In this current research I sought to answer two questions; 1) Do individuals have the capacity to recognize when they are being justly or unjustly socially excluded or conversely socially included? 2) Do the consequences of just and unjust social exclusion or social inclusion vary? In efforts to address these questions, I used perceptions of burden (i.e., participant’s overall contribution to a group task) to manipulate the perceived fairness of one’s inclusionary status to see how this affects the participants’ emotional and behavioral reactions. In Study 1, participants engaged in an imaginary group interaction in which they were burdensome (performing worse than the group) or non-burdensome (performing equal to the group) on a group-task while either being included or rejected. For Study 2, participants were randomly assigned to be burdensome versus non-burdensome, in a similar fashion as Study 1, and then ostracized or included by confederate players in a computerized group word game (i.e., Atimia). Participants in both studies reported their levels of perceived justice, needs satisfaction, social pain, negative affect, and aggressive behavior temptations. Participants in Study 2 also completed a behavioral aggression measure (i.e., candy allocation task). In Study 1, perceptions of justice had no impact on the consequences of social exclusion; rejected participants felt bad regardless of the fairness of their rejection. For included participants, unjust, compared to just, inclusion induced thwarted needs, increased social pain, negative affect, and aggressive behavior temptations (consequences similar to that of social exclusion). In Study 2 almost no differences emerged within the affective state of included individuals. Based primarily on the results of Study 1, it appears that burden may play a critical role in the ostracism experience. Further research is recommended to better understand this relationship.

In a world where artificial intelligence and machine learning aregrowing and spreading in society, its impact and consequence forpeople is increasing. The technology is used in services that peopleuse every day. Both privately but also in a commercial context, forexample social media and to identify fraud in the banking sector.Previous studies show that machine learning models can givediscriminatory outcomes when it comes to, among other things,gender and ethnicity. This study aims to investigate how, in systemdevelopment projects where machine learning is used, one works tocounteract discriminatory outcomes. The study examines both thefactors that contribute to the emergence of discriminatoryoutcomes, as well as the solutions that exist to counteract theproblem. The study is conducted at a global IT consultingcompany.To investigate the area, a study, with qualitative researchmethodology, has been conducted. The empirical material has beencollected through six semi-structured interviews. All respondentswho participated in the study work within the same organization, indifferent projects and with varying experiences in the area. Therespondents have been selected through a subjective selectionbased on their experience in the field in relation to the purpose ofthe study.The results of the study show that the decisive factor for theemergence of discrimination is the training data which the modelsare trained with. The majority of solutions to counteractdiscriminatory outcomes have also been identified. The results ofthe study differ to some extent from the previous research done inthe field. Regarding factors, previous research and the results of thestudy agree that data is the decisive factor that contributes todiscriminatory outcomes arising from machine learning models.The main difference among the solutions is that previous researchshows more specific techniques, which are used to identify ormitigate discriminatory outcomes, while the results of the studyshow softer values and almost no specific techniques at all. In theresults of the study, for example, the individual is seen as a centralpart of the process instead of automatic techniques and tools.The study concludes that data is the most decisive factor indiscriminatory outcomes in machine learning models. The modelsare not discriminatory in themselves, they only reflect the trainingdata. If the data contains discrimination, the model will learn thisand ultimately give discriminatory outcomes. The very basicproblem for this is the human being, who creates the prejudices thatexist in society and from which the data is collected. At the sametime, man is a central part of the process of reducing discriminatoryoutcomes and is needed to counteract this problem. / I en värld där artificiell intelligens och maskininlärning växer ochsprids i samhället ökar samtidigt dess påverkan och konsekvens förmänniskor. Tekniken används i tjänster som människor användervarje dag. Både privat men även i ett kommersiellt sammanhang,exempelvis sociala medier och för att identifiera bedrägerier inombanksektorn. Tidigare studier visar att maskininlärningsmodellerkan ge diskriminerande utfall när det kommer till bland annat könoch etnicitet. Denna studie syftar till att undersöka hur man, isystemutvecklingsprojekt där maskininlärning används, arbetar föratt motverka diskriminerande utfall. Studien undersöker både vilkafaktorer som bidrar till att diskriminerande utfall uppstår, samtvilka lösningar som finns för att motverka problemet. Studiengenomförs på ett globalt IT-konsultbolag.För att undersöka området har en studie, med kvalitativforskningsmetodik genomförts. Det empiriska materialet harsamlats in via sex stycken semistrukturerade intervjuer. Samtligarespondenter som deltagit i studien arbetar inom sammaorganisation i olika systemutvecklingsprojekt samt med varierandeerfarenheter inom området. Respondenterna har valts ut genom ettsubjektivt urval baserad på deras erfarenhet inom området samt irelation med studiens syfte.Studiens resultat visar att den mest avgörande faktorn för uppkomstav diskriminering är träningsdatat som modellerna tränas med.Flertalet lösningar för att motverka diskriminerande utfall har ävenidentifierats i studien. Studiens resultat skiljer sig till viss del motden tidigare forskning som gjorts inom området. Gällande faktorerär tidigare forskning och studiens resultat eniga om att datat är denavgörande faktorn som bidrar att diskriminerande utfall uppstårfrån maskininlärningsmodeller. Den största skillnaden blandlösningarna är att tidigare forskning visar på mer specifika teknikeroch verktyg som används för att identifiera eller mildradiskriminerande utfall, medan resultatet i studien visar mer mjukavärden och nästan inga specifika tekniker alls. I studiens resultatses exempelvis den enskilda individen som en central del iprocessen istället för automatiska tekniker och verktyg. Vidareframkommer det i resultatet blandade åsikter gällande ansvaret förmaskininlärningsmodeller samt behov av regleringar på området.Studiens slutsats är att datat är den mest avgörande faktorn till attdiskriminerande utfall uppstår i maskininlärningsmodeller.Modellerna är inte diskriminerande i sig, utan de speglar bara8. Handledare9. Examinator10. Termin11. Övrigt/AnmärkningKomplettera i alla blanka fält. Gråmarkerade fält skall kompletteras när det finns anledning. I annatfall ska de avlägsnas. För mer information se ”HANDLÄGGNING AV RAPPORT, DEL AV SJÄLVSTÄNDIGT ARBETE(EXAMENSARBETE), INOM NMT”, MIUN 2015/XXX. Det är examinator som är ansvarig för innehållet idetta dokument.träningsdatat. Om datat innehåller diskriminering kommermodellen att lära sig detta och slutligen ge diskriminerande utfall.Själva grundproblemet till detta är människan som skapat defördomar som finns i samhället vilket är där träningsdatat samlas infrån. Samtidigt visar studiens resultat att människan idag är encentral del i processen med att både motverka och identifieradiskriminerande utfall från maskininlärningsmodeller

AI spås få lika stor påverkan på samhället som elektricitet haft och avancemangen inom till exempel maskininlärning och neurala nätverk har tagit AI in i sektorer som rättsväsende, rekrytering och hälso- och sjukvård. Men AI-system är, precis som människor, känsliga för olika typer av snedvridningar, vilket kan leda till orättvisa beslut. En alarmerande mängd studier och rapporter visar att AI i flera fall speglar, sprider och förstärker befintliga snedvridningar i samhället i form av fördomar och värderingar vad gäller könsstereotyper och könsdiskriminering. Algoritmer som används i bildigenkänning baserar sina beslut på stereotyper om vad som är manligt och kvinnligt, röstigenkänning är mer trolig att korrekt känna igen manliga röster jämfört med kvinnliga röster och röstassistenter som Microsoft:s Cortona eller Apple:s Siri förstärker befintlig könsdiskriminering i samhällen. Syftet med denna studie är att undersöka hur könsdiskriminering kan uppstå i AI-system generellt, hur relationen mellan gender bias och AI-system ser ut samt hur ett företag som arbetar med utveckling av AI resonerar kring relationen mellan gender bias och AI-utveckling. Studiens syfte uppfylls genom en litteraturgenomgång samt djupintervjuer med nyckelpersoner som på olika sätt arbetar med AI-utveckling på KPMG. Resultaten visar att bias i allmänhet och gender bias i synnerhet finns närvarande i alla steg i utvecklingen av AI och kan uppstå på grund av en mängd olika faktorer, inklusive men inte begränsat till mångfald i utvecklingsteamen, utformningen av algoritmer och beslut relaterade till hur data samlas in, kodas, eller används för att träna algoritmer. De lösningar som föreslås handlar dels om att adressera respektive orsaksfaktor som identifierats, men även att se problemet med gender bias och könsdiskriminering i AI-system från ett helhetsperspektiv. Essensen av resultaten är att det inte räcker att ändra någon av parametrarna om inte systemets struktur samtidigt ändras. / Recent advances in, for example, machine learning and neural networks have taken artificial intelligence into disciplines such as justice, recruitment and health care. As in all fields subject to AI, correct decisions are crucial and there is no room for discriminatory conclusions. However, AI-systems are, just like humans, subject to various types of distortions, which can lead to unfair decisions. An alarming number of studies and reports show that AI in many cases reflects and reinforces existing gender bias in society. Algorithms used in image recognition base their decisions on character stereotypes of male and female. Voice recognition is more likely to correctly recognize male voices compared to female voices, and earlier 2019 the United Nations released a study showing that voice assistants, such as Microsoft's Cortona or Apple's Siri, reinforce existing gender bias. The purpose of this study is to investigate how gender discrimination can appear in AI-systems, and what constitutes the relationship between gender bias, gender discrimination and AI-systems. Furthermore it addresses how a company that works with the development of AI reason concerning the relationship between gender bias, gender discrimination and AI development. The study contains a thorough literature review, as well as in-depth interviews with key persons working with various aspects of AI development at KPMG.  The results show that bias in general, and gender bias in particular, are present at all stages of AI development. It can occur due to a variety of factors, including but not limited to the lack of diversity in the workforce, the design of algorithms and the decisions related to how data is collected, encoded and used to train algorithms. The solutions proposed are partly about addressing the identified factors, but also about looking at the problem from a holistic perspective. The significance of seeing and understanding the links between gender bias in society and gender bias in AI-systems, as well as reconsidering how each factor depends on and correlates with other ones, is emphasized. The essence of the results is that it is not enough to alter any of the parameters unless the structure of the system is changed as well.

No description available.

Nell’ambito della psicologia fiscale, che studia gli aspetti psicosociali alla base del comportamento dei contribuenti, un numero crescente di ricerche sottolinea l’importanza della buona relazione tra i cittadini, professionisti e autorità fiscali, ai fini della tax compliance. Il progetto studia il “contraddittorio”, uno dei principali touchpoint tra funzionari e professionisti fiscali che si svolge in Agenzia delle Entrate. La ricerca è multi-metodo. Il fenomeno è stato dapprima esplorato con l’“etnografia focalizzata”: sono state individuate quattro strategie di engagement utilizzate dalle autorità durante gli incontri. È stato inoltre trovato che l’incontro vis a vis con le autorità stimola fiducia e la percezione di equità del sistema. Nella seconda parte del lavoro, censuaria, diversi aspetti del contraddittorio sono stati analizzati sul piano quantitativo, sulla popolazione dei funzionari in contraddittorio, attraverso una batteria di item (scale già in uso e scale costruite appositamente). Sono stati individuati i principali aspetti di variazione nell’uso delle strategie. È stata studiata anche la relazione tra appartenenza organizzativa e comportamento sul posto di lavoro dei funzionari. I risultati suggeriscono molteplici aree di indagine rispetto al fare compliance, sia “interna” (verso il personale) che esterna (contribuenti), e numerosi spunti per migliorare l’efficacia dell’azione amministrativa. / In the field of tax psychology, which studies psychosocial antecedents of taxpayer behavior, a growing number of researches emphasize the importance of e good relationship between citizen, tax professionals and tax authorities, for enhancing tax compliance. The project explores the “cross-examination” (“contraddittorio”), one of the main touchpoints between tax officials and taxpayers that takes place in Revenue Agency, thought a multi-method research. The phenomenon was first explored in a “focused ethnography”: four engagement strategies used by authorities during the meeting were identified. The effects of vis a vis with the authorities on trust and the perception of fairness towards the tax system were confirmed. In the second part of the work (census), the several features of the cross examination was quantitatively analyzed on the entire population of tax-officials. A battery of items (made up of validated scales and ad hoc scales developed for the purpose) was used. The main aspects of variation in the use of strategies was identified. The relationship between the tax officials organizational belonging and their behavior at work was also studied. The results suggest multiple areas of investigation with respect of engagement, both “internal” (toward personnel) and external (toward clients) of revenue agency, and ideas for improving the effectiveness of tax administration.

Même si les réseaux de données ont beaucoup évolué au cours des dernières décennies, les paquets sont presque toujours transmis d’un nœud à l’autre comme des blocs de données inaltérables. Cependant, ce paradigme fondamental est aujourd’hui remis en question par des techniques novatrices comme le codage réseau, qui promet des améliorations de performance et de fiabilité si les nœuds sont autorisés à mixer des paquets entre eux. Les réseaux sans fil manquent de fiabilité en raison des obstacles ou interférences que subissent les liens sans fil, et ces problèmes peuvent empirer dans des topologies maillées avec de multiples relais potentiels. Dans ce travail, nous nous concentrons sur l’application du codage réseau intra-flux aux flux unicast dans les réseaux sans fil, avec pour objectif d’améliorer la fiabilité des transferts de données et de discuter des opportunités de déploiement et des performances. Tout d’abord, nous proposons une borne inférieure pour la redondance, puis un algorithme opportuniste distribué, pour adapter le codage aux conditions du réseau et permettre la livraison fiable des données dans un réseau sans fil maillé, tout en prenant en compte les besoins de l’application. En outre, puisque les opérations requises pour le codage réseau sont coûteuses en termes de calcul et de mémoire, nous étendons cet algorithme pour s’adapter aux contraintes physiques de chaque nœud. Ensuite, nous étudions les interactions du codage intra-flux avec TCP et son extension MPTCP. Le codage réseau peut en effet améliorer les performances de TCP, qui ont tendance à être plus faibles sur les liens sans fil, moins fiables. Nous observons l’impact des problèmes d’équité qui se posent quand des flux codés fonctionnent en parallèle avec des flux traditionnels non codés. Pour finir, nous explorons deux manières différentes d’améliorer les performances de MPTCP dans les environnements sans fil : le faire fonctionner sur du codage réseau, et implémenter directement le codage directement dans le protocole MPTCP tout en préservant sa compatibilité avec TCP / Even if packet networks have significantly evolved in the last decades, packets are still transmitted from one hop to the next as unalterable pieces of data. Yet this fundamental paradigm has recently been challenged by new techniques like network coding, which promises network performance and reliability enhancements provided nodes can mix packets together. Wireless networks rely on various network technologies such as WiFi and LTE. They can however be unreliable due to obstacles, interferences, and these issues are worsened in wireless mesh network topologies with potential network relays. In this work, we focus on the application of intra-flow network coding to unicast flows in wireless networks. The main objective is to enhance reliability of data transfers over wireless links, and discuss deployment opportunities and performance. First, we propose a redundancy lower bound and a distributed opportunistic algorithm, to adapt coding to network conditions and allow reliable data delivery in a wireless mesh. We believe that application requirements have also to be taken into account. Since network coding operations introduce a non negligible cost in terms of processing and memory resources, we extend the algorithm to consider the physical constraints of each node. Then, we study the interactions of intra-flow coding with TCP and its extension MPTCP. Network coding can indeed enhance the performances of TCP, which tends to perform poorly over lossy wireless links. We investigate the pratical impact of fairness issues created when running coded TCP flows besides legacy non-coded TCP flows. Finally, we explore two different ways to enhance the performance of MPCTP in wireless environments : running it over network coding, and implementing the coding process directly in MPTCP while keeping it fully TCP-compatible.

Droughts are occurring globally and should be recognized as a global issue and drought planning should use a proactive approach on the part of the world community. However, much drought planning, even in developed and highly developed countries, is reactive and programs are often poorly coordinated sometimes with unforeseen negative consequences for marginalized and disenfranchised populations. Literature pertaining to planning strategy for existing, drought crises is nominal and often contributes to patterns of reactiveness and resulting inequity. To gain a better understanding of crisis-driven planning and the participatory process, this gap was viewed through the lenses of institutional analysis and development and procedural justice and fairness. Specifically, this study was designed to determine how procedural justice and fairness, and the institutional analysis and development framework delineates participatory roles during reactive, crisis-driven planning versus proactive, preparedness planning. A multi-case/within-case analysis was conducted. Six publicly-available documents were selected using provisional and sequence coding lists; emerging themes were also identified at this time. The within-case analysis showed discernable differences between reactive and proactive participatory processes. These findings were used to conduct a cross-case analysis; this analysis indicated that commitment to the participatory process and to change were the keys elements in producing fair and just policies. Drought events can be widely divergent and dynamic, no two being alike; however, the spirit of procedural justice must be part of governance that brings public participation within the reactive planning process into better alignment with proactive planning.

L'industrie des télécommunications sans fil fait actuellement face à une croissance considérable pour des débits toujours plus hauts, stimulée par le développement des services mobiles de données. Ce développement rend le spectre disponible de plus en plus rare et nécessite des solutions afin d'optimiser l'usage de ses ressources limitées.Le principal défi auquel les réseaux sans fils font face est de maximiser la disponibilité, la résilience et la qualité de service, tout en minimisant les coûts et en assurant entre les utilisateurs une allocation de ressources équitable. Cette thèse tente de présenter des solutions à ces problèmes et se focalise sur trois thèmes.Sur le premier thème, le but est de trouver la topologie en anneau qui optimise la disponibilité. Il est montré que les algorithmes développés dans le cadre de la théorie des graphes peuvent être utilisés de manière efficace pour définir en temps polynomial la topologie en anneau optimale si les anneaux sont petits (deux nœuds en plus du nœud d'agrégation). Pour les anneaux plus grands, le problème est NP-hard.Le deuxième thème concerne la polarisation. Nous proposons une solution innovante qui peut améliorer efficacité spectrale jusqu'à 50% par comparaison avec l'état de l'art. Le paradigme proposé introduit de nouvelles perspectives au sujet de l'optimisation de la topologie et de l'allocation de canal.Le troisième thème concerne l'allocation de ressources. Nous remettons en question l'approche présente, basée sur l'optimisation de l'efficacité du réseau. Nous montrons que cette approche est similaire au modèle d'utilité espérée de Bernoulli, qui a été réfuté par les paradoxes d'Allais. C'est pourquoi nous introduisons le concept d'aversion au manque d'équité et considérons la question d'allocation de ressources comme un compromis entre efficacité du réseau et équité. / The wireless telecommunication sector is presently facing a tremendous growth of demand for higher data rates, driven by the development of mobile data services. This development makes the available spectrum scarcer and scarcer and requires solutions in order to optimize the use of its limited resources.The main challenge wireless networks are facing is to maximize availability, resiliency and Quality of Service, while minimizing costs and ensuring fair resource allocation among users.The present thesis will try to present solutions to these issues and will focus on three topics.On the first topic, the purpose is to find the ring-based topology which optimizes availability. It will be shown that algorithms which have been developed in the field of graph theory can be used efficiently to define in polynomial time the optimal ring network topology if the rings are small (two nodes in addition to the aggregation node). For bigger rings, the problem will be NP-hard. The second topic deals with polarization. We propose an innovative solution which can improve spectral efficiency in wireless ring networks by up to 50% in comparison with the state of the art. The proposed paradigm brings new perspectives regarding topology optimization and channel allocation.The third topic deals with resource allocation. We question the present approach based on optimization of network effciency. We show that this approach is similar to Bernoulli's expected utility model, which has been disproved by Allais' paradoxes. For this reason, we introduce the concept of unfairness aversion and consider the question of resource allocation as a trade-off between network efficiency and fairness.

Nous étudions des modèles mathématiques issus de la théorie des files d’attente pour évaluer et optimiser les performances des mécanismes de partage de ressources entre flots dans les réseaux. Dans une première partie, nous proposons des approximations simples et explicites des principales métriques de performance des flots élastiques dans les réseaux à partage de bande passante opérant sous le mode ”équité équilibré”. Nous étudions ensuite le partage de bande passante entre flux élastiques et flux de streaming en supposant que le nombre de ces derniers est limité par un mécanisme de contrôle d’admission et proposons des approximations de performance basées sur une hypothèse de quasi stationnarité. Les résultats de simulation montrent le bon niveau de précision des approximations proposées.Dans une deuxième partie, nous étudions le compromis entre délai et énergie dans les réseaux à partage de bande passante dont les noeuds peuvent réguler leur vitesse en fonction de la charge du système. En supposant que le réseau est initialement dans un état de congestion, on s’intéresse à la politique optimale d’allocation de débit permettant de le vider à coût minimal. L’analyse de la politique stochastique optimale via la théorie des processus de décision markoviens étant extrêmement difficile, nous proposons de l’approximer en utilisant un modèle fluide déterministe qui peut être résolu grâce à des techniques de contrôle optimal. Pour le cas d’un seul lien partagé par plusieurs classes de trafic, on montre que la politique optimale correspond à la règle cμ et on propose une expression explicite de la vitesse optimale. Enfin, dans une troisième partie, on s’intéresse aux plateformes de Cloud Computing dans le cadre du modèle SaaS. En supposant un partage équitable des ressources physiques entre machines virtuelles s’exécutant de manière concurrente, nous proposons des modèles de file d’attente simples pour prédire les temps de réponse des applications. Les modèles proposés prennent explicitement en compte le comportement des différentes classes d’application (tâches interactives, de calcul ou permanentes). Les expérimentations menées sur une plateforme réelle montrent que les modèles mathématiques obtenus permettent de prédire les temps de réponse avec une bonne précision. / We study queueing-theoretic models for the performance evaluation and optimization of bandwidth-sharing networks. We first propose simple and explicit approximations for the main performance metrics of elastic flows in bandwidth-sharing networks operating under balanced fairness. Assuming that an admission control mechanism is used to limit the number of simultaneous streaming flows, we then study the competition for bandwidth between elastic and streaming flows and propose performance approximations based on a quasi-stationary assumption. Simulation results show the good accuracy of the proposed approximations. We then investigate the energy-delay tradeoff in bandwidth-sharing networks in which nodes can regulate their speed according to the load of the system. Assuming that the network is initially congested, we investigate the rate allocation to the classes that drains out the network with minimum total energy and delay cost. We formulate this optimal resource allocation problem as a Markov decision process which proves tobe both analytically and computationally challenging. We thus propose to solve this stochastic problem using a deterministic fluid approximation. For a single link sharedby an arbitrary number of classes, we show that the optimal-fluid solution follows thewell-known cμ rule and give an explicit expression for the optimal speed. Finally, we consider cloud computing platforms under the SaaS model. Assuming a fair share of the capacity of physical resources between virtual machines executed concurrently, we propose simple queueing models for predicting response times of applications.The proposed models explicitly take into account the different behaviors of the different classes of applications (interactive, CPU-intensive or permanent applications). Experiments on a real virtualized platform show that the mathematical models allow to predict response times accurately

Governance is derived from the French word "gouverner", so Good Governance refers to the way to control, steer, rule, and direct the organizations by the individuals who are in charge of the management affairs. The study uses the theory of Alexis de Tocqueville (1956), which emphasizes the importance of the presence of democracy to manage the internal tasks of the organizations, so these NGOs can play a role in fostering democracy in civil society. The study intends to find out the imperatives of good NGOs' governance through linking the daily operational tasks and the applications of democratic principles inside NGOs by using the qualitative research method to collect information and compare the implementation of the imperatives of good NGOs' governance in organizations in Berlin, Cairo, and Prague. Thus, the study presents two normative frameworks; the first one is to conceptualize and operationalize the imperatives of good NGOs' governance through integrating democratic theory with the representation and participation schools, and the second one is to examine the influence of the internal and external factors on the implementation of these imperatives in NGOs. The thesis categorizes the "Good NGOs' Governance Imperatives" into managerial imperatives and democratic imperatives. The managerial...