With the rapid growth of computing power, approaches based on deep learning algorithms have achieved remarkable results in solving computer vision classification problems. These performance improvements are achieved by assuming the source and target data are collected from the same probability distribution. However, this assumption is usually too strict to be satisfied in many real-world applications, such as big data analysis, natural language processing, and computer vision classification problems. Because of distribution discrepancies between these domains, directly training the model on the source domain cannot be expected to generate satisfactory results on the target domain. Therefore, the problem of minimizing these data distribution discrepancies is the main challenge with which modern machine learning is now faced. To address this problem, domain adaptation (DA) aims to identify domain-invariant features between two different but related domains. This thesis proposes a state-of-the-art DA approach that overcomes the limitations of traditional DA methods. To capture fine-grained information for each category, I deploy centroid-to-centroid alignment to perform domain adaptation. An Exponential Moving Average strategy (EMA) is used to ensure we can form robust source and target centroids. A Gaussian-uniform mixture model is trained using an Expectation-Maximization (EM) algorithm to infer the robustness of the target pseudo-labels. With the help of target pseudo-labels, I propose two novel types of classifiers: (1) a target-oriented classifier (TO); and (2) a centroid-oriented classifier (CO). Extensive experiments show that these two classifiers exhibit superior performance on a variety of DA benchmarks when compared to standard baseline methods. / Master of Science / Approaches based on deep learning algorithms have achieved remarkable results in solving computer vision classification problems. These performance improvements are achieved by assuming the source and target data are collected from the same probability distribution; however, in many real-world applications, such as big data analysis, natural language processing, and computer vision classification problems, this assumption is usually too strict to be satisfied. For example, these two domains may have the same types of classes, but the objects in each category of these different domains can vary in shape, color, background, or even illumination. Because the probability distributions are slightly mismatched, directly training the model on one domain cannot achieve a satisfactory result on the other domain. To address this problem, domain adaptation (DA) aims to extract common features on both domains to transfer knowledge from one domain to another. In this thesis, I propose a state-of-the-art DA approach that overcomes the limitation of the traditional DA methods. To capture the low-level information of each category, I deploy centroid-to-centroid alignment to perform domain adaptation. An Exponential Moving Average (EMA) strategy is used to ensure the generation of robust centroids. A Gaussian-Uniform Mixture model is trained by using the Expectation-Maximization (EM) algorithm to infer the robustness of the target sample pseudo-labels. With the help of robust target pseudo-labels, I propose two novel types of classifiers: (1) a target-oriented classifier (TO); and (2) a centroid-oriented classifier (CO). Extensive experiments show that the proposed method outperforms traditional baseline methods on various DA benchmarks.

The Swedish school system is based on basic, democratic values. A fairly new addition to the value system is that the education should provide attention to equal rights between men and women. A way of doing this can be to use fiction in the Swedish language classroom. It can allow students to reflect on other’s living conditions and develop their critical thinking. But fiction has an uncertain future in the language classroom. Studies show that working specifically with the value system is often not prioritized by teachers. Instead, the education focuses on qualities that can be measured. This kind of priority makes fiction lose its potential to provide an opportunity for students to have meaningful conversations about human society. The aim for the study is to explore the didactic potential of the gothic literature, a genre know for its hidden critique towards society. Two gothic short stories from the 1800s are analyzed based on how the characters engage with each other regarding gender power and gender relations. This is conducted through a two part analysis. The first part is a close reading of the short stories based on a gender perspective and the other from a sociohistorical perspective. The analysis show that the characters from both stories uphold gender norms from when the stories were written. The stories’ potential is that they encourage discussions about the interaction between fiction and the living conditions during the 1800s. To further explore the didactic potential of the gothic literature, studies should be conducted where either the students’ experiences of the literature is examined or swedish teachers viewon the gothic literature.

The doctoral thesis deals with an analysis of thermoplastics pipes in a soil environment, in particularly, pipes performed by using trenchless technologies. In the doctoral thesis, experiments of polyethylene pipes loaded by external hydrostatic pressure were performed. The aim of the experiments was to determine a behaviour and load-bearing capacity of the polyethylene pipes loaded by external hydrostatic pressure. For the experiments, a steel pressure chamber was designed and produced which allowed loading and observing a pipe during its loss of the stability. Mechanical properties of the pipe material were determined based on the tensile and bending tests. The series of detailed numerical analyses of the pipe experiments were performed. Various material models which take into account elastic, plastic and viscous behaviour of thermoplastic materials were used. In addition, series of standardized calculations of polyethylene pipes installed using trenchless technologies were performed. The aim was to create new design diagrams for practical assessment of these pipes placed in the partially deteriorated old pipes. In the case of the deteriorated old pipe, a numerical model for determination of design coefficients was created. These design coefficients were verified with standard values. Then, the results of the numerical model were compared with the results of the more complex numerical model which better take into account pipe-soil interaction of various soil types.

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2018-12-21T13:22:18Z No. of bitstreams: 1 José Henrique Nazzi Ehms_.pdf: 4808489 bytes, checksum: b590c723912ffe1d5d90006182f8cff2 (MD5) / Made available in DSpace on 2018-12-21T13:22:18Z (GMT). No. of bitstreams: 1 José Henrique Nazzi Ehms_.pdf: 4808489 bytes, checksum: b590c723912ffe1d5d90006182f8cff2 (MD5) Previous issue date: 2018-10-31 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Armazenamento térmico apresenta grande potencial de utilização em diversas aplicações, como energia solar, climatização, conservação de alimentos ou aproveitamento de calor residual em processos industriais. O armazenamento térmico de calor latente é realizado com materiais de mudança de fase (PCM), através dos processos de fusão e solidificação. A representação destes processos através de simulação numérica é realizada com o acréscimo ao modelo matemático básico, composto pelas equações da conservação da massa, quantidade de movimento e energia, modelos para descrever o calor latente e a transição na velocidade entre as fases, tais como: Darcy STM (source term method), VVM (variable viscosity method) e SOM (switch-off method). No entanto, a grande maioria das pesquisas nesta área utiliza o primeiro método. Além disso, são poucos os estudos comparativos de diferentes métodos para descrever processos de mudança de fase. Assim, o objetivo do presente estudo é comparar os métodos Darcy STM, VVM e misto (utilizando elementos dos dois anteriores) na simulação numérica de processos de fusão e de solidificação do PCM RT27 no interior de uma esfera. O estudo foi realizado utilizando-se fluidodinâmica computacional, através do método dos volumes finitos. O Modelo numérico foi validado com resultados experimentais da literatura. Os resultados quantitativos e qualitativos de fração líquida mostram que o método Darcy STM é mais adequado ao processo de solidificação, enquanto o método VVM produz resultados mais próximos aos experimentais no processo de fusão. O custo computacional foi menor para o método Darcy STM e maior para o método VVM, enquanto o método misto apresenta custo computacional pouco inferior ao do método VVM. Na análise da camada líquida na fusão de contato, foram analisadas a velocidade descendente do sólido, a espessura da camada e a vazão na camada. Os resultados referentes a camada líquida indicaram significativa influência das configurações do método Darcy STM. No entanto, são pouco influenciados pelas configurações do método VVM. / Thermal energy storage presents great potential of utilization in several applications, such as solar energy, HVAC systems, food conservation or waste heat recovery in industrial processes. Latent heat thermal energy storage is realized with phase change materials (PCM), through solidification and melting processes. Representation of such processes through numerical simulation is performed with the addition to the basic numerical model, composed of the conservation equations of mass, momentum and energy, models to account the latent heat and the velocity transition between the phases, such as: Darcy STM (source term method), VVM (variable viscosity method) and SOM (switch-off method). However, the large majority of the research on such area employ the first method. Besides that, there are few comparative studies of different methods to describe phase change processes. Thus, the objective of the present work is compare Darcy STM, VVM and mixed method (using elements of the two prior) in the simulation of melting and solidification processes of PCM RT27 inside a sphere. The study was realized using computational fluid dynamics, with the finite volume method. The numerical model was validated with experimental results from literature. Quantitative and qualitative results of liquid fraction show that Darcy STM is most suitable to solidification process, while VVM produces results closer to experimental in the melting process. Computational cost was smaller for Darcy STM and greater for VVM, while mixed method presents computational cost slightly lower than the one of VVM. In the analysis of the liquid layer in close contact melting, were analyzed descending velocity of the solid, liquid layer thickness and the flow in the liquid layer. The results regarding the liquid layer denote significant influence of the configurations for Darcy STM. However, such results are little influenced by the configurations of VVM.

Denna uppsats inriktar sig på rekvisitet verksam i betydande omfattning (verksamhetsrekvisitet) i relation till ”vanliga anställda” i personalägda fåmansföretag. Det framgår i förarbetena att verksamhetsrekvisitet anses vara uppfyllt när delägarens arbetsinsatser har stor betydelse för vinstgenereringen. Den avgörande faktorn här är om delägaren genom arbetsinsatser genererat intäkter till företaget. Arbetsinsatser av ”vanliga anställda”, dvs. anställda utan någon ledarbefattning eller utan kvalifikationer för företagets kärnverksamhet, anses i vanliga fall inte generera vinster till företaget. Det avgjordes dock några rättsfall där domstolen ansåg att ”vanliga anställda” var verksamma i betydande omfattning. Det är därför relevant att utreda om domstolens bedömningar i de här rättsfallen avviker från syftet bakom verksamhetsrekvisitet. Uppsatsens syfte är att utreda om domstolens bedömningar i rättspraxis som berör rekvisitet ”verksam i betydande omfattning” i relation till ”vanliga anställda” i personalägda företag avviker från lagstiftarens syfte bakom rekvisitet. Lagtexten ger ingen förklaring till hur verksamhetsrekvisitet ska tolkas. I uppsatsen läggs därför vikt vid analysen av förarbeten för att fastställa motivet till lagstiftningen. För att uppnå uppsatsens syfte görs därefter en jämförande analys av praxis med förarbetena. Författaren finner att domstolens bedömningar avviker från syftet bakom verksamhetsrekvisitet beträffande ”vanliga anställda” i personalägda företag. Uppsatsen visar att domstolen stegvis har slutat ta ställning till om arbetsinsatser av sådana anställda har någon betydelse för vinstgenereringen. Domstolen betraktar ”vanliga anställda” som aktivt verksamma utan att ge övertygande skäl till sitt avgörande. Det visas även att domstolens avgöranden påverkar framtida bedömningar i likartade mål. / This thesis focuses on the criterion significantly active in relation to ordinary employees in employee-owned private companies. Shareholders in close companies are considered to be significantly active when their undertakings are essential for the company’s profit. The decisive factor here is whether the shareholders contribute profits to the company. Ordinary employees usually are not concidered to generate profits to the company. Nevertheless, there have been several legal cases in which the Court stated that ordinary employees were significantly active. The question that arises is if the Court’s judgement in these cases deviates from what the legislator meant with the criterion “significantly active” concerning ordinary employees. The purpose of this thesis is to investigate whether the Courts’s judgement in the case law that concerns the criterion “significantly active” in relation to ordinary employees in employee-owned private companies deviates from what the legislator meant with this criterion. The law does not provide any guidance to how the criterion shoud be interpreted. The thesis is based therefore on the comparative analysis of the preparatory work with the relevant case law in order to answer the purpose of this thesis. The conclusion made in this thesis is that the Court’s judgement does deviate from what the legislator meant with the criterion “significantly active” concerning ordinary employees in employee-owned private companies. The analysis shows that the Court does not take into consideration whether such employees’ contribution is profitable to the company. The Court states that ordinary employees are significantly active without giving any convincing justification for its decision. Furthermore, the Court’s decisions affect future judgements in similar cases.

Syftet med denna uppsats var att genom en genusvetenskaplig och socialkonstruktivistisk ansats undersöka hur kvinnor framställs i de svenska hårdrocksmagasinen Close-Up Magazine och Sweden Rock Magazine. Då majoriteten av den genusrelaterade medieforskningen har intresserat sig för reklam- och modebilder finns en kunskapslucka kring hur kvinnor framställs i bilder inom genrer vid sidan av dessa. Denna studie avsåg att genom semiotisk textanalys undersöka bilder i dessa magasin, för att identifiera de kvinnliga stereotyperna i dem och sedan jämföra dessa med etablerade stereotyper som tidigare forskning påvisat. Sammanlagt analyserades 28 bilder i 15 magasin från 2013 utifrån feministisk blickteori och Goffmans teori om könsframställning. Tre tydliga kvinnliga stereotyper framträdde i resultatet. Dessa benämnde vi som Den hårda, Den medkännande och Den objektifierade. De två sistnämnda är väl kända sedan innan inom forskningssammanhang medan Den hårda kvinnan, som bland annat uppvisar tecken på aggressivitet, avsky och överlägsenhet, inte påträffats i liknande studier innan. Denna nyfunna stereotyps blick, som inte definierats i tidigare forskning, namngav vi The Glare. Slutligen påvisar studiens resultat en större nyansering av de stereotyper som tidigare har diskuterats i genusrelaterad medieforskning. Resultaten tyder på att det inte finns en enhetlig kvinnlig stereotyp, utan att de kan variera beroende på genre och kulturell kontext. / The aim of this paper was to examine how women are portrayed in the Swedish metal magazines Close-Up Magazine and Sweden Rock Magazine from a gender theoretical and social constructivist perspective. Since the majority of the gender related media research have focused on pictures in fashion and advertising, there is a lack of awareness regarding how women are portrayed in other genres. This study’s objective was to examine pictures in these magazines in order to identify female stereotypes and compare them to established stereotypes from previous studies. In total 28 pictures from 15 magazines, that were published in 2013, were analyzed based on the understanding of feminist theory and Goffman’s theory of gender display. Three distinctive female stereotypes appeared in the results. These were titled The tough, The compassionate and The objectified. The last two are well known in the media research field while The tough woman, who show signs of aggression, contempt and superiority, has not appeared in previous studies. This new stereotype's view of the observer, which has not been defined in previous studies, we decided to call The glare. The results show a greater variety of stereotypes than what has been discussed in prior media related research. They also indicate that there is no uniform female stereotype and that they vary depending on genre and cultural context.

<p dir="ltr">Accurate 3D point clouds are crucial in various fields, and the advancement of software algorithms has facilitated the reconstruction of 3D models from high-quality images. Notably, both single-camera and multi-camera systems have gained popularity in obtaining these images. While single-camera setups offer simplicity and cost-effectiveness, multi-camera systems provide a broader field of view and improved coverage. However, a crucial gap persists, a lack of direct comparison and comprehensive analysis regarding the quality of point clouds acquired from each system. This thesis aims to bridge this gap by evaluating the point cloud quality obtained from both single-camera and multi-camera systems, considering various factors such as lighting conditions, camera settings, and the stability of multi-camera setup in the 3D reconstruction process. Our research also aims to provide insights into how these factors influence the quality and performance of the reconstructed point clouds. By understanding the strengths and limitations of each system, researchers and professionals can make informed decisions when selecting the most suitable 3D imaging approach for their specific applications. To achieve these objectives, we designed and utilized a custom rig with three vertically stacked cameras, each equipped with a fixed camera lens, and maintained uniform lighting conditions. Additionally, we employed a single-camera system with a zoom lens and non uniform lighting conditions. Through noise analysis, our results revealed several crucial findings. The single-camera system exhibited relatively higher noise levels, likely due to non-uniform lighting and the use of a zoom lens. In contrast, the multi-camera system demonstrated lower noise levels, which can be attributed to well-lit conditions and the use of fixed lenses. However, within the multi-camera system, instances of significant instability led to a substantial increase in noise levels in the reconstructed point cloud compared to more stable conditions. Our noise analysis showed the multi-camera system preformed better compared to the single-camera system in terms of noise quality. However, it is crucial to recognize that noise detection also revealed the influence of factors like lighting conditions, camera calibration and camera stability of multi-camera systems on the reconstruction process.</p>

[pt] Entre 2004 e 2020, 18,9 por cento das cidades brasileiras tiveram ao menos um prefeito que também era empresário. No Brasil, esse cargo é relevante para alocação de recursos governamentais e execução de políticas públicas. Com essa competência constitucional, o representante eleito pode direcionar recursos para sua firma, direta ou indiretamente. Mesmo se isso é feito legalmente, o desempenho empresarial do político eleito pode ser relevante para os eleitores. Por isso, esse trabalho usa 3 bases de dados administrativas, sobre candidatos, donos de empresas e contratos de emprego formal, para responder se as firmas do prefeitos cresceram desproporcionalmente durante seus mandatos. A fim de estabelecer uma interpretação causal, o método de close election é aplicado. Esse trabalho conclui que, ao longo dos quatro anos do mandato, as empresas dos prefeitos cresceram, aproximadamente, 25 por cento a mais do que cresceriam se seu dono tivesse perdido a eleição. / [en] Between 2004 and 2020, 18.9 percent of the Brazilian municipalities had at least one mayor that was also a business owner. In Brazil, this office is relevant for government spending allocation and public policy decisions. With this constitutional competence, the elected official could swing resources to their firm directly or indirectly. Even if this is done within the borders of legality, information on the performance of politician-owned enterprises may be relevant for voters. Therefore, this work uses three administrative data sources on candidates, firm ownership, and formal employment contracts to answer if mayor-owned firms grew disproportional during their owner s term. To provide causal interpretation, the estimations are undertaken applying a close election discontinuity design. This work, therefore, compares firms from barely elected mayors with companies owned by almost victorious candidates. It concludes that mayor-owned companies grew approximately 25 percent more than they would if their owner had lost the election during the four years of the term.

In this work cadmium telluride thin film photovoltaic devices have successfully been produced using a novel closed-field magnetron sputtering technique. This technique offers the possibility of producing cells in an all-in-one vacuum process with the potential to provide a new lower cost production route. The sputtered cadmium telluride layers were characterised in detail using a range of advanced microscopy based techniques both in the as deposited and after the cadmium chloride treated state, a treatment that is necessary to produce a working cell. In the as deposited condition the cadmium telluride layer was seen to have a fine-grained columnar structure containing a high density of stacking faults. After the cadmium chloride treatment these grains recrystallized and the new grains were equiaxed with a much lower density of intragranular defects. Similar effects were also observed in samples prepared using close space sublimation. To understand this recrystallization behaviour during the cadmium chloride treatment, the key treatment parameters were systematically varied. Chemical analysis in Scanning Transmission Electron Microscopy (STEM) showed that chlorine travelled down the cadmium telluride grain boundaries and accumulated adjacent to the cadmium telluride/cadmium sulphide interface. This interface is where the cadmium telluride grains were found to recrystallise first during interrupted cadmium chloride treatments. The nature of the stacking faults was examined using High Resolution Transmission Electron Microscopy (HR-TEM). This showed that in localised regions up to one plane of atoms per sequence was missing based on the expected zinc blende structure. This changed the packing of the atoms such that a local change in crystal structure occurred. This local change in phase was successfully mapped using Electron Backscatter Diffraction in planar section produced using Focused Ion Beam milling. This was subsequently studied in more detail using Transmission Electron Backscatter Diffraction in the Scanning Electron Microscope, where the intra-granular arrangement of the phases was observed. HR-TEM was used to quantitatively measure the linear defects in the cadmium telluride layer after thermal annealing with and without the cadmium chloride present. This showed that annealing alone resulted in only a modest reduction in the density of linear defects and grain recrysallisation only occurred in the presence of cadmium chloride. Cadmium magnesium telluride (CMT) was successfully grown epitaxially onto the cadmium telluride as an electron reflector layer to improve cell performance. During deposition the cell experienced high temperatures and this caused the stacking faults to return in a cell that had been previously cadmium chloride treated. This resulted in a reduction in cell efficiency, providing another link between linear defects and a degradation in cell performance.

This thesis portrays new information concerning the cloud-to-ground (CG) lightning flashes or ground flashes produced by thunderclouds. It emphasizes the importance of characterizing lightning studies as the relationship between lightning mechanisms, and of incorporating the influence of geographical location, latitude and storm type. Sweden, Malaysia and USA were chosen as the main locations for field experiments in 2009 to 2011 to gather a significant number of negative and positive CG flashes. This work provided data on a total of 1792 CG lightning flashes (1685 negative and 107 positive ones) from a total of 53 thunderstorms by monitoring both the slow and the fast electric field and the narrowband radiation field at 3 and 30 MHz signals simultaneously. This thesis is comprised of: (i) the relationship of the Low Positive Charge Region (LPCR) and Preliminary Breakdown Pulse (PBP) trains to the occurrence of negative CG, (ii) slow field changes generated by preliminary breakdown processes in positive and negative ground flashes, and (iii) the occurrence of positive and negative ground flashes. It was revealed that the PBP train appeared have a higher strength in the in Sweden. The strength of the PBP train was caused by the LPCR; in contrast,  weak PBP trains were characteristic in tropical countries constituting insignificant LPCR and needing little energy to break the “blocking” agent to allow the flash to propagate downward to the ground. The second contribution concerns the characteristics of the PBP train mentioned; this includes novel information for Malaysia. Further, it is stated that there are some different characteristics in the PBP trains in Johor, Malaysia and Florida, USA. The studies of slow field changes generated by preliminary breakdown processes clarifies unclear features concerning the starting position of slow field changes generated by preliminary breakdown processes in positive and negative ground flashes. It was found that the slow field changes did not occur before the initial process of the commencement of preliminary breakdown. Single-station electric field measurements incorporating narrowband radiation field measurement and high resolution transient recording (12 bits) with an accuracy of several nanoseconds, allows one to distinguish between the intracloud activities and the preceding processes of ground flashes. The results for the interstroke intervals, amplitude distribution of subsequent return-stroke (SRS) and the number of strokes per flash in the tropics, subtropics and northern regions were similar. Finally, a significant number of positive return-stroke (RS) electric fields provided statistically significant information on the characteristics of these strokes.