux objectifs principaux : offrir une fenêtre sur le système de justice et permettre aux gens d'acquérir une meilleure compréhension du processus judiciaire. Parmi les documents essentiels au système de justice figure le plumitif : un document qui détaille le déroulement de chaque dossier ouverts devant les tribunaux. Malgré tout, il a été démontré que le plumitif est un document difficile à comprendre, tant pour les citoyens que les praticiens. Dans cette thèse, nous concentrons nos efforts sur le plumitif criminel, et nous proposons d'améliorer l'accès à ce registre juridique à l'aide de techniques du traitement automatique de la langue naturelle. Premièrement, nous proposons un nouveau jeu de données pour la génération des descriptions de plumitifs. Ce jeu de données est utilisé pour entraîner des générateurs de texte neuronaux afin de fournir des descriptions intelligibles des plumitifs criminels. Nous proposons ensuite une nouvelle métrique robuste d'évaluation de génération textuelle qui quantifie les omissions et les hallucinations des générateurs textuels neuronaux, un problème de grande importance dans le domaine juridique. Nous avons ensuite mené une évaluation manuelle des générations faites par différents modèles de réseaux de neurones, pour mieux caractériser le comportement de ceux-ci. Finalement, nous proposons un nouvel algorithme de décodage pour les générateurs textuels neuronaux de types "data-to-text" qui améliore la fidélité du texte généré par rapport aux données d'entrée. / As in many other democracies, Canada has a right of access to court information. It is a fundamental element of any judicial process. This right has two main purposes: to provide a window on the justice system and to allow people to gain a better understanding of the court process. One of the essential documents in the justice system is the docket; a document that details the progress of each case before the courts. Despite this, it has been shown that the docket is a document difficult to understand for both citizens and practitioners. In this thesis, we focus our efforts on the criminal docket, and we propose to improve access to this legal record using automatic natural language processing techniques. To this end, we propose a new dataset for generating docket descriptions. This dataset is used to train neural text generators to provide intelligible descriptions of criminal dockets. We then propose a new robust text generation evaluation metric that quantifies omissions and hallucinations of neural text generators, a problem of great importance in the legal domain. We then conduct a manual evaluation of generations made by neural networks, to better characterize their behavior. Finally, we propose a new decoding algorithm for data-to-text neural generators that improves the faithfulness of generated text with respect to the input.

The legal nature of the bill of lading as a negotiable document of title has allowed it to provide the basis of a system in which bankers provide credit for the financing of international sales contracts on the strength of the security afforded by the goods represented in the bill. The sea waybill has appeared as a substitute for the bill of lading and, despite its nature as a nonnegotiable document, it can be employed in a manner which allows it to provide collateral security to banks. Multimodal transport documents which may be issued in negotiable or non-negotiable form assume the same legal role as the bill oflading or sea waybill respectively. The inclusion of specific articles in the 1993 Revision of the UCP relating to non-negotiable sea waybills and multimodal transport documents affirms their acceptability to banks financing international sales contracts under documentary letters of credit. / Law / LL.M.