The purpose of this thesis is to examine the potential impact of High Voltage Direct Current (HVDC) innovations on the power industry. A short historical review is provided on the so-called ‘War of the Currents’, which has placed alternating current (AC) as the dominant platform for power transmission. The revival of DC in high voltage transmission is here examined as a reverse salient, as various indicators show that the role of HVDC in the power industry seems to be constantly growing. In this thesis the potential of HVDC to drive industrial change is put to question, and an effort is made to define that change; questions are being addressed such as ‘how likely is HVDC to gain a more important role in power transmission, and what may that role be?’ and ‘how may the power market look like after a wider implementation of HVDC in power systems?’. The competition between ABB and Siemens (and at a lower level Alstom), market leaders in HVDC technologies, is analyzed with specific regard to the pursuit of inventing a HVDC circuit breaker; a technological leap that has been considered necessary for overcoming a number of obstacles to creating a DC grid. As of November 2012, ABB has developed the world’s first HVDC circuit breaker, acknowledged as a technological breakthrough. An attempt is being made to evaluate the impact of this innovation in terms of enhancing ABB’s entrepreneurial activity and granting the company with a competitive edge. Furthermore, the vision of developing DC grids – and thus R&D investments for the invention of a HVDC circuit breaker – is approached as a case of climate change being a main driving force for innovation, since this novel invention promises to make easier the integration of more renewable energy sources in power systems. This co-evolution of environment policy and innovation strategy is examined under the scope of Erik Dahmén’s theory of development blocks. Finally, ABB’s recent innovation is considered a disturbance in the system capable of bringing implications on the market that are here interpreted as a case of creative destruction, based upon Schumpeter’s terminology. In the conclusion section possible threats for Siemens and Alstom are also realized, and the need for them is questioned to go through changes in order to remain competitive, a situation that is here regarded as a case of transformation pressure.

No description available.

Yes / In wireless visual sensor networks (WVSNs), streaming all imaging data is impractical due to resource constraints. Moreover, the sheer volume of surveillance videos inhibits the ability of analysts to extract actionable intelligence. In this work, an energy-efficient image prioritization framework is presented to cope with the fragility of traditional WVSNs. The proposed framework selects semantically relevant information before it is transmitted to a sink node. This is based on salient motion detection, which works on the principle of human cognitive processes. Each camera node estimates the background by a bootstrapping procedure, thus increasing the efficiency of salient motion detection. Based on the salient motion, each sensor node is classified as being high or low priority. This classification is dynamic, such that camera nodes toggle between high-priority and low-priority status depending on the coverage of the region of interest. High-priority camera nodes are allowed to access reliable radio channels to ensure the timely and reliable transmission of data. We compare the performance of this framework with other state-of-the-art methods for both single and multi-camera monitoring. The results demonstrate the usefulness of the proposed method in terms of salient event coverage and reduced computational and transmission costs, as well as in helping analysts find semantically relevant visual information. / Supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2013R1A1A2012904).

Fashion understanding is a hot topic in computer vision, with many applications having a great business value in the market. It remains a difficult challenge for computer vision due to the immense diversity of garments and a wide range of scenes and backgrounds. In this work, we try to remove the background of fashion images to boost data quality and ultimately increase model performance. Thanks to the fashion image consisting of evident persons in full garments visible, we can utilize Salient Object Detection (SOD) to achieve the background removal of fashion data to our expectations. The fashion image with removing the background is claimed as the “rembg” image, contrasting with the original one in the fashion dataset. We conduct comparative experiments between these two types of images on multiple aspects of model training, including model architectures, model initialization, compatibility with other training tricks and data augmentations, and target task types. Our experiments suggested that background removal can significantly work for fashion data in simple and shallow networks that are not susceptible to overfitting. It can improve model accuracy by up to 5% in the classification of FashionStyle14 when training models from scratch. However, background removal does not perform well in the deep network due to its incompatibility with other regularization techniques like batch normalization, pre-trained initialization, and data augmentations introducing randomness. The loss of background pixels invalidates many existing training tricks in the model training, adding the risk of overfitting for deep models. / Modeförståelse är ett hett ämne inom datorseende, med många applikationer som har ett stort affärsvärde på marknaden. Det är fortfarande en svår utmaning för datorseende på grund av den enorma mångfalden av plagg och ett brett utbud av scener och bakgrunder. I det här arbetet försöker vi ta bort bakgrunden från modebilder för att öka datakvaliteten och i slutändan öka modellens prestanda. Tack vare modebilden som består av synliga personer i helt synliga plagg, kan vi använda framträdande objektivdetektion för att uppnå bakgrundsborttagning av modedata enligt våra förväntningar. Modebilden med att ta bort bakgrunden hävdas vara “rembg”-bilden, i kontrast till den ursprungliga i modedatasetet. Vi genomför jämförande experiment mellan dessa två typer av bilder på flera aspekter av modellträning, inklusive modellarkitekturer, modellinitiering , kompatibilitet med andra träningsknep och dataökningar och måluppgiftstyper. Våra experiment antydde att bakgrundsborttagning avsevärt kan fungera för modedata i enkla och ytliga nätverk som inte är mottagliga för överanpassning. Det kan förbättra modellens noggrannhet med upp till 5 % i klassificeringen av FashionStyle14 när man tränar modeller från grunden. Bakgrundsborttagning fungerar dock inte bra i det djupa nätverket på grund av dess inkompatibilitet med andra regulariseringstekniker som batchnormalisering, förtränad initialisering och dataförstärkningar som introducerar slumpmässighet. Förlusten av bakgrundspixlar ogiltigförklarar många befintliga träningsknep i modellträningen, lägg till risken för övermontering för djupa modeller.

Mental health challenges among Canadian post-secondary students have been on a steep upward trend in recent years, exacerbated by the COVID-19 pandemic. Yet, many students do not reach out to the mental health services available to them. To lessen or remove actual and perceived barriers, research has been exploring how to predict their intentions to seek help for their personal mental health challenges. Much of this work applies the widely used Theory of Planned Behaviour (TPB; Fishbein & Ajzen, 1980) which posits that behavioural intentions are best predicted by attitudes toward the behaviour; perceived normative expectations; and perceived behavioural control. Though the theory also posits that beliefs underlie and are formative of these three predictors, studies have largely neglected their measurement. This is problematic as it deviates from the TPB and frustrates efforts toward the development of interventions to enhance behaviours towards seeking help for mental health challenges. If they are to be effective, such interventions must be directed at changing salient beliefs. This research addresses this gap through a mixed-method sequential design. It provides a unique and valuable contribution to scholarship and practice by identifying and examining the role of students’ attitudinal, normative, and control beliefs with respect to their intentions to seek help. This is examined separately and together with the more traditionally studied direct predictors of attitudes, subjective norms, and perceived behavioural control (i.e., the formative constructs defined by the beliefs). Students relied on six salient beliefs concerning their intentions to seek mental health help: two behavioural, two normative, and two control beliefs. Student attitudes, subjective norms, and PBC mediated the links between salient beliefs and intentions, with PBC being the strongest predictor of intentions to help-seek. The findings help inform interventions to change the beliefs most associated with low intentions to get mental health support. / Thesis / Doctor of Philosophy (PhD) / This dissertation aims to advance the understanding of student intentions or—lack thereof—to seek mental health help following the onset of COVID-19. Salient beliefs that underlie student attitudes, subjective norms and perceived behavioural control are examined, using a Theory of Planned Behaviour approach. The research provides valuable theoretical and practical contribution. It identifies the beliefs held by post-secondary students post COVID-19, which can inform interventions intended to increase mental health service use. In Phase 1 of this study, students’ salient beliefs were examined through survey and interview methods. In Phase 2, I assessed these beliefs as direct predictors of their attitudes, subjective norms, perceived behaviour control, and as indirect predictors of mental health help-seeking intentions. Six beliefs were identified as relevant to student intentions to seek help, and findings lend support for the applied and scholarly value of the TPB to this area of study.

This dissertation is intended to further technology studies by analyzing some of its important methodological tools and using those tools in combination to study complex technological systems in an historical context. The first chapter of the dissertation examines in detail four influential models by which complex technological systems have been analyzed: Hughes's system model, Bijker's social construction model, Latour, Callon, and Law's actor-network model, and Ruth Schwartz Cowan's consumption junction model. For each model, I summarize the seminal works, analyze the uses of the model in the literature, and offer some refinements to the models based on that analysis. Chapter 2 presents three case studies applying these models two different technological systems. First, the early development of the American automobile industry, 1895 to 1940, is studied using Hughes's concepts of technological momentum and reverse salients. Second, the automobile's impact on American society is explored over the same time period relying on Ruth Schwartz Cowan's consumer-oriented perspective and Wiebe Bijker's concept of technological frames. The third case study examines the technological means by which, over a long period of time, American cities were rendered impervious to huge conflagrations—commonplace until the end of the 19th century. For the analysis of this system, I use actor-network theory, Wiebe Bijker's technological frames, and Hughes's reverse salients. / Ph. D.

This dissertation focuses on exploring individual and strategic decision problems in Economics. I take a different approach in each chapter to capture various aspects of decision problems. An overview of this dissertation is provided in Chapter 1. Chapter 2 studies an individual's decision making in extensive-form games under ambiguity when the individual is ambiguous about an opponent's moves. In this chapter, a player follows Choquet Expected Utility preferences, since the standard Expected Utility cannot explain the situations of ambiguity. I raise the issue that dynamically inconsistent decision making can be derived in extensive-form games with ambiguity. To cope with this issue, this chapter provides sufficient conditions to recover dynamic consistency. Chapter 3 analyzes the strategic decision making in signaling games when a player makes an inference about hidden information from the behavioral hypothesis. The Hypothesis Testing Equilibrium (HTE) is proposed to provide an explanation for posterior beliefs from the player. The notion of HTE admits belief updates for all events including zero-probability events. In addition, this chapter introduces well-motivated modifications of HTE. Finally, Chapter 4 examines a boundedly rational individual who considers selective attributes when making a decision. It is assumed that the individual focuses on a subset of attributes that stand out from a choice set. The selective attributes model can accommodate violations of choice axioms of Independence from Irrelevant Alternative (IIA) and Regularity. / Doctor of Philosophy / This dissertation focuses on exploring individual and strategic decision problems in Economics. I take a different approach in each chapter to capture various aspects of decision problem. An overview of this dissertation is provided in Chapter 1. Chapter 2 studies an individual’s decision making in extensive-form games under ambiguity. Ambiguity describes the situation in which the information available to a decision maker is too imprecise to be summarized by a probability measure (Epstein, 1999). It is known that ambiguity causes dynamic inconsistency between ex-ante and interim decision making. This chapter provides sufficient conditions under which dynamic consistency is maintained. Chapter 3 analyzes the strategic decision making in signaling games in which there are two players: informed sender and uninformed receiver. The sender has a private information about his type and the receiver makes an inference about hidden information. This chapter suggests a notion of the Hypothesis Testing Equilibrium (HTE), which provides an alternative explanation for the receiver’s beliefs. The idea of the HTE can be used as a refinement of Perfect Bayesian Equilibrium (PBE) in signaling games to cope with the known limitations of PBE. Finally, Chapter 4 examines a boundedly rational individual who considers only salient attributes when making a decision. The individual considers an attribute only when it stands out enough in a choice set. The selective attribute model can accommodate violations of choice axioms of Independence from Irrelevant Alternative (IIA) and Regularity.

No description available.

Notre thèse s’intéresse à la restructuration du service social d’une plate-forme téléphonique de santé – Santé Info. Au moment de l’étude le service social de ce centre de contact venait de faire l’objet d’une importante réorganisation, qui s’est concrétisée par la création d’une équipe de télé-opérateurs spécialisés. Dans ce contexte de changement organisationnel, notre travail vise à comprendre les incidences du développement de l’activité sur les pratiques situées des acteurs et saisir comment celles-ci nourrissent ce développement en retour. Nos réflexions autour de cette question sont inspirées par les théories de l’activité ainsi que les courants de la cognition et de l’action située. Nous nous appuyons sur la notion de perspective professionnelle et proposons la notion de saillance afin de rendre compte du développement de l’activité en pratique. Nous avons conduit une enquête ethnographique de près de deux ans au sein de cette organisation. Le dispositif méthodologique que nous avons mis en œuvre privilégie l’observation in situ des pratiques. La stratégie observationnelle choisie était celle du suivi, celui i) des acteurs, ii) des affaires c’est-à-dire des demandes traitées par plusieurs intervenants et iii) des situations de communication collective (formations, réunions d’équipe). Nous avons par ailleurs mobilisé la méthode de l’autoconfrontation à partir d’enregistrements audiovisuels. Nos résultats mettent en évidence deux axes de développement contradictoires de l’organisation, à savoir, d’une part, une tendance à l’intégration aux services des mutuelles fondatrices et une tendance à l’autonomisation de Santé Info en tant que prestataire de service externe. D’autre part, la restructuration du service social peut être appréhendée comme actualisant une contradiction sous-jacente au système d’activité, celle entre travail d’information (informer) et conseil (conseiller). Nous analysons les pratiques de prise et de construction de configurations d’indices informationnels des professionnels que nous qualifions de saillances. Nos analyses mettent en évidence que connaître son travail revient à connaître le travail des autres c’est-à-dire à anticiper leurs contraintes, leurs éventuelles difficultés, à reconnaître et tenir compte des impératifs auxquels ils doivent faire face (temporels et de production). La restructuration du service social de Santé Info au travers de la création d’une équipe spécialisée de téléconseillers – le pôle social - contribue à transformer ces pratiques. Par ailleurs, les nouvelles exigences que cette restructuration fait peser sur les professionnels participent de la redéfinition des enjeux de leurs actions en situation. En d’autres termes connaître le travail d’autrui est une ressource pour le collectif mais aussi pour le sujet. Les contributions de ce travail de thèse se situe à trois niveaux : théorique, méthodologique et disciplinaire. L’apport théorique réside dans la recherche d’une articulation entre une analyse orientée par les théories de l’activité et les approches situées de la cognition et de l’action. Ces questions nourrissent des enjeux méthodologiques pour notre travail, qui constituent le second apport de notre recherche. Il s’agit également de construire un dispositif d’observation et de collecte systématique de données dans un environnement professionnel exigeant où le travail est, à la fois, intellectuel et relationnel, la coopération latente et différée, et qui s’étayent sur une infrastructure sociotechnique complexe. Du point de vue de la psychologie du travail, notre étude ouvre à une réflexion autour des compétences collectives et du bien-être au travail. / Our research deals with the organizational change within Health Info, the inbound call center for a number of mutual health insurance companies. Health Info provides users with health information and referral for the purpose of helping them to make better choices in healthcare and to reduce their expenses. In order to improve the service delivered to the clients of the mutual insurance companies, call operators’ activity was redesigned. Some of the most experienced call center agents formed a new team (referred as the « social team »), which had the responsibility to process the information requests of users in need of financial support, different kind of allowances, within outbound calls. In this context, the goals of this doctoral thesis is to understand this organizational change as a development, that of Health Info’ collective activity but also that of the local individual practices. We build upon three approaches to the study of context - activity theory, situated action and situated cognition. Our research is based on a two years ethnographic study. We have used various data collection techniques. Those included ethnographic observation, video recordings of naturally occurring situations as well as interviews with staff members. We systematically attended and video recorded the regular staff meeting of the « social team ». We also shadowed staff members and followed specific « client cases » as first identified by call operators, then processed by the members of the « social team » and eventually transferred for further analysis to the social worker. Our results highlight several contradictions underlying Health Info’ current development. On one hand, Health Info is an additional service designed to fit within the those already existing of the mutual insurance companies. But it is also an autonomous service provider whose clients are the mutual companies. On the other hand, the « social team » reveals another contradiction between different work perspectives, namely that of informing users and helping them or providing them with support. Furthermore, our study underlines the practical dimensions of Health Info’ restructuring. The analysis of situated practices shows that works perspectives are grounded on salient configurations which are both emergent in situ and collectively elaborated cultural resources. The contributions of this doctoral thesis are to address some of the a) methodological and b) conceptual challenges arising from the use of activity theory as a theoretical framework in empirical studies, c) triggers off some interesting considerations on collective competencies and on well-being in the workplace.

Cette thèse est centrée sur le problème de la détection d'objets saillants et de leur segmentation dans une vidéo en vue de détecter les objets les plus attractifs ou d'affecter des identités cohérentes d'objets à chaque pixel d'une séquence vidéo. Concernant la détection d'objets saillants dans vidéo, outre une revue des techniques existantes, une nouvelle approche et l'extension d'un modèle sont proposées; de plus une approche est proposée pour la segmentation d'instances d'objets vidéo. Pour la détection d'objets saillants dans une vidéo, nous proposons : (1) une approche traditionnelle pour détecter l'objet saillant dans sa totalité à l'aide de la notion de "bordures virtuelles". Un filtre guidé est appliqué sur la sortie temporelle pour intégrer les informations de bord spatial en vue d'une meilleure détection des bords de l'objet saillants. Une carte globale de saillance spatio-temporelle est obtenue en combinant la carte de saillance spatiale et la carte de saillance temporelle en fonction de l'entropie. (2) Une revue des développements récents des méthodes basées sur l'apprentissage profond est réalisée. Elle inclut les classifications des méthodes de l'état de l'art et de leurs architectures, ainsi qu'une étude expérimentale comparative de leurs performances. (3) Une extension d'un modèle de l'approche traditionnelle proposée en intégrant un procédé de détection d'objet saillant d'image basé sur l'apprentissage profond a permis d'améliorer encore les performances. Pour la segmentation des instances d'objets dans une vidéo, nous proposons une approche d'apprentissage profond dans laquelle le calcul de la confiance de déformation détermine d'abord la confiance de la carte masquée, puis une sélection sémantique est optimisée pour améliorer la carte déformée, où l'objet est réidentifié à l'aide de l'étiquettes sémantique de l'objet cible. Les approches proposées ont été évaluées sur des jeux de données complexes et de grande taille disponibles publiquement et les résultats expérimentaux montrent que les approches proposées sont plus performantes que les méthodes de l'état de l'art. / This thesis focuses on the problem of video salient object detection and video object instance segmentation which aim to detect the most attracting objects or assign consistent object IDs to each pixel in a video sequence. One approach, one overview and one extended model are proposed for video salient object detection, and one approach is proposed for video object instance segmentation. For video salient object detection, we propose: (1) one traditional approach to detect the whole salient object via the adjunction of virtual borders. A guided filter is applied on the temporal output to integrate the spatial edge information for a better detection of the salient object edges. A global spatio-temporal saliency map is obtained by combining the spatial saliency map and the temporal saliency map together according to the entropy. (2) An overview of recent developments for deep-learning based methods is provided. It includes the classifications of the state-of-the-art methods and their frameworks, and the experimental comparison of the performances of the state-of-the-art methods. (3) One extended model further improves the performance of the proposed traditional approach by integrating a deep-learning based image salient object detection method For video object instance segmentation, we propose a deep-learning approach in which the warping confidence computation firstly judges the confidence of the mask warped map, then a semantic selection is introduced to optimize the warped map, where the object is re-identified using the semantics labels of the target object. The proposed approaches have been assessed on the published large-scale and challenging datasets. The experimental results show that the proposed approaches outperform the state-of-the-art methods.