abstract: Designing a hazard intelligence platform enables public agencies to organize diversity and manage complexity in collaborative partnerships. To maintain the integrity of the platform while preserving the prosocial ethos, understanding the dynamics of “non-regulatory supplements” to central governance is crucial. In conceptualization, social responsiveness is shaped by communicative actions, in which coordination is attained through negotiated agreements by way of the evaluation of validity claims. The dynamic processes involve information processing and knowledge sharing. The access and the use of collaborative intelligence can be examined by notions of traceability and intelligence cohort. Empirical evidence indicates that social traceability is statistical significant and positively associated with the improvement of collaborative performance. Moreover, social traceability positively contributes to the efficacy of technical traceability, but not vice versa. Furthermore, technical traceability significantly contributes to both moderate and high performance improvement; while social traceability is only significant for moderate performance improvement. Therefore, the social effect is limited and contingent. The results further suggest strategic considerations. Social significance: social traceability is the fundamental consideration to high cohort performance. Cocktail therapy: high cohort performance involves an integrative strategy with high social traceability and high technical traceability. Servant leadership: public agencies should exercise limited authority and perform a supporting role in the provision of appropriate technical traceability, while actively promoting social traceability in the system. / Dissertation/Thesis / Doctoral Dissertation Business Administration 2015

Computational linguistics is a sub-field of artificial intelligence; it is an interdisciplinary field dealing with statistical and/or rule-based modeling of natural language from a computational perspective. Traditionally, fuzzy logic is used to deal with fuzziness among single linguistic terms in documents. However, linguistic terms may be related to other types of uncertainty. For instance, different users search ‘cheap hotel’ in a search engine, they may need distinct pieces of relevant hidden information such as shopping, transportation, weather, etc. Therefore, this research work focuses on studying granular words and developing new algorithms to process them to deal with uncertainty globally. To precisely describe the granular words, a new structure called Granular Information Hyper Tree (GIHT) is constructed. Furthermore, several technologies are developed to cooperate with computing with granular words in spam filtering and query recommendation. Based on simulation results, the GIHT-Bayesian algorithm can get more accurate spam filtering rate than conventional method Naive Bayesian and SVM; computing with granular word also generates better recommendation results based on users’ assessment when applied it to search engine.

Cette thèse CIFRE menée au sein de l'équipe « Veille Innovation » d'EDF R&D décrit le processus de constitution de l'activité de veille collaborative sur le portail collaboratif Hermès. Nous analysons la tension entre, d'une part, un objectif de déploiement basé sur un usage volontaire et émergent valorisant le mode communautaire et, d'autre part, le contexte de l'entreprise marqué par une organisation hiérarchique et en mode projet. Suivant une approche de recherche pragmatique, nous mobilisons un cadre théorique et analytique double : la « Communication Constitutive de l'Organisation » (CCO) et la « Sémiotique des Transactions Coopératives » (STC).Dans la première partie, nous présentons les caractéristiques de l'activité de veille réalisée au sein des entreprises et des outils 2.0 dédiés. Nous décrivons en détail le portail collaboratif de veille Hermès et le travail de son équipe-projet. L'analyse des traces d'usage du portail nous permet de mieux appréhender l'appropriation d'Hermès par ses utilisateurs ainsi que les deux modes de coopération en présence : la coopération indirecte et la coopération directe.Dans la deuxième partie, nous présentons les cadres théoriques qui nous permettent d'introduire la notion de communicologue et de proposer une démarche pour l'engagement des acteurs dans la veille collaborative. En utilisant une démarche de théorisation ancrée, nous analysons les entretiens individuels réalisés avec des veilleurs appartenant à quatre « communautés » Hermès. Nous reconstituons quatre cas de coopération à partir de leur compréhension de l'activité de veille collaborative et de leurs usages du portail.La persistance du mode d'organisation par projets au sein des « communautés » de veille Hermès indique que les changements envisagés par l'équipe-projet de conception ne sont pas produits. Pourtant, nous observons que les « communautés » de veille peuvent construire leur propre mode d'organisation de l'activité communautaire tout en restant en phase avec les besoins et les lignes directrices des projets. Nous faisons l'hypothèse qu'en mettant en place la démarche du communicologue, l'équipe-projet peut aider les communautés Hermès à concevoir leur activité collective selon les caractéristiques d'un régime de coopération communautaire. Enfin, cette recherche nous permet de distinguer les activités de veille collaborative et de veille communautaire au sens propre. / This CIFRE thesis, conducted within the team "Veille Innovation" at the R&D direction of EDF, describes the constitution process of the collaborative intelligence activity through the collaborative portal Hermes. We analyze the tension between, on one hand, a deployment strategy based on a voluntary and emerging use that values community based organization and, secondly, the business context marked by a hierarchical and project based organization. Following a pragmatic research approach, we mobilize a double theoretical and analytical framework: the "Communicative Constitution of Organizations" (CCO) and "Semiotics of Cooperative Transactions" (STC).In the first part, we present the characteristics of the competitive intelligence activity carried out within companies using dedicated 2.0 tools. We describe in detail the collaborative portal Hermes and the work of its project team. The analysis of the portal's usage quantitative data allows us to better understand the appropriation of Hermes for its users and the two modes of cooperation involved: indirect cooperation and direct cooperation.In the second part, we present the theoretical frameworks that allow us to introduce the concept of “communicologue” and to propose an approach for the engagement of stakeholders in collaborative intelligence activity. Using a grounded theory approach, we analyze individual interviews with users belonging to four Hermès "communities". We reconstruct four cases of cooperation from their understanding of the collaborative intelligence activity and their use of the portal.The persistence of the project mode organization in the Hermès "communities" indicates that the changes envisioned by the project team are not produced. However, we observe that the "communities" can build their own community based organization while remaining in line with the needs and guidelines of the projects. We assume that by the implementation of the “communicologue” approach, the project team can help Hermes communities design their collective activity following the characteristics of a community cooperative regime. Finally, this research allows us to distinguish the activities of collaborative intelligence and community watch literally.

The onset of Industry 4.0 has provided considerable benefits to Intelligent Cyber-Physical Systems (ICPS), with technologies such as internet of things, wireless sensing, cognitive computing and artificial intelligence to improve automation and control. However, with increasing automation, the “human” element in industrial systems is often overlooked for the sake of standardization. While automation aims to redirect the workload of human to standardized and programmable entities, humans possess qualities such as cognitive awareness, perception and intuition which cannot be automated (or programmatically replicated) but can provide automated systems with much needed robustness and sustainability, especially in unstructured and dynamic environments. Incorporating tangible human skills and knowledge within industrial environments is an essential function of “Human-in-the-loop” (HITL) Systems, a term for systems powerfully augmented by different qualities of human agents. The primary challenge, however, lies in the realistic modelling and application of these qualities; an accurate human model must be developed, integrated and tested within different cyber-physical workflows to 1) validate the assumed advantages, investments and 2) ensure optimized collaboration between entities. Agricultural Robotic Systems (ARS) are an example of such cyber-physical systems (CPS) which, in order to reduce reliance on traditional human-intensive approaches, leverage sensor networks, autonomous robotics and vision systems and for the early detection of diseases in greenhouse plants. Complete elimination of humans from such environments can prove sub-optimal given that greenhouses present a host of dynamic conditions and interactions which cannot be explicitly defined or managed automatically. Supported by efficient algorithms for sampling, routing and search, HITL augmentation into ARS can provide improved detection capabilities, system performance and stability, while also reducing the workload of humans as compared to traditional methods. This research thus studies the modelling and integration of humans into the loop of ARS, using simulation techniques and employing intelligent protocols for optimized interactions. Human qualities are modelled in human “classes” within an event-based, discrete time simulation developed in Python. A logic controller based on collaborative intelligence (HUB-CI) efficiently dictates workflow logic, owing to the multi-agent and multi-algorithm nature of the system. Two integration hierarchies are simulated to study different types of integrations of HITL: Sequential, and Shared Integration. System performance metrics such as costs, number of tasks and classification accuracy are measured and compared for different collaboration protocols within each hierarchy, to verify the impact of chosen sampling and search algorithms. The experiments performed show the statistically significant advantages of HUB-CI based protocol over traditional protocols in terms of collaborative task performance and disease detectability, thus justifying added investment due to the inclusion of HITL. The results also discuss the competitive factors between both integrations, laying out the relative advantages and disadvantages and the scope for further research. Improving human modelling and expanding the range of human activities within the loop can help to improve the practicality and accuracy of the simulation in replicating an HITL-ARS. Finally, the research also discusses the development of a user-interface software based on ARS methodologies to test the system in the real-world.<br>

As the threat of supply-chain attacks grows, the need for techniques to protect software integrity likewise increases. The concept of reproducible builds is one such protection. By ensuring that a package can be rebuilt in the exact same way every time, reproducible builds allow users to notice when a package has changed even though its source code stays the same. Thus, the knowledge of which packages are reproducible and therefore easier to trust is a crucial part of this protection mechanism. Current strategies for validating and distributing this information rely on the work of a small number of individual entities with limited coordination in-between them, leading to user confusion because of the lack of a central authority. This work describes a protocol for decentralized coordination and validation of package reproducibility based on hidden votes to limit collusion and a reward scheme to ensure collaboration. The protocol uses the Hyperledger Fabric blockchain as supporting infrastructure, gaining the benefits of high availability, integrity of results and decision traceability from its decentralized nature. To test the protocol, a formal specification was written in TLA+ and validated through model checking. The results showed that, at least for the tested networks, the protocol produces valid results and enforces collaboration between users. Next steps for the project would be to build a functional prototype of the system to test its performance characteristics as well as studying the system actor assumptions made in the protocol design. / Likt hotet från leveranskedjeattacker har ökat, ökar även behoven av skyddstekniker för att säkerställa riktigheten hos mjukvara. Ett sådant typ av skydd ges av reproducerbara byggen. Om ett mjukvarupaket kan byggas exakt likadant varje gång så möjliggör det för användare att upptäcka om paketet har förändrats trots att dess källkod inte har gjort det. Att kunna veta vilka paket som är reproducerbara och därmed lättar att lita på är således en central del i denna skyddsmekanism. Nuvarande strategier för validering och distribution av sådan information bygger på arbete från ett fåtal individer och organisationer med begränsad koordinering däremellan. Detta leder till förvirring för användare på grund av bristen av en central tillitspunkt eller auktoritet. Detta arbete beskriver ett protokoll för decentralizerad koordinering och validering av paketreproducerbarhet baserat på hemliga röster för att begränsa otillåtet samarbete och ett belöningssystem för att motivera önskat samarbete. Protokollet använder blockkedjan Hyperledger Fabric som grund, med fördelarna av att få hög tillgänglighet, resultatsriktighet och spårbara beslut. En formel specifikation skrevs i TLA+ för att testa protokollet och validerades med modeltestning. Testresultatet för de testade nätverkskonfigurationerna visade att protokollet genererar valida resultat och garanterar samarbete mellan användare. De nästa stegen i projektet skulle vara att bygga en funktionell prototyp av systemet för att testa dess prestanda såväl som att studera de antaganden protokollet är designat runt.

Emerging advances in the realm of Artificial Intelligence (AI) have had a tremendous impact on various fields around us and society as a whole. As technologies continue to evolve, so will the role of designers when it comes to using AI. It has the potential to be the next tool designers use to create human-centered products and services. To truly understand AI and harness its capabilities, it is crucial to demystify the term and its inner workings. This thesis is explorative research to shed light on collaborative intelligence and how designers can harness the capabilities of AI. It further explores how to integrate the principles of design and AI to create AI-driven products and services. In addition to background research conducted on both design and AI, the importance of both these fields’ intersection was also researched upon. The project followed the Double Diamond design process principles, consisting of four phases: discover, define, develop, and deliver. This process was then used again to design a framework that bridges the gap between AI &amp; design principles. This research aimed to explore how designers could use AI to develop new products and services. The project resulted in a framework that guides designers on how to get acclimated to AI and uses a specific set of principles to design for AI. It contains concepts necessary to understand the different aspects of AI and aims to build a common language amongst all AI practitioners. The framework also serves as a basic outline of a workshop that provides various design methods that AI practitioners can use to ideate AI-driven solutions. / Framsteg inom Artificiell Intelligens (AI) har redan haft enorma effekter på diverse ämnesområden med direkt eller indirekt påverkan på oss människor. I och med att teknologin som utnyttjar AI kommer att utvecklas, kommer även rollen för designers att förändras. AI innehar potentialen att bli nästa verktyg som kan brukas för att skapa människocentrerade produkter och tjänster. För att förstå och nyttja AI och dess förmåga är det kritiskt att avmystifiera termen och dess potential. Detta explorativa arbete syftar till att nysta upp kollaborativ intelligens samt att undersöka hur det kan användas av designers för att nyttja AIs fulla potential. Därtill utforskar arbetet hur AI och designprinciper kan integreras för att skapa AI-baserade produkter och tjänster. Utöver forskningen inom design och AI undersöks även fältens skärningspunkter. Följande arbete använder sig av Double Diamond designprocessen och dess principer: discover, define, develop, and deliver. Denna process kommer att användas för att konstruera ett ramverk som binder samman AI och designprinciper. Arbetet syftar till att utforska hur designers kan använda AI för att skapa nya produkter och tjänster. Resultatet är ett ramverk som kan vägleda designers att acklimatisera sig med AI och dess specifika principer för att kunna applicera AI i sitt arbete. Ramverket innefattar nödvändiga koncept för att förstå olika aspekter av AI och strävar efter att bygga ett gemensamt språk för alla utövare av AI. Ramverket ger dessutom riktlinjer för att strukturera workshops som förser alla möjliga AI användare med designmetoder för att skapa AI baserade lösningar.

De nos jours, il y a une demande croissante pour fournir de l'information temps réel à partir de l'environnement, e.g. état infectieux de maladies, force du signal, conditions de circulation, qualité de l'air. La prolifération des dispositifs de capteurs et la mobilité des personnes font de la Mobile Collaborative Sensing (MCS) un moyen efficace de détecter et collecter l'information à un faible coût. Dans MCS, au lieu de déployer des capteurs statiques dans une zone, les personnes disposant d'appareils mobiles jouent le rôle de capteurs mobiles. En général, une application MCS exige que l'appareil de chacun ait la capacité d'effectuer la détection et retourne les résultats à un serveur central, mais également de collaborer avec d'autres dispositifs. Pour que les résultats puissent représenter l'information physique d'une région cible et convenir, quel type de données peut être utilisé et quel type d'information doit être inclus dans les données collectées? Les données spatio-temporelles peuvent être utilisées par des applications pour bien représenter la région cible. Dans des applications différentes, l'information de localisation et de temps sont 2 types d'information communes, et en les utilisant la région cible d'une application est sous surveillance complète du temps et de l'espace. Différentes applications nécessitent de l'information différente pour atteindre des objectifs différents. E.g. dans cette thèse: i- MCS-Locating application: l'information de résistance du signal doit être incluse dans les données détectées par des dispositifs mobiles à partir d'émetteurs de signaux ; ii- MCS-Prédiction application : la relation entre les cas d'infection et les cas infectés doit être incluse dans les données par les dispositifs mobiles provenant des zones de flambée de la maladie ; iii- MCS-Routing application : l'information routière en temps réel provenant de différentes routes de circulation doit être incluse dans les données détectées par des dispositifs embarqués. Avec la détection de l'information physique d'une région cible, et la mise en interaction des dispositifs, 3 thèmes d'optimisation basés sur la détection sont étudiés et 4 travaux de recherche menés: -Mobile Collaboratif Détection Cadre : un cadre mobile de détection collaborative est conçu pour faciliter la coopérativité de la collecte, du partage et de l'analyse des données. Les données sont collectées à partir de sources et de points temporels différents. Pour le déploiement du cadre dans les applications, les défis clés pertinents et les problèmes ouverts sont discutés. -MCS-Locating : l'algorithme LiCS (Locating in Collaborative Sensing based Data Space) est proposé pour atteindre la localisation de la cible. LiCS utilise la puissance du signal reçu dans tous les périphériques sans fil comme empreintes digitales de localisation pour les différents emplacements. De sorte LiCS peut être directement pris en charge par l'infrastructure sans fil standard. Il utilise des données de trace d'appareils mobiles d'individus, et un modèle d'estimation d'emplacement. Il forme le modèle d'estimation de localisation en utilisant les données de trace pour atteindre la localisation de la cible collaborative. Cette collaboration entre périphériques est au niveau des données et est supportée par un modèle. -MCS-Prédiction: un modèle de reconnaissance est conçu pour acquérir dynamiquement la connaissance de structure de la RCN pertinente pendant la propagation de la maladie. Sur ce modèle, un algorithme de prédiction est proposé pour prédire le paramètre R. i.e. le nombre de reproduction qui est utilisé pour quantifier la dynamique de la maladie pendant sa propagation. -MCS-Routing : un algorithme de navigation écologique ‘eRouting’ est conçu en combinant l'information de trafic temps réel et un modèle d'énergie/émission basé sur des facteurs représentatifs. Sur la base de l'infrastructure standard d'un système de trafic intelligent, l'information sur le trafic est collectée / Nowadays, there is an increasing demand to provide real-time information from the environment, e.g., the infection status of infectious diseases, signal strength, traffic conditions, and air quality, to citizens in urban areas for various purposes. The proliferation of sensor-equipped devices and the mobility of people are making Mobile Collaborative Sensing (MCS) an effective way to sense and collect information at a low deployment cost. In MCS, instead of just deploying static sensors in an interested area, people with mobile devices play the role of mobile sensors to sense the information of their surroundings, and the communication network (3G, WiFi, etc.) is used to transfer data for MCS applications. Typically, a MCS application not only requires each participant's mobile device to possess the capability of performing sensing and returning sensed results to a central server, but also requires to collaborate with other mobile and static devices. In order to make sensed results well represent the physical information of a target region, and well be suitable to a certain application, what kind of data can be used for different applications, and what kind of information needs to be included into the collected sensing data? Spatio-temporal data can be used by different applications to well represent the target region. In different applications, location and time information is two kinds of common information, and by using such information, the target region of an application is under comprehensive monitoring from the view of time and space. Different applications require different information to achieve different sensing purposes. E.g. in this thesis: i- MCS-Locating application: signal strength information needs to be included into the sensed data by mobile devices from signal transmitters; ii- MCS-Prediction application: the relationship between infecting and infected cases needs to be included into the sensed data by mobile devices from disease outbreak areas; iii- MCS-Routing application: real-time traffic and road information from different traffic roads, e.g., traffic velocity and road gradient, needs to be included into the sensed data by road-embedded and vehicle-mounted devices. With sensing the physical information of a target region, and making mobile and static devices collaborate with each other in mind, in this thesis three sensing based optimization applications are studied, and following four research works are conducted: - a MCS Framework is designed to facilitate the cooperativity of data collection, sharing, and analysis among different devices. Data is collected from different sources and time points. For deploying the framework into applications, relevant key challenges and open issues are discussed. - MCS-Locating: an algorithm LiCS (Locating in Collaborative Sensing based Data Space) is proposed to achieve target locating. It uses Received Signal Strength that exists in any wireless devices as location fingerprints to differentiate different locations, so it can be directly supported by off-the-shelf wireless infrastructure. LiCS uses trace data from individuals' mobile devices, and a location estimation model. It trains the location estimation model by using the trace data to achieve collaborative target locating. Such collaboration between different devices is data-level, and model-supported. - MCS-Prediction: a recognition model is designed to dynamically acquire the structure knowledge of the relevant RCN during disease spread. On the basis of this model, a prediction algorithm is proposed to predict the parameter R. R is the reproductive number which is used to quantify the disease dynamics during disease spread. - MCS-Routing: an eco-friendly navigation algorithm, eRouting, is designed by combining real-time traffic information and a representative factor based energy/emission model. Based on the off-the-shelf infrastructure of an intelligent traffic system, the traffic information is collected