Discovering new crystalline materials lies at the frontier of modern materials science, driving innovation in energy storage, catalysis, semiconductors, and beyond. The vastness of the chemical and structural space poses a profound challenge: the number of possible atomic arrangements grows prohibitably large with system size and composition. Traditional first-principles methods such as density functional theory (DFT) have revolutionized materials discovery, but their high computational cost limits large-scale exploration. This work addresses the combinatorial bottleneck by bringing together two complementary dimensions of modern materials discovery: data-driven predictions using machine learning and high-performance computing.

The work presented in this thesis builds on a symmetry-aware representation of crystal structures called protostructures, based on Wyckoff positions: a coordinate free description of symmetry related atomic sites. This formulation transforms the continuous space of atomic coordinates into a discrete and combinatorially enumerable one. We developed a machine learning model, Wren, which is trained on this representation to provide fast estimates of stability and guide exploration toward promising regions of structural space. A GPU-accelerated workflow using machine-learning-based interatomic potentials and parallelized screening allows for the evaluation of billions of candidate structures within practical timeframes.

Building on this framework, the presented work enumerates 39 billion binary and ternary compounds spanning the chemical space from lithium to bromine, identifying over 88,000 new structural prototypes, and about half a million new crystal structures within a stability limit of 100 meV/atom. The approach is further applied to experimentally unresolved powder diffraction data, where it reconstructs crystal structures consistent with measured patterns, demonstrating the workflow’s ability to uncover physically realizable materials beyond known prototypes.

To explore even broader regions of structural complexity, this work introduces WyckoffDiff, a diffusion-based generative model that produces novel, symmetry-consistent protostructures beyond the training distribution, some predicted to be thermodynamically stable.

Since pretrained interatomic potentials form the foundation of this work, their quality was examined through two complementary studies. The first benchmarks their accuracy in reproducing mixing enthalpies across disordered alloys. The second investigates how these potentials capture the topology of potential energy surfaces by probing energy variations along symmetry-constrained pathways, showing how different machine-learning potentials represent local minima and saddle points, and other artifacts. These two benchmarks provides insight into their reliability for structure prediction, and the resulting findings informed the selection and parametrization of models used throughout our screening framework.

Altogether, the work presented in this thesis demonstrates that the combination of coarse grained screening, ML-based interatomic potentials, and high-performance computing can dramatically accelerate the discovery of previously unseen crystal structures. The framework presented in the thesis expands the boundaries of computational materials discovery and represents a step toward a large-scale, perhaps even comprehensive, mapping of all stable crystal structures permitted by chemistry and symmetry.

The empirical success of deep learning in a wide range of applications over the last decade has been remarkable. Neural networks can now achieve human-like or superhuman performance at tasks such as image recognition and segmentation,speech recognition, and natural language generation.

Despite decades of research dedicated to understanding how such models learn,there are still many unresolved questions. For instance, neural networks are often severely overparameterized, sometimes with many more parameters than training samples, which according to intuition from classical theory should lead to high sensitivity to noise and poor performance when encountering new data. Yet with enough parameters or training, one can overcome this issue, even without explicit regularization. Understanding implicit biases in training and the induced behavior of neural networks is an important puzzle piece towards understanding how these models learn so efficiently.

This thesis emphasizes the ‘network’ part of neural networks, and uses tools from graph theory to view this class of models from a new perspective that adds to our understanding of their inner workings.

The first paper treats deep linear neural networks, which are neural networks where the nonlinear activations have been removed. The gradient flow equations describing the network’s learning process is an analytically treatable dynamical system, and although it is a simplified model, a deep linear network shares several interesting features with its nonlinear counterpart, such as a non-convex loss function and nonlinear dynamics induced by the overparameterization. The network is considered as a directed acyclic graph and the learning dynamics are described in terms of its adjacency matrix. This reformulation simplifies the gradient flow equations and provides insight into the system properties. For instance,it allows us to highlight an equivalence relation among adjacency matrices, and to investigate stable and unstable manifolds at the critical points of the system without needing to compute the Hessian of the loss function.

The second paper uses the concept of frustration from statistical physics in the context of deep neural networks, and relates frustration to monotonicity of the network when viewed as a function. It is shown that state-of-the-art convolutional neural networks trained on image classification tasks are less frustrated,and thus closer to monotone functions, than what is expected from null models. This suggests an implicit bias in the kind of function that they learn.

For the signed graph associated to a deep neural network, one can compute the frustration level, i.e., test how close or distant the graph is to structural balance. For all the pretrained deep convolutional neural networks we consider, we find that the frustration is always less than expected from null models. From a statistical physics point of view, and in particular in reference to an Ising spin glass model, the reduced frustration indicates that the amount of disorder encoded in the network is less than in the null models. From a functional point of view, low frustration (i.e., proximity to structural balance) means that the function representing the network behaves near-monotonically, i.e., more similarly to a monotone function than in the null models. Evidence of near-monotonic behavior along the partial order determined by frustration is observed for all networks we consider. This confirms that the class of deep convolutional neural networks tends to have a more ordered behavior than expected from null models, and suggests a novel form of implicit regularization.

The aim of this project was to use high-density SNP data to develop a method using four statisticalmodels to predict biogeographical ancestry (BGA) on an intra-continental level. The method aimsto generate DNA intelligence that can provide investigative leads in cases where no missing-personmatches or other indications of the individual’s identity are available. The approach consistedof collecting and processing publicly available reference data from several sources, hierarchicallyclustering populations into regions, and lastly training and testing the four classification models:PCA-ellipse classification, nearest centroid classification, naive Bayes classification, and UMAPellipse classification. For each model, the accuracy was determined and the sensitivity, specificity,and positive predictive value (PPV) were calculated across all regions. The models were also testedto predict BGA for individuals of non-European ancestry and Swedish ancestry to further evaluatethe models. During this study, the nearest centroid classification model had the highest predictiveperformance due to high accuracy, as well as high sensitivity, specificity, and PPV values across allregions. The results obtained from this model are also easy to interpret, which is convenient in aroutine work situation.

This literature review examines high school students’ learning of logarithms and identifies common procedural errors and conceptual misconceptions, as well as teaching strategies that previous research suggests can support learning. The analysis shows that students often make procedural mistakes, such as incorrect manipulation of logarithmic laws, and develop misconceptions where the logarithm is perceived as a symbol rather than a function. To address these difficulties, research highlights four teaching strategies: historical anchoring, linguistic precision, concrete and representative contexts, and reasoning through discussion. These strategies can help strengthen mathematical competencies such as reasoning, handling symbols, and communicating mathematics, and may reduce the risk of mechanical rule application. The study emphasizes the importance of multifaceted instruction that combines clear conceptual explanations, contextualization, and dialogue to promote a deep and sustainable understanding of logarithms.

This study investigates the difficulties students encounter in understanding and using the unit circle in trigonometry instruction at the upper secondary and university levels. The study is a systematic literature review in which the literature is identified through database searches and snowball sampling. An inductive thematic analysis shows that students’ difficulties can be grouped into three themes: difficulties in understanding the properties of the unit circle, difficulties in using the unit circle and connecting it to other areas of trigonometry, and difficulties regarding how and when the unit circle can be used in trigonometry instruction. The results indicate that students who struggle to understand the properties of the unit circle—such as the length of the radius or how points along the circumference are projected onto the x- and y-axes—often experience difficulties in using the unit circle as a broader tool when solving various trigonometric tasks. The findings suggest that a root cause of these difficulties is students’ limited prior knowledge in the area, particularly their lack of understanding of the relationship between the unit circle and the right-angled triangle.

This study scrutinises the shaping and discursive effects of care in preschool teacher education textbooks in Sweden. Drawing on a poststructural approach, regularities of description were analysed, deepening our understanding of how textbooks shape ideas about preschool care practices and preschool teacher professionalism, and how these processes are informed by dominant discourses operating in this field. The study shows how three dominant regularities emerge in the shaping of preschool care: how it is shaped as existing in its own right; how it is shaped through system criticism; and how it is shaped as a societal benefit, aligning with neoliberal views. The study shows how predominant discourses dictate the conditions for preschool care to be shaped, but it also shows possibilities for resistance.

As bicycling becomes an integral part of sustainable mobility, reliable planning tools are essential to ensure bicycling as an efficient mode of transport. The growing bicycle demand requires not only expanding infrastructure, but also ensuring that such infrastructure supports well-functioning traffic under high demands. Given the high heterogeneity in bicyclist characteristics, the use of microscopic traffic simulation, which explicitly considers individual properties and preferences, becomes particularly useful for evaluating bicycle traffic performance. While traffic simulation has been extensively utilized for traffic planning of various modes of transport, this type of modeling support is largely lacking in the planning of bicycle traffic. Although most commercial simulators allow multi-modal traffic analysis, bicycle traffic is often modeled by adjusting parameters in models originally designed for other modes, even though bicyclists may exhibit distinct characteristics and behaviors. Consequently, the proper inclusion of bicyclists into various traffic simulation analyses is difficult, and often inaccurate. The objective of this thesis is to develop and evaluate mathematical models for accurate microscopic simulation of bicycle traffic, with a focus on developing empirically well-founded models that capture the heterogeneity in bicyclists’ characteristics and preferences, as well as their interactions with the built environment and with each other. The thesis delivers an empirical characterization of bicycle traffic in diverse contexts, describing the heterogeneity in characteristics and preferences of bicyclists—including disaggregated analyses by bicycle type— that potentially influence traffic performance. Methods for processing and validating bicycling data are developed to support this characterization. Furthermore, the thesis demonstrates that bicyclist speeds are highly context-dependent and proposes simulation models for context-related features of bicycling trips, such as topography, curvature, and wind, that integrate heterogeneous and adaptive free riding behavior to improve the accuracy of simulated speeds and the reliability of bicycle traffic simulations. This thesis advances the accuracy and applicability of microscopic simulation of bicycle traffic for its use in the planning of well-functioning bicycle traffic.

System of Systems (SoS) architectures rely heavily on information exchange to coordinate the independent constituent systems (CS) towards a shared mission goal. However, communication dynamics are often overlooked in the early stage of SoS modelling. This thesis explores how communication and control configurations influence the mission outcomes and individual system behaviour within an SoS. 

A forest wildfire scenario is modelled using Agent-Based Modelling (ABM) in NetLogo, where multiple firefighting assets such as UAVs, helicopters, water tankers and bulldozers collaborate to suppress the active fire. Three different scenarios of communications were evaluated: a fully decentralised mode where agents act on their own discretion and possess global knowledge of fires, a centralised command where the agents rely on an incident commander for orders and a third scenario, which is similar to the second scenario but adds a delay to regulate the information flow.

The obtained results showcase that excessive communication causes command bottlenecks and incomplete missions, thereby affecting the operational outcomes. By throttling communications, the mission completion rates go up, as well as the overall outcomes of the operation.

The findings also validate low-fidelity agent-based models as an effective tool to assess SoS architectures and highlight the strategic role of communication in carrying out a successful operation. The work provides a methodological foundation for future architecting of an SoS applicable to any mission-driven operations.

Social robots are expected to become ubiquitous across different areas of both the private and public spheres of human life, assisting in care, education, and daily tasks. Hopes for meaningful, efficient, and enjoyable interactions with these kinds of robots drive social robotics research, but there are also ethical concerns and critiques concerning negative social consequences. Both sides involve assumptions about the boundaries around what it means to be human in relation to the non-human. In order to disentangle optimistic and pessimistic notions about this relationality, there needs to be a reconsideration of how the boundaries constituting the human/non-human divide are understood and figured in encounters with robot bodies. Following Dominika Lisy’s personal experiences that have been documented as reflections, diary entries, and by photographing moments of meeting different social robots, this thesis starts at the first point of contact: the skin.

But how might the skin be used to rethink boundaries between humans and social robots? The research in this thesis illustrates an interdisciplinary endeavour to weave together insights from feminist theory and methodology, research on affective and tactile robots, and the biology and neurophysiology of the skin. Grounded in feminist new materialism, which embraces the entanglement of matter and discourse, the thesis develops a figuration of the skin through which both the form and content of the thesis’ text illustrate what it means to pay attention to boundaries during encounters with affective and tactile robots. Figuring boundaries in this way, they can be described as multilayered, flexible yet sturdy, hardening over time, and de-pendent on un/noticeable sensations, just like the skin. This thesis aims to contribute to feminist theories of human/non-human relationality. The thesis presents a situated feminist account of making sense amidst diverse knowledges through the skin and personal embodied experiences in order to develop a sensitivity and ethical responsibility for human–robot relations.

Understanding how circulating proteins and metabolites co-vary at the population level is essential for characterising systemic physiology and for interpreting how diseases perturb biological signalling. Yet, most studies analyse proteomics and metabolomics separately, leaving the structure of their joint variation—and the signalling modules that connect them—largely unresolved. This thesis addresses the problem of how to construct a statistically principled, demography-adjusted circulating protein–metabolite signalling network that can serve as a reference framework for future biomedical and epidemiological research.

We integrate large-scale affinity-based plasma proteomics measured using the Olink platform (approximately 3,000 proteins) together with nuclear magnetic resonance(NMR)–based plasma metabolomics (covering lipids, lipoprotein subclasses, amino acids,and related metabolic traits) obtained from a population-based cohort (blood samples) comprising over ten thousand individuals after quality control. Using these data, we develop a fully transparent statistical workflow that combines: (i) rigorous preprocessing and demographic adjustment; (ii) large-scale association discovery with false-discovery rate control; (iii) bipartite network construction and community detection; and (iv) multivariate integration through classical, ridge-regularised, and sparse canonical correlation analysis. The methods are chosen to maximise statistical robustness, interpretability, and reproducibility in a high-dimensional setting.

The resulting network reveals coherent biological modules—most prominently lipid transport, inflammatory signalling, and amino-acid metabolism—and identifies multivariate axes of shared protein–metabolite variation that are stable across CCA formulations. Collectively, the thesis provides a statistically grounded blueprint for quantifying cross omic structure in human populations and offers a reproducible foundation for future work on mechanistic pathways, risk stratification, and disease-related perturbations.

The increasing demand for customized products, together with the need for flexible, human-centric, and resilient manufacturing systems has intensified the need for automation solutions capable of operating in dynamic and unstructured industrial environments. This dissertation shows how automation methodologies evolve and where they fail as complexity and variability increase across the design and production domains.

The research begins by addressing time-consuming and iterative engineering tasks through design automation. Using Knowledge-Based Engineering (KBE) approaches, automated frameworks were developed to streamline engineering workflows and support consistent decision-making in structured industrial settings. However, when extending the focus to real-world production, the growing complexity and uncertainty of the environment create substantial challenges for both design and production automation.

While KBE provides structure and consistency, its reliance on predefined rules necessitates standardization, resulting in a rigid design space and limited adaptability. To overcome these inherent restrictions, the research integrates complementary tools and techniques that enable more flexible and adaptive automation. Camera vision captures real-world conditions and tracks changes in the environment, while large language models, combined with an agent-based approach, provide reasoning capabilities that interpret variations in products or processes and generate adaptive decision-making strategies. Digital twin simulations validate and predict the outcomes of these variations in a virtual environment, allowing the system to respond proactively and safely by reconciling real-time data with simulation outcomes.

Overall, this work contributes a holistic and scalable automation methodology that unifies design automation, adaptive digital twins, and knowledge-driven reasoning. The results demonstrate how structured engineering knowledge, combined with reasoning and adaptive technologies, enables the development of resilient automation solutions for the increasingly unstructured landscape of future Industry.

This industry white paper presents a research design for studying Sweden's first deployment of unaccompanied autonomous delivery robot fleets in Stockholm, marking an important milestone in Swedish urban robotics implementation. Through a two-year longitudinal study combining ethnographic field observations, video data collection, and multi-stakeholder workshops, we will document the practical realities of integrating delivery robots into dense urban environments. We aim to identify recurring operational barriers including permit uncertainties, challenges around community integration, and robot-human interaction patterns across different seasons and contexts. We approach this complex problem through cross-sector collaboration between academia, industry, municipalities, and regulatory bodies, with the goal to develop actionable policy implications for cities, robot operators, and national policymakers as well as contributing to international standards development. 

The heavy-duty freight sector finds itself in a state of change. Legislation and customer demand pushes the industry towards electrification, where the pure battery and fuel cell electric hybrid have emerged as the top contending technologies to replace the conventional diesel powertrain, and although pure battery powertrains dominate the light-duty sector, projections indicate that fuel cell hybrids will show superior performance in long-range missions with heavy cargo. Particularly so in the context of future autonomous vehicles, considering the potential for continuous, non-stop driving. Regardless, several techno-economic challenges remain before widespread adoption of either pure battery powertrains or fuel cell hybrids in the heavy-duty sector. These challenges include but are not limited to the weight of lithium cells, elevated hydrogen prices, insufficient recharging/refuelling infrastructure, durability, as well as thermal management related issues.

A model-based approach can be used to target these challenges, which motivated the development of the Electrochemical Commercial Vehicle (ECCV) platform; a model library tailored for controls algorithm development and rapid virtual prototyping of electrified trucks. While auto-manufacturers use in-house and proprietary software to solve similar tasks, the open-source nature of the ECCV-platform allows for collaboration within academia and industry without issues relating to intellectual property rights, a type of collaboration which was demonstrated in a benchmark competition held at the IFAC World Congress 2023, where six teams from universities all over the world contributed their solutions to the fuel cell hybrid energy management problem.

Although the development of the ECCV-platform constitutes a major part of the thesis, further work was done to improve its capacity. First, the platform was extended with the capability to model state-of-the-art thermal systems, specifically through the inclusion of refrigerant models. This extension allowed for an investigation into the effectiveness of various heat pump systems for pure battery trucks. Secondly, the platform was extended with a fuel cell model validated in the intermediate temperature range, which enabled an investigation of how elevated stack temperatures may benefit overall system efficiency. Thirdly, the model library was used to develop a real-time energy management algorithm for fuel cell trucks, demonstrating the value of the platform also in the context of optimal control.

Centrum för klimatpolitisk forskning (CSPR) genomförde under 2023 en utvärdering av Linköpings universitets (LiU) klimatarbete. I detta brief pekar vi på de mest akuta utmaningarna med nuvarande klimatarbete och presenterar fyra rekommendationer för hur LiU:s klimatarbete kan utformas i linje med dess vision om att bidra till en bättre värld (LiU:s vision 2030). 

Idiopathic normal pressure hydrocephalus (iNPH) is a common and treatable neurological disorder in older adults with ventriculomegaly due to disturbed cerebrospinal fluid (CSF) dynamics. The pathophysiology of iNPH is still incompletely understood, and there is growing evidence that neuroinflammation is partly responsible. Ventricular size is radiologically assessed with linear measures such as the Evans index (EI), but these measures are insensitive to clinically relevant changes in ventricular volume, so postoperative monitoring remains challenging. Shunt treatment is frequently complicated by overdrainage, underdrainage and suspected malfunction.

The overall aim of this thesis was to develop and evaluate a clinically feasible quantitative tool for monitoring shunt treatment and apply it in a clinical cohort of iNPH patients with a special focus on the following: 1. Ventricular volumetry based on 3D quantitative MRI (qMRI). 2. Very early postoperative ventricular change in relation to initial valve setting and in vivo MRI resistance of a modern adjustable shunt valve. 3. Long-term relationships between volumetric change, clinical response and overdrainage 4. Inflammatory and neurodegenerative biomarker profiles in iNPH patients across plasma, lumbar CSF and ventricular CSF.

The LiNPH study, a prospective controlled clinical longitudinal cohort study with a randomized controlled trial component, provided the data for the four papers in this thesis. Paper I was a methodological reliability study (45 qMRI examinations from iNPH patients and controls) comparing manual, automated and reviewed ventricular segmentations. Paper II was a randomized, double-blinded clinical study (n=45) with qMRI volumetry preoperatively and ~36 hours post-operatively that compared Certas Plus shunt valve settings 4 (open shunt) versus 8 (virtually closed) and evaluated the in vivo MRI stability of the valve. Paper III was a prospective cohort study (50 enrolled; follow-up at 3, 12 and 36 months) relating ventricular volumetry to clinical outcomes, valve adjustments and complications. Paper IV utilized CSF and plasma samples collected preoperatively from iNPH patients (n=56) and neurologically healthy controls (n=19) to profile 71 inflammatory cytokines, chemokines, growth factors and classical neurodegenerative markers, including paired plasma, lumbar CSF and ventricular CSF, via multiplex immunoassays and multivariate modelling.

qMRI-based ventricular volumetry demonstrated excellent intra- and interrater reliability (intraclass correlation coefficient 0.999--1.000), whereas EI failed to reflect several clinically important volume differences. Early after shunting, ventricular volumes decreased significantly in both randomized groups, but the reduction was greater at setting 4 than at setting 8 (mean 16±9 mL vs 5±5 mL), whereas linear measures were equal except for THC (tight high convexity, the amount of CSF in the parasagittal convexity sulci). There were no inadvertent valve adjustments after 156 MRI examinations were performed. Over three years, the mean ventricular volume decreased from 134±35 mL preoperatively to 97±31 mL at 36 months, whereas gait and the iNPH scale improved markedly early and remained improved at the group level. Compared with nonresponders, responders presented greater reductions, and patients with overdrainage symptoms (orthostatic headaches and more) presented greater reductions than asymptomatic patients did. Biomarker analyses revealed a robust iNPH inflammatory profile compared with that of the controls and strong compartment concentration separation.

In conclusion, qMRI-based ventricular volumetry is a sensitive, reliable and clinically feasible method for longitudinal iNPH follow-up, providing clinically valuable information on shunt effects, valve adjustments, suspicion of malfunction and overdrainage risk. The Codman Certas Plus valve is MRI resistant in vivo. There is a clear inflammatory component in iNPH pathophysiology, and there are substantial differences in cytokine and neurodegenerative biomarker concentrations among ventricular CSF, lumbar CSF and plasma, which are relevant to future study designs.

The project investigated the potential of using Direct Air Capture (DAC) integrated into ventilation systems to capture metabolic and atmospheric CO₂ in university environments. An upgraded laboratory-based test facility using zeolite 13X was developed, with particular emphasis on studying the impact of humidity on capture efficiency through the installation of a dehumidifier that enables controlled regulation of air moisture levels.

The experimental results demonstrate that reduced humidity is critical for achieving high capture efficiency. A relative humidity of approximately 8% enabled CO₂ capture rates exceeding 90%, compared with around one-third without dehumidification. The techno-economic assessment included all campus areas of Linköping University and shows a total annual capture potential of 8,031 tonnes of CO₂. The associated energy use amounts to 4,015 MWh of electricity and 12,046 MWh of heat per year, corresponding to annual energy costs of approximately SEK 10 million. Potential revenues from the sale of captured CO₂ are estimated at SEK 20 million per year, resulting in an annual net return of SEK 10 million when energy costs are considered. Although installation costs for the DAC units and potential revenues from carbon credits were not included in the analysis, the results indicate that DAC technology can represent a significant component in efforts toward carbon-neutral university environments.

This thesis addresses the longstanding problem of joint state and parameter estimation in state-space models, with particular emphasis on the small-sample regime where the classical maximum likelihood (ML) can become biased and statistically inefficient. To overcome the drawbacks of ML under such scenarios, we analyze the joint maximum a posteriori-maximum likelihood (JMAP-ML) principle and show that, under the state-space setting, the parameter update step in JMAP-ML induces an errors-in-variables (EIV) structure because it uses the first-order moment of state estimation while ignoring the covariances. This observation motivates incorporating EIV-based estimators in the parameter estimation to reduce bias and improve robustness. Building on this, we develop an iterative alternating framework that unifies expectation maximisation (EM)/ML, standard JMAP-ML, and EIV-enhanced JMAP-ML variants, enabling principled choices across different data regimes.

On the state-estimation side, we study smoothing with uncertain model parameters and propose the Partial EIV (PEIV)–based and the bias-compensated least squares (BCLS)-based Kalman smoother, which explicitly consider parametric uncertainty in the smoothing step. Simulations show that this approach yields better accuracy over traditional smoothers when parameter uncertainty is non-negligible.

In terms of the parameter estimation problem, we review and integrate methods that either compensate for or explicitly model uncertainty in the regressors. In particular, we study bias-compensated least squares and a suite of total least squares (TLS) methods (weighted, constrained, partial). Under the assumption that the noise in the regressor is independent of the regressors themselves, both of these two approaches can improve the estimation accuracy compared with the standard least squares method. The bias-compensated methods rely on probability-limit approximation to the analytical solution of bias and therefore requires sufficient data to be reliable. In terms of the comparison between the TLS-type methods and the BCLS method when the noise in the regressor is dependent on the regressor itself, it is shown that the BCLS method mitigates the effect of this dependence more effectively than the TLS-type methods when the bias approximation is accurate (enough data).

Beyond JMAP-ML and its variants, maximum likelihood—implemented directly or via EM under a state-space formulation—is a popular choice owing to its asymptotic optimality. To assess batch-size effects, JMAP-ML-type methods are compared with EM, highlighting their practical differences. In joint state parameter estimation, a key difficulty is that the state estimates are noisy and are correlated with the process and/or measurement noise. A central implication for TLS-based identification is that the regressor (the state sequence) and its perturbation are driven by the same process noise, violating the independence assumption and thereby explaining the degraded performance of TLS in this setting.

Extensive Monte Carlo simulations support several main conclusions. First, there is no universal winner across all scenarios. Second, with small samples, standard and TLS-enhanced JMAP-ML are typically the most competitive, whereas in the middle-sample regime BCLS-enhanced JMAP-ML outperforms them, which reflects its probability-limit based bias correction. Third, in the large-data regime, maximum likelihood (direct or via EM) can surpass the alternatives due to its asymptotic optimality. Overall, the thesis advocates explicit uncertainty modeling in both the parameter- and state-estimation stages and presents a unified framework that practitioners can tune to data size, noise characteristics, and model fidelity.

This thesis investigates the 2nd-year (11th grade, aged 16-17) High School students’ experiences and attitudes concerning high-stakes testing (HST) and shadow education, in Greece. The Greek education system requires students of the 3rd year of High School to participate in the Panhellenic Exams in order to secure a position in public universities. These are highly competitive high-stakes exams in four core subjects and carry significant academic and social weight. In order to achieve high scores, students turn to private tutoring, commonly referred to as shadow education.Although students are the ones mostly affected by the above phenomena, their perspectives and lived experiences have received limited attention in existing research. The specific study, with the adoption of a child-centered approach and through the lens of interpretivism and phenomenology as well as the ontological framework of constructionism, perceives children as active agents who construct their educational experiences. For this reason, data was collected through semi-structured interviews with six 2nd-year High School students, who provided a range of experiences and insights into their exam preparation process.The study explores how these students describe and talk about their experiences with HST and shadow education in their everyday school life, motivation, engagement and emotional well-being. For this reason, thematic analysis was used to offer a nuanced understanding of the way they experience and make sense of their educational reality. With the amplification of students’ voices, the thesis wishes to contribute to the fields of Child Studies and Sociology of Childhood, through its emphasis on the importance of recognizing students as competent participants in the educational process, whose perspectives are central to discussions about educational policy.The results of the study identified four key findings. First, students reported structural deficiencies in public education (overcrowded classrooms, limited instruction time, lack of personalized support) that minimize effective learning and their motivation. Second, they expressed disengagement from formal schooling, especially for non-core subjects and underscored a narrowed, exam-led curriculum, repetitive and superficial. Third, shadow education was described as the primary and normalized means for academic success, considered more important than[4]mainstream schooling. Finally, the emotional consequences of this dual system were noted - in particular anxiety, stress, social sacrifices and an internalized obligation to meet expectations.

Denna rapport tar sin utgångspunkt i frågan om konsumtion för hållbar omställning med fokus på hållbar matkonsumtion. Inom detta område finns olika nationella mål och överenskommelser. Dessa omfattar bland annat Mil­jömålsberedningen förslag att införa nationella konsumtionsbaserade miljömål som innebär att Sverige senast år 2045 ska ha netto noll konsumtionsbaserade utsläpp (SOU 2022:15). Matkonsumtion står för omkring 20 procent av Sveriges konsumtionsbaserade utsläpp och är ett område där Livsmedelsverket och Folkhälsomyndigheten (2023) ställt upp mål och insatsområden. Dessa siktar mot en hållbar och hälsosam matkonsumtion med syfte att ge en tydlig inriktning för folkhälso- och hållbarhetsarbetet på lokal, regional och nationell nivå.

Emellertid är ansvarsfördelningen mellan dels de politiska nivåerna, dels mellan samhällssektorerna (den offentliga och privata sektorn liksom civilsamhället) för att genomföra insatser för att uppnå uppsatta mål och att bidra till minskad miljöpåverkan från matkonsumtion otydlig. Den här rapporten studerar hur kommunerna ser på sitt ansvar i det här sammanhanget och väljer att genomföra insatser för hållbar matkonsumtion bland invånare och besökare i lokalsamhället.

Vi har besökt alla 290 kommuners webbplatser för att söka efter om man beskriver policy, insatser eller ger information med koppling till hållbar matkonsumtion. Resultatet från detta visar att omkring 20 procent av kommunerna tar upp den här frågan från något perspektiv. Med detta som utgångspunkt har därefter sju kommuner valts ut för en fördjupad studie. I dessa har vi genomfört intervjuer, främst med tjänstepersoner, samt studerat kommunala styrdokument. Vi har kodat materialet utifrån en modell om hållbar omställning som skiljer ut tre sfärer: den politiska, den praktiska och den personliga. Inom den politiska sfären har vi tagit del av kommunernas styrdokument. Den praktiska sfären omfattar olika typer av aktiviteter som kan vara kampanjer, event, samt dialogbaserade initiativ. Den personliga sfären, avslutningsvis, handlar om individers föreställningar om hur kommunen kan vara en möjliggörare för hållbar matkonsumtion. Det kan vara i olika roller som exempelvis markägare, upphandlare och kunskapsspridare.

Sammantaget finner vi att frågor om konsumtion generellt och hållbar matkonsumtion specifikt, kan ses som ett område som är under kontinuerlig förändring samt att de organisatoriska strukturerna och sambanden har en betydelse för hur insatser för hållbar matkonsumtion genomförs. Variationer mellan kommunerna motiverar framtida studier i flera av Sveriges kommuner för att få en bredare bild av hur insatser för hållbar matkonsumtion genomförs genom lokala initiativ.

Artificial intelligence (AI) has advanced rapidly across engineering and industrial domains, yet its adoption in production environments is often constrained by the need for higher system reliability, limited availability of high-quality data, and the challenge of embedding tacit engineering knowledge into learning-based models. These limitations hinder the broader industrial push toward flexible, data-driven automation capable of handling increasing product variability and shorter production cycles.

This thesis investigates how learning-based methods can be designed and integrated as re-liable perception components within production workflows. Through four complementary case studies, the work demonstrates how visual representations—ranging from Computer-Aided Design (CAD)-derived images and engineering drawings to point clouds and RGB-D images—can be leveraged to address concrete industrial challenges across multiple stages of the manufacturing pipeline.

The first case study predicts fixturing clamp configurations for welding operations in automotive manufacturing by learning geometric patterns from CAD-derived representations. The second applies optical character recognition to engineering drawings to accelerate quality-control and documentation workflows. The third examines scene reconstruction and 3D object detection from point clouds, using synthetic data generation to mitigate data scarcity. The fourth develops a fast, zero-shot pose estimation approach for robotic manipulation, enabling reliable object localization in dynamic industrial environments.

Taken together, these studies show how AI methods informed by structured engineering knowledge can increase process efficiency, reduce manual workload, and help resolve persistent automation bottlenecks in modern manufacturing.

This study emphasizes the importance of automatic synthetic data generation in data-driven applications, especially in the development of a 3D computer vision system for engineering contexts such as brownfield factory projects, where no data is readily available. Key points: (1) A successful integration of a synthetic data generator with the S3DIS dataset, leading to a significant enhancement in object detection of previous classes and enabling recognition of new ones; (2) A proposal for a CAD-based configurator for efficient and customizable scene reconstruction from LiDAR scanner point clouds.

This thesis proposes a Nietzschean alignment and transvaluation of Artificial Intelligence (AI) ethics through the novel ethical framework of Biocracy — aligning AI to foster an environment selecting the strengthening of our biological capacities towards human flourishing. Current dominant approaches in AI ethics emphasize fairness, accessibility, expediency or transhumanism — yet these frameworks fundamentally, even if unintentionally — accelerate the degeneration of human cognition as phenotypic skill atrophy in the short term, and in the longer term, if such cultural environments persist, relaxed selection will contribute to genotypic shifts across generations. The erosion of those very capacities — capacities that made humans distinctive as a species and enabled the development of AI itself (such as large language models, search engines, tutoring models or navigational aids which may either complement or compete with our biological cognition) — first leads to human decline before possibly into their subjugation or even extinction in the long run. Drawing from evolutionary theory and the philosophy of Nietzsche, I argue that AI should be designed as a complementary technology (eutech) — one that challenges users to strengthen their biological cognitive abilities (phenotypically) and fosters an environment selecting for them across generations (genotypically) — rather than as a competitive technology (dystech) that augments users with artificial cognitive abilities while diminishing and selecting against their biological ones. Through Biocratic Principles (adaptive challenge scaling, selective pressure implementation, and competence expiration), I advance a normative framework in which access to AI's augmenting capabilities is proportionally granted by the user's demonstrated biological competence. In this alignment, AI becomes a testing apparatus rather than a debilitating crutch — an evolutionary environment catalyzing human flourishing instead of decline in both short- and long-terms, fostering against competing alternatives not only our survival and dominance but, above all, life affirmation. 

Matematisk  problemlösning utgör en central del av matematikundervisning i grundskolan och ställer höga krav på elevers begreppsförståelse, resonemangsförmåga och språklig kompetens. Nationella prov används som viktigt verktyg för att bedöma elevers kunskap, men hur resultaten faktiskt används i undervisningen och för att identifiera elevers svårigheter är mindre utforskat särskilt i mellanstadiet. Mot denna bakgrund syftar studien till att undersöka hur lärare i årskurs 4–6 arbetar med matematisk problemlösning samt hur de använder resultat och information från nationella prov för att identifiera elevers svårigheter och utveckla undervisningen. Med en kvalitativ ansats genomfördes intervjuer med sex mellanstadielärare som analyserades genom tematisk analys. Resultatet visar att nationella prov fungerar som ett viktigt men begränsat stöd för att bedöma elevers problemlösningsförmåga. Lärarna betonar att provresultaten behöver kompletteras med egna observationer, diagnoser och formativ bedömning för att ge en mer nyanserad bild av elevernas kunskaper. En central utmaning som framkommer är bristande överlämning mellan lågstadiet och mellanstadiet, där kvalitativ information om elevers strategier, resonemang och språkliga förmågor ofta saknas. Vidare framhålls språkets och läsförståelsens avgörande betydelse för elevers framgång i matematisk problemlösning. Många elevers svårigheter kopplas främst till tolkning av textuppgifter snarare än till matematiska begrepp. Lärarna arbetar därför språkutvecklande och använder varierade strategier såsom gemensamma genomgångar, visuellt stöd, strukturerade lässtrategier och muntligt resonemang.

En central slutsats som kan dras är att nationella prov kan bidra till undervisningsutveckling på både individ- och gruppnivå, men att deras värde är beroende av hur resultaten tolkas och kombineras med annan pedagogisk information. För att stärka elevers problemlösningsförmåga efterfrågar lärarna mer systematiska överlämningsrutiner, ökat kollegialt samarbete och ett tydligare fokus på språk och resonemang i matematikundervisningen.

Background: As project management evolves towards hybrid approaches that combine traditional and agile principles, project managers face increasing complexity in navigating the tension between them. While research has explored managerial solutions, little attention has been given to leadership perspectives. Existing frameworks solely reflect agile leadership, offering little guidance on how to manage the tensions between the agile and traditional principles. This leads to a disconnect between theory and practice as hybrid projects are not simply about embracing one approach over another, but rather about integrating both in a way that addresses the demands of each. 

Purpose: This study aims to explore how project managers in hybrid projects navigate leadership tensions, and whether their leadership style is agile, traditional, or a hybrid of both. Ultimately, the purpose is to build a theoretical framework that reflects the experiences of project managers and supports both explanatory depth and practical guidance.

Methodology: This study adopts a qualitative research design, including semi-structured interviews with eight project managers with leadership responsibility in hybrid projects. The analysis follows an inductive approach, allowing themes and patterns to emerge from the data, as well as a thematic approach to interpret recurring themes.

Empirical findings: This study shows that all project managers adopted hybrid leadership to navigate the contrasting leadership demands in hybrid projects. Furthermore, the empirical findings revealed underlying logics guiding leadership behavior, shaped by various contextual and interpersonal influential factors. 

Conclusion: Building on the empirical findings, this study provides a framework that highlights how project managers face multiple contrasting leadership demands, with all but one giving rise to leadership tensions. Instead of offering a one-size-fits-all solution, this framework contributes by illustrating the dynamic nature of hybrid leadership. 

En uppsats med fokus på hur lärare undervisar om nationella minoriteter, vad de använder för material samt vilka uppfattningar lärare har om undervisningen i området. Grunden för analysen är en enkätundersökning som genomfördes år 2020 där 33 svensklärare deltagit. Materialet analyserades därefter med hjälp av tematisk analys och med utgångspunkt i teacher beliefs samt postkolonial teori och erfarenhetspedagogiska perspektiv. Resultatdelens har delats i de tre delarna: Minoritetsspråkens plats, Först fiktion och sen fakta samt Relevant för vem? 

This commentary poses the question: how do subjects recognise and evaluate what they are able tochange for themselves under algorithmic conditions that increasingly shape attention, desire, and anticipa-tion? Bringing Bernard Stiegler into dialogue with Ricœur, I consider what becomes of freedom in digitalsocieties if one is also to recognise Stiegler’s argument that human life is technologically constituted at itsgenesis. We were never voluntary, following Stiegler, since there was never a notion of voluntary humanchoice that could be bifurcated from involuntary acts of technologies from the start. This techno-geneticline of thought insists instead that freedom, needs, and desires at the level of the human subject concern atask of enacting a liberty to think differently given an originary human-technological entwinement.

Denna uppsats syftar till att utreda utmaningarna som eleverna i högstadiet möter i sina studier i biologi, samt hur eleverna upplever att läraren anpassar sin undervisning för att möta deras behov och väcka deras motivation att studera ämnet. I denna studie har kvalitativ metod använts, data samlades in genom semistrukturerade intervjuer med 12 elever från årskurs 7 till 9 från olika skolor i syfte att få fram olika perspektiv. Intervjufrågorna var relaterade till undervisningsmetoder och elevers upplevelser, motivation, engagemang och interaktion med lärare och elever. Tematiska analysmetoden användes för att analysera data. Resultaten presenterades genom olika teman bland annat påverkan av undervisningsstruktur på lärandet, resurser såsom andra personer och digitala verktyg, samt aktiviteter och konkreta föremål. Diskussionen kopplade resultaten till tidigare forskning och teorier. Slutsatsen som drogs av resultaten är att praktisk, visuell och varierad undervisning upplevs vara framgångsrik på förbättrad undervisning och ökad förståelse. Därmed har praktiska moment ökat motivationen till studier. 

This study aimed to understand how death becomes entertainment in relation to the conflict in Ukraine on Reddit’s subforum UkraineWarVideoReport, with the purpose of deepening understanding of how authentic representations of death are perceived and interpreted by users, as well as how these representations are mediated and aestheticized. The study collected data through shadowing netnographic observations on the subforum. The analysis employed an abductive approach, in which theories by Baudrillard (1994), Debord (2002), and Goffman (1959) were continuously consulted alongside the empirical material in accordance with a phenomenological methodology. The main findings show that mediated death carries entertainment value. The videos that were analyzed were often aestheticized in ways that encouraged or evoked entertainment value among viewers, while the comment sections were frequently characterized by themes such as humor, admiration, and fictional references, as well as moral concerns. The conclusions were that users often perceive and interpret the video material as entertainment, and that the videos themselves are often conveyed on an idealized, aestheticized, and fictionalized level; the interaction between comments and videos enabled a reinforcement of the inherent entertainment meaning within the context.

Background and aims: Aortic stenosis (AS) is the most common valve disease in Europe. It is associated with adverse myocardial remodelling and impaired exercise capacity, reflecting limited ability to increase cardiac output during physical exertion. Resting measures of AS do not fully capture functional limitations and despite practical recommendations, cardiopulmonary exercise testing (CPET) remains underused in this patient population. In patients with severe AS, important knowledge gaps remain regarding how circulating biomarkers reflect structural myocardial remodelling, how exercise capacity and systolic blood pressure (SBP) change before and after surgical aortic valve replacement (AVR), and how structured postoperative exercise training influences functional recovery. This thesis aimed to improve pre- and postoperative evaluation of patients with severe AS by exploring exercise physiology methods, focusing on CPET and circulating biomarkers.

Methods: Four studies were conducted within two prospective cohorts of patients with severe AS referred for AVR at a tertiary centre in Sweden.

In paper I, associations between circulating biomarkers and diffuse myocardial fibrosis assessed by myocardial biopsies and by imaging were explored.

In paper II, maximal and submaximal CPET variables were evaluated before and one year after AVR surgery.

In paper III, SBP responses during incremental CPET were analyzed before and after AVR, including categorization of SBP responses and work-load indexed SBP-slopes.

In paper IV, a randomized controlled trial assessed the effect of a 12-week aerobic exercise training intervention compared to usual care on CPET-derived functional outcomes following AVR.

Results: In Paper I (n=36), circulating biomarkers showed distinct associations with DMF assessed by cardiac magnetic resonance imaging (CMR) and histopathology. Biomarkers related to extracellular matrix turnover (Galectin-3, MMP-2) and myocardial stress (Troponin-I, NT-proBNP) were most strongly linked to CMR-derived measures (T1-time, extracellular volume), while NT-proBNP correlated with biopsy-derived collagen volume fraction. These findings indicate that biochemical markers reflect complementary aspects of myocardial remodelling.

In paper II (n=30), maximal exercise capacity was relatively preserved preoperatively (mean peak VO₂ ≈95% of predicted). One year after AVR, maximal workload increased modestly (+8%), peak VO₂ showed no significant group-level change, and submaximal aerobic capacity (VO₂ at ventilatory threshold) increased significantly (+9%). Improvements in peak VO₂ were heterogeneous and inversely related to baseline values. We found a strong significant correlation between submaximal oxygen uptake efficiency slope (OUES) and peak VO₂ (r = 0.89, p<0.05). 

In Paper III (n=45), an exercise-induced drop in SBP was uncommon preoperatively (<10%). Following AVR, SBP responses showed overall improvement, with no postoperative SBP drops observed. Among patients achieving maximal effort at preoperative and postoperative assessments (n=31), the SBP/Watt slope increased by approximately 48%.

In Paper IV (n=12), patients randomized to supervised aerobic exercise training demonstrated significant improvements in maximal workload (+28%) and peak VO₂ (+23%) after 12 weeks of intervention, whereas no significant changes were observed in the control group. Submaximal CPET variables showed non-significant trends toward improvement.

Conclusion: This thesis demonstrates that CPET-derived functional and hemodynamic responses provide clinically relevant information beyond traditional resting measures in severe AS. Circulating biomarkers were related to DMF. By capturing complementary aspects of myocardial remodelling and cardiopulmonary function before and after valve replacement, these methods support a more integrated approach to patient evaluation and postoperative management.

Detta examensarbete undersöker hur laddinfrastruktur för elektrifierade tungafordon bör dimensioneras vid byggarbetsplatser. Arbetet är baserat på verklig da-ta från Skanskas projekt E02 i Slakthusområdet och syftar till att ge insikter somkan användas som underlag för framtida tillämpningar av datadrivna beslut in-om verksamheten. Genom simuleringar i Python utvecklas en optimeringsmodellsom analyserar hur laddningsrutiner påverkar elbehovet och dimensioneringenav energilagring. Med hjälp av Monte Carlo-simuleringar genereras olika ladd-profiler som sedan integreras i optimeringsmodellen. Resultaten visualiseras iform av uppskattningar av hur energilagringsbehovet i en verksamhet varierarvid förändringar av antalet fordon och hur laddningsrutiner av dessa fordon kanse ut för optimera behovet av energilagring.Utifrån optimeringen kan slutsatsen dras att kapacitetskravet på batterilag-ring kan minska om laddningstider kan justeras. Studien visar att de metodersom företaget Skanska använder sig av idag för att dimensionera system för ladd-ning av elektrifierade fordon riskerar att leda till överdimensionering av kostsambatteriinfrastruktur. Arbetet föreslår istället ett helhetsperspektiv vid dimensio-nering av laddinfrastruktur, där laddningsmönster och effektprofiler optimerasför att uppnå ett mer kostnadseffektivt system. Sammanfattningsvis utvecklas idetta arbete förslag på metoder som kan användas vid analys av mätdata frånelektrifierade bygg- och anläggningsprojekt. Genom att kombinera verklig mät-data med Monte-Carlo simulerade värden genereras insikter i verksamhetensenergianvändning och hur laddinfrastruktur och rutiner för laddning kan opti-meras för att minimera behovet av kostsamma energilagringssystem.

Studiens syfte är att öka kunskapen om förskollärares perspektiv på hur muntligt berättande kan fungera som ett pedagogiskt redskap i förskolan. Studien undersöker därför förskollärares uppfattningar om hur barns agens, delaktighet och inflytande möjliggörs eller begränsas i berättande. Studien utgår från sociokulturell teori och barndomssociologi, där verbalt och icke-verbalt språk, samspel och förskollärares responsivitet ses som centrala för barns möjligheter att påverka sin vardag i förskolan. Data består av intervjuer med utbildade förskollärare verksamma i svenska förskolor. Studiens resultat visar att ett lyhört och följsamt förhållningssätt, responsiv pedagogik och kompetens i strategier för interaktivt muntligt berättande är betydande för att möjliggöra barns agens, delaktighet och inflytande i berättande. Förskollärarens medvetenhet om sin egen barnsyn är av vikt för att se och stötta barns agens. Samtidigt är en positiv inställning och självsäkerhet i den egna berättarförmågan avgörande för berättarformens användning som ett pedagogiskt verktyg i förskolan. Även kunskap om vad muntligt berättande är framstår som betydande för att öka förskollärares intresse och motivation att medvetet och pedagogiskt välja att använda muntligt berättande i förskolan. Vidare visar resultaten att muntligt berättande kan främja inkludering av alla barn, även barn med mindre utvecklat verbalt språk, när icke-verbalt språk och rekvisita används.

The study examines how sexual abuse in schools, committed by a teacher against a student, is discursively represented in two documentaries. Previous research shows that the media has significant power to shape the understanding of sexual violence, while the field of research is limited, which motivates further studies. The study focuses on factors portrayed as enabling the abuse, proposed solutions, and constructions of perpetrator and victims. The theoretical framework is based on feminist theory, social constructionism, and a Foucault-inspired toolkit, with discourse analysis as the method.

The findings identify several coexisting discourses: ignorance and irresponsibility, distrust and trivialization, and silence and loyalty, which collectively facilitate the abuse structurally. The perpetrators are portrayed both as anyone and as empathy-free monsters, while the victims are constructed as small and vulnerable but simultaneously as empowered. The study provides a more nuanced picture than previous research and can encourage critical reflection on media representations.

Sweden implemented a unique policy instrument over two consecutive five-year periods aimed at promoting industrial energy efficiency: a voluntary agreement programme (VAP) known as the Programme for Improving Energy Efficiency in Energy-Intensive Industries (PFE). This paper evaluates the second programme period using process and impact evaluation approaches. The impact evaluation revealed that the programme generated numerous benefits beyond the implementation of energy efficiency measures. The second programme period demonstrated impacts comparable to the first, underscoring the pivotal role of energy management within policy frameworks. Moreover, the findings suggest that VAPs may be among the few policy instruments capable of achieving substantial impact within industrial production processes. The process evaluation identified two critical gaps: a policy gap and a knowledge gap. Many of the positive elements present in the evaluated programme are currently absent from Sweden’s industrial policy mix. The role of the policy operator—as facilitator or intermediary—proved essential, particularly in the context of policies involving energy audits and certified energy management systems. In this case, the Swedish Energy Agency played a central role. The study concludes that continuous energy management, when integrated throughout the entire organization, significantly enhances companies’ capacity to identify and implement energy efficiency measures. This approach not only deepens internal knowledge but also fosters the development of more structured and effective routines. Furthermore, the findings suggest that financial incentives may be essential to ensure the long-term persistence of energy efficiency practices. Consequently, future policy design should include a strong facilitator and incorporate mechanisms that both reinforce organizational knowledge of energy end-use processes and promote sustained engagement over time, preferably involving a financial incentive.

The thesis employs theory alongside textual evidence to demonstrate how Go Set a Watchman impacts a nostalgic reading of To Kill a Mockigbird. This was accomplished by studying the characterization of Atticus Finch as he appears in the two novels through intertextuality. Similarities and differences have been found in comparison to prior research. The study finds that the nostalgic image of Atticus Finch is shattered in Go Set a Watchman, thereby negatively impacting a nostalgic read of To Kill a Watchman.

KLIMATFÖRÄNDRINGEN påverkar jord- och skogsbruket i Sverige på komplexa sätt. En högre medeltemperatur och längre växtsäsong skulle kunna innebära nya möjligheter för växtproduktion, men samtidigt finns klimatfaktorer som begränsar möjligheterna att utnyttja denna potential, såsom förändrade nederbördsmönster, frosthändelser och en ökad risk för extrema väderhändelser som värmeböljor, torka och skyfall. För att förstå, möta och hantera dessa utmaningar krävs underlag som bygger på relevant, användbar och användaranpassad klimatinformation.

Den här studien syftar till att undersöka hur lärare och speciallärare resonerar kring hur de arbetar med att förbereda eleverna i den anpassade gymnasieskolan för de sociala situationer de kommer möta i arbetslivet. Lärarna i studien arbetar med elever som är inskrivna i den anpassade gymnasieskolan varför den här studien utgår från specialpedagogiska dimensioner. En kvalitativ studie har genomförts där empiri samlats in genom semistrukturerade intervjuer med fem lärare eller speciallärare från fyra olika skolor där de arbetar i den anpassade gymnasieskolan.

De teoretiska perspektiv som använts i studien är det sociokulturella och specialpedagogikens relationella perspektiv där de didaktiska frågorna ligger till grund för frågeställningarna.

Resultatet visar att det på olika skolor arbetas på olika sätt med förberedelserna för arbetslivet. Studien visar att skolorna sällan arbetar systematiskt med förberedelserna för arbetslivet och att punktinsatser i samband med arbetsplatsförlagt lärande (APL) är den mest förekommande förberedelsen. Dessa förberedelser handlar om bland annat hygien, tid och hur man förväntas bete sig i ett fikarum. Samtliga informanter uttrycker att de sociala situationerna är en större utmaning för eleverna än arbetsuppgifterna i sig.

Education is the bedrock in building a strong, vibrant, and sustainable nation. However, the traditional classroom teaching approach remains the foundation of education in Liberia, and integrating outdoor learning presents challenges. This limitation makes it harder for students to learn and develop both intellectually and holistically. The study investigates the impact of outdoor education (OE) on students’ academic performance, social interaction, and personal development. It identifies barriers to the effective implementation of outdoor education, including opportunities for incorporating outdoor learning. Data were collected from six teachers using a qualitative research methodology and were analyzed thematically. Findings reveal that outdoor education strengthens students’ ability to work collaboratively, think independently, nurture creativity, regulate emotions, and achieve better learning outcomes. Yet, the teacher-centered approach, limited resources, and lack of structured support hinder its practice. Despite these challenges, teachers are willing to integrate outdoor learning. The study emphasizes the importance of integrating outdoor education into the national curriculum and strengthening teacher professional development in outdoor pedagogies to ensure its effective implementation.

I ett samhälle som utvecklas snabbt och ständigt pga information- och kommunikationsteknik och den ökade digitaliseringen, har skolan en viktig roll i att skapa digital kompetens och förbereda eleverna för deras framtida utmaningar i arbetslivet. Skolverket har förtydligat att digitalisering går som en röd tråd genom hela skolväsendet. Digitala verktyg inkluderas i alla kursplaner, i alla årskurser. Syftet med denna studie är att analysera hur användning av digitala verktyg ser ut i bildsalen utifrån elevernas perspektiv och i vilken mån bildundervisningen är i linje med Skolverkets ambitioner av digitalisering. 223 elever från åtta olika skolor har svarat på en enkätundersökning. Resultatet visar elevernas uppfattningar om digitalisering i bildämnet samt ger en överblick över vilken utrustning som finns tillgänglig i klassrummen, vilka digitala tekniker som verkligen används i undervisningen och hur stor del av bildundervisningen anser de har varit digital under deras år på högstadiet. Studien visar också vilken typ av bildprojekt där digitala verktyg används som eleverna brukar jobba med och lyfter upp vilka digitala tekniker som kanske inte ingår i undervisningen och som de skulle vilja jobba med på bildlektioner. Studiens utgångspunkt är en nationell ämnesutvärdering i bild som Skolverket anordnade 2013 och vars resultat jämförs med svaren givna i min enkät 2022. Resultatet 2022 visar att majoriteten av deltagarna har en positiv inställning till digitalisering, har anammat användningen av digitala verktyg i bildundervisningen och skulle vilja jobba i större utsträckning med digitala verktyg i ämnet. Många tycker att digitalisering ökar deras motivation och intresse i bildundervisningen. Elever som har en negativ inställning till digitalisering och som inte vill använda digitala tekniker i bildämnet är en klar minoritet.

The aim of this study is to investigate how study guidance teachers perceive that the organization of mother tongue study guidance affects their opportunities to support students language development. The study is based on semi structured interviws with eight study guidance teachers worknig across differnt educational levels, from compulsory scholl to upper secondary school, in various schools and with diverse linguistic backgronds. 

The results show that the organization of study guidance varies considerably between schools and that there are no established routines for how the work should be carried out in order to achieve its intended aims. insufficient organzation limits study guidanceteachers opportunities to work systematically with language development. the findings also indicate that the study guidance teachers competence in language developmented influences how language support is implemented and to what extent it contributes to students linguistic progression. The atudy demonstrats that when study guidance teacher has copetence in language development, most participants perceive that theguidance makes a positive difference for students learning and language development.

The study highlights the need for study guidance as well as targeted professional development initives for study guidance teachers.  

I samband med att den kyrkoantikvariska inventeringen av äldre boksamlingar i Linköpings stifts församlingar avslutades i december 2019, uppkom frågan hur de data som framtagits vid genomlysningarna bäst skulle tas tillvara. Projektet hade då pågått i två år, och under arbetets gång hade ett omfattande, och bitvis helt okänt material av kulturhistoriskt värde och med hög forskningspotential blottlagts.

Våren 2020 inleddes en uppföljning av projektet Kyrkornas böcker, som vid samma tid rönt viss medial uppmärksamhet, bland annat via ett reportage i Vetenskapsradion historia. Genom ett samarbete mellan projektet och Linköpings universitetsbibliotek (LiUB) skapades databasen Kyrkornas böcker som sjösattes i maj 2021 genom förlaget LiU Electronic Press.

Samma år inleddes ett nytt kyrkoantikvariskt bokinventeringsprojekt, denna gång i Växjö stift, där kompetens lånades in från Linköping. Därvid uppkom frågan om inte även data från denna inventering, och från andra stift som har genomfört eller planerar att genomföra liknande inventeringar, borde kunna inlemmas i den befintliga databasen och göras sökbara. Diskussionerna fortgick mellan LiUB och Linköpings stift, och 2023 beviljades medel från den kyrkoantikvariska ersättningen (KAE) till ett nationellt kunskapsutvecklingsprojekt om det kyrkliga kulturarvet, i syfte att bygga en databas som på sikt är tänkt att bli rikstäckande.

I projektet ingår, förutom LiUB och Linköpings stift, stiftsorganisationerna i Växjö, Skara, Strängnäs och Stockholm, som alla genomfört genomlysningar av kyrkornas boksamlingar i sina pastorat.

As conversational agents become increasingly prevalent in educational contexts, it remains unclear how text-based chatbots and embodied social robots differ in terms of user experience. This study addresses that precisely that, using a conversational agent called Bookbot, which is designed to have book discussions with 10- to 11- year-olds. Bookbot was applied to two interfaces, one text-based chatbot interface and the other a social robot interface called Furhat. Ten university students was recruited were they had one interaction with each interface. Employing a mixed-methods design, quantitative data were collected via the User Experience Questionnaire (UEQ) and qualitative insights through thematic analysis through semi-structured interviews. Although no statistically significant differences emerged on any UEQ scale, descriptive trends indicated superior performance of the chatbot on pragmatic dimensions (perspicuity, efficiency, dependability) and overall attractiveness, whereas the robot excelled on novelty and stimulation. Thematic analysis revealed that the chatbot’s clarity and predictable pacing fostered ease of comprehension and control, while the robot’s face and voice generated stronger social presence but suffered from turn-taking breakdowns and speech delivery issues. These findings highlight complementary strengths in the two interfaces: chatbots deliver functional stability and clarity, while social robots enable richer social engagement. Accordingly, design implications include integrating minimalist social cues into chatbots and enhancing the conversational robustness of social robots. Future research should employ issue‐based metrics and longitudinal social‐presence measures to distinguish novelty effects from sustained user engagement.

This study investigates the current and projected needs for heavy truck parking spaces within Norrköping municipality, emphasizing accessibility and safety. Norrköping’s strategic location along Sweden’s major highways, E4 and E22, positions it as a critical logistics hub. Through GIS-based spatial analysis, semi-structured interviews with logistics operators, and field observations, the study reveals a pronounced shortage of adequate parking facilities, particularly near industrial areas and the harbor. The lack of sufficient parking forces many trucks to park in unsuitable or illegal locations along roadways, increasing traffic hazards, accident risks, and negatively affecting the occupational environment for professional drivers. Furthermore, many existing parking sites suffer from insufficient lighting, lack of surveillance, and limited amenities such as restrooms and break areas, which contribute to increased insecurity and vulnerability to theft and vandalism. The study proposes two complementary solutions to address these challenges. The primary recommendation is the establishment of a centrally located logistics center designed to provide secure parking, comprehensive driver services, and future-proofed infrastructure for electric truck charging. The secondary recommendation involves developing smaller, decentralized parking facilities close to transport-intensive zones, such as industrial districts and the harbor, to relieve pressure on central parking areas and improve convenience for drivers. Overall, the findings underscore the importance of a well-planned parking infrastructure as a key component of an efficient and sustainable logistics chain. The study highlights the need for a coordinated, long-term strategy developed collaboratively by Norrköping municipality, logistics companies, and private stakeholders. Such an approach can strengthen the city’s appeal to businesses, enhance traffic safety, improve working conditions for drivers, and contribute significantly to sustainable urban development and transport electrification goals.

Neural rendering methods such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting have transformed novel view synthesis, yet their development relies heavily on standardized datasets for training and evaluation. This thesis addresses the dataset gap by developing a multi-camera capture system and collecting two neural rendering datasets: a studio object dataset and a large-scale outdoor heritage site reconstruction. We designed and implemented an automated capture system using twelve synchronized industrial cameras arranged on a quarter-hemisphere arc with a motorized turntable. The system captures 432 images per object across 36 rotation positions. Custom C++ software controls camera synchronization, turntable automation, and data organization. The studio dataset contains 15 objects spanning diverse materials including diffuse, glossy, transparent, and reflective surfaces. For large-scale capture, we documented Gränsö Castle in Sweden using 5,262 images from drone and SLR photography. This dataset tests neural rendering scalability on architectural heritage sites with varying lighting conditions. We compared traditional photogrammetry (RealityCapture) against neural rendering methods (Nerfacto, Splatfacto) on both datasets. Results show that Splatfacto outperforms Nerfacto with average PSNR of 33.27 dB versus 22.07 dB on studio objects. Neural methods successfully reconstruct challenging materials (glass, metal) where photogrammetry fails due to feature matching limitations on specular surfaces. For large-scale scenes, photogrammetry remains practical for complete geometric models while neural rendering excels in bounded high-quality visualization. Both datasets and the capture software are released as open-source resources for the research community.

Condensation trails are the white line-shaped clouds that can appear after aircraft as their exhaust mixes with the ambient air. In certain conditions, these can survive for hours and develop into cirrus-like clouds. These are believed to be one of the largest contributors to non-CO2 emissions from aviation, impacting global warming. They can have both a warming and a cooling effect, depending on the balance of inbound shortwave radiation and outbound longwave radiation. The net total is, however, a warming effect. Various mitigation strategies have been proposed. In this work, navigational avoidance by altitude adjustments is studied within the context of Swedish airspace. This is achieved by analysing historical flight trajectory data and meteorological data for the year 2024 using open-source contrail models, aircraft performance models, and aircraft trajectory optimisation models. Two contrail models — the Contrail Cirrus Prediction Model (CoCiP) and an algorithmic Climate Change Function (aCCF) from CLIMaCCF — are compared against each other. As expected, there are clear seasonal and diurnal effects, both regarding which altitudes that are contrail-prone, and which hours that see a net warming and net cooling effect. Geographical effects are more difficult to discern, especially due to the sharp decrease in air traffic north of 60°N. The relationship between navigational avoidance and other mitigation strategies is also discussed based on findings from related works.

ABSTRACT This report investigates technical solutions for reducing road traffic noise under Swedish conditions. Traffic noise is one of the most widespread environmental problems in urban areas, affecting millions of people and contributing to sleep disturbances, cardiovascular diseases, and reduced quality of life. The aim of this study is to evaluate the efficiency, durability, and adaptability of established noise reduction measures such as low-noise asphalt, noise barriers, and vegetation-based solutions. The research is based on a qualitative literature review combined with two case studies: Uppsala Municipality’s Noise Action Plan and the Swedish Transport Administration’s Action Programme for Road Traffic Noise 2024–2028. The results show that low-noise asphalt can reduce sound levels by approximately 3–5 dB(A), although its long-term performance decreases in cold climates due to clogging and wear. Noise barriers provide the most consistent reduction (5–10 dB(A)) but require space and maintenance. Vegetation-based solutions offer limited acoustic performance (2–5 dB(A)) but significant environmental and aesthetic benefits. The study concludes that combining technical and nature-based measures, adapted to local climate and traffic conditions, is the most effective strategy for long-term noise reduction. Integrated noise management that links engineering, urban planning, and environmental considerations is essential for achieving sustainable and quieter urban development in Sweden.

When many people choose to switch to electric cars, there are significant demands placed on the infrastructure to easily charge their vehicles wherever they are. One company contributing to this transition is CTEK. Their charging stations and Nanogrid systems help property owners, individuals, and others to provide a charging solution that works without overloading the power grid. To reduce excess work for the support department, a user interface was developed, allowing installers to configure and build a system themselves. The developed user interface focused on simplicity, scalability, and speed. Through an iterative, user-centered design process, which included a preliminary study, prototype development, implementation, and user testing, an interface with an interactive tree diagram was created. The preliminary study consisted of a literature review where theories on usability, UX design, interaction design, user-centered design, iterative design processes, impact mapping, and visualization methods were analyzed. A Lo-Fi prototype was developed and used in several user tests to ensure the new interface met the established expectations. Based on the user tests, an interface was developed. The implementation is a web application with an interactive tree diagram that allows users to add, edit, and remove charging stations and fuses in the configuration. The final product shows promising results, but parts of the interface require further development to handle error management and reduce repetitive tasks.

This candidate thesis aims to identify the rhetorical strategies Winston Churchill employs in his speech ‘We Shall Fight on the Beaches’, focusing specifically on how Churchill’s speech attempts to shape social reality, rather than measuring its actual effects on the audience. Drawing on Aristotle’s concepts of ethos, pathos, and logos, the study analyses how Churchill establishes credibility, regulates emotions, and frames resistance as the only rational and moral course of action. In addition, a metaphorical discourse analysis is applied to examine how metaphors such as “we” versus “the enemy” shape collective identity and social norms under crisis. 

The study further situates Churchill’s rhetoric within his literary background and rhetorical self awareness, drawing on his early essay The Scaffolding of Rhetoric to contextualize stylistic choices such as rhythm, repetition, accumulation of argument, and analogy. Influences from Shakespearean drama and historical narrative are shown to enhance the persuasive force and emotional resonance of the speech. By integrating classical rhetoric with theories of social reality, including perspectives from Boroditsky, Cassirer and Castoriadis the study demonstrates how political rhetoric functions not only as a tool of persuasion but as a mechanism for constructing social order, moral obligation, and collective resolve in times of crisis.

Denna litteraturstudie ämnar undersöka hur samhällskunskapslärare balanserar mellan objektivitet och fostransuppdraget vid undervisning om kontroversiella frågor. Studien utgår från ett antal vetenskapliga artiklar som på olika sätt behandlar lärares objektivitet i undervisningen. Artiklarna tar sin utgångspunkt i olika geografiska kontexter, vilket stärker bredden, samtidigt som det begränsar generaliserbarheten eftersom resultaten är beroende av olika samhälleliga och utbildningsmässiga förutsättningar. Artiklarna har jämförts tematiskt i syfte att identifiera skillnader och likheter angående vilka strategier som förekommer vad gäller objektivitet och fostran. Studien kommer fram till att samhällskunskapslärare ständigt balanserar mellan kraven på att vara opartiska och samtidigt fostra demokratiska samhällsmedborgare. Den visar också på att en vanlig tendens hos flertalet lärare är att distansera sig från undervisningen, vilket leder till att ansvaret för att säkerställa en optimal undervisningssituation överlåts till eleverna själva. Studien indikerar också på att det saknas tydliga riktlinjer för hur lärare bör agera vid undervisning om kontroversiella frågor och att det därför skapas en stor variation mellan svenska klassrum.