Publikationer från Uppsala Universitet
Not a member yet
127603 research outputs found
Sort by
Planting Crops and Co-creating Worlds : Garden practices in the Basque Country
No full textThis study examines the social relationships and entanglement taking form between different species where gardening and other agricultural practices take place. A fieldwork of five and a half weeks was conducted in the Basque Country, at an urban vegetable garden and a rural vegetable farm. The research is based on participant observation taking place in the vegetable fields, three recorded interviews, and fieldnotes. The aim of the thesis is to show how life is constituted through entanglements between humans, plants, animals, and soil when growing vegetable crops. Drawing on Tim Ingold’s concepts of dwelling and sentient ecology, the research shows how working and spending time with other beings in gardens creates embodied knowledge, caring relationships, and well-being
Hierarchical Control and Inverter Functionalities in Microgrids : Modeling, Simulation, and Case Studies
No full textAs the impacts of climate change intensify, the urgency of achieving carbon neutrality has never been greater. To achieve this goal, many countries are turning to distributed energy resources to reduce their reliance on fossil fuels and accelerate the transition to renewable energy sources. A promising approach is the microgrid, a small-scale, self-sufficient energy system that can operate isolated or in connection with the main grid. With autonomous control of voltage, frequency, and power flow, microgrids enhance reliability and resilience, support faster recovery during disturbances, and improve power quality. Inverters are central to microgrid operation, as they enable the integration of diverse energy sources while supporting grid stability. They can be classified into three categories: grid-feeding, grid-supporting, and grid-forming. Each fulfills a distinct function, from injecting power into the grid to providing stability during disturbances and even enabling independent islanded operation. Together, these functionalities expand the adaptability and reliability of microgrids across a wide range of scenarios. The control architecture of microgrids is commonly structured in hierarchical levels: inner, primary, secondary, and tertiary. At the inner level, fast loops regulate current and voltage, ensuring basic power quality. Primary control, typically implemented through droop strategies, allows proportional power sharing among distributed generators without communication. Secondary control restores nominal voltage and frequency, while tertiary control manages power exchange and optimization with the main grid. Building on this framework, this thesis examines inverter functionalities and hierarchical control strategies across different microgrid configurations. The analysis combines theoretical discussion with practical evaluation through representative case studies. The outcomes were developed and validated using MATLAB/Simulink and PSCAD. The discussion is organized according to the hierarchical control levels, starting with the inner control level and progressing towards the secondary level. The analysis at the levels includes all inverter operation modes, covering cases where inverters operate independently in a single mode, in dual modes, and with a configuration that combines up to three operational modes, illustrating the progressive enhancement of control autonomy and system support capabilities. This approach provides a comprehensive view of how inverter functionalities and control hierarchies interact, highlighting their impact on system performance, adaptability, and stability under different operating conditions
Explaining Expected Threat inSoccer Passes: AnInterpretable Deep Learning Approach with TabNet
No full textMachine learning has become an integral part of sports analytics, offering new possibilities fortactical decision-making and performance evaluation. In football analytics, explainable modelsare especially valuable for helping coaches, scouts, and analysts interpret model outputs in clear,football-specific language. This thesis develops models to predict the expected threat (xT) ofpasses by combining event and tracking data from the Swedish Allsvenskan. Two complementarymachine learning approaches are explored: TabNet, applied to a handcrafted feature dataset of19 variables, and a Convolutional Neural Network (CNN) trained on positional player–matrixrepresentations.The findings highlight a trade-off between interpretability and predictive performance. TabNetprovides strong transparency through feature attributions and Integrated Gradients analysis, whileCNNs achieve higher accuracy by leveraging spatial context. To enhance accessibility, numericaloutputs from TabNet are further translated into natural-language explanations using awordalisation framework powered by Large Language Models (LLMs), enabling footballpractitioners to derive actionable insights and facilitating communication between data scientistsand non-technical experts.By integrating interpretability, predictive power, and communicability, this thesis contributes alayered framework for advancing football analytics, demonstrating how explainable AI can bridgethe gap between complex statistical modelling and real-world practice. Keywords: Sports analytics, TabNet, Convolutional Neural Networks (CNN), Explainable AI,Integrated Gradients, Large Language Models (LLMs
Pythia - Legal analysis with AIefficiency
No full textPythia, a web application designed to enhance document management for legal entities. Navigating through legal documents traditionally involves manual processes, such as keyword searches or skimming through pages of text, which are time-consuming and prone to human error. Pythia addresses this issue through an AI-powered tool, tailored specifically for extracting information from large legal documents. Through Pythia’s development, insights have been made about the potential opportunities and challenges of using AI-tools for processing large Swedish legal documents
A mixed methods study of self-directed learning in clinical practice using a mobile skills training system
No full textBackground The challenge of increasing demands on healthcare and a less available workforce requires new approaches to retain and develop healthcare professionals. One way of addressing this challenge is by instituting competency-based continuing professional development based on self-directed learning (SDL) principles. We investigated the feasibility of a mobile training system in a clinical and peer-to-peer, SDL setting. Methods The study used mixed methods with surveys, observations, and interviews to gain comprehensive insights into how healthcare professionals (nurse, assistant nurse) and managers experienced the investigated educational tool. Results Healthcare professionals' experiences of SDL were illustrated in four main themes: (1) leadership required for learning, (2) conditions for learning, (3) effects of continuing professional development, and (4) suggestions for improving training methods. In surveys, healthcare professionals self-assessed their confidence in executing care tasks before and after undergoing SDL for those tasks. The results indicated a significant increase in self-efficacy post-SDL training. Engaging in care tasks through SDL, which includes checklists aligned with evidence-based practices, enhances healthcare professionals' awareness of their own competence. Additionally, it provides managers with the means to ensure that employees meet core competence requirements. The role of a clinical skills centre to support managers and staff in their learning efforts was seen as important for sustained learning. Conclusion A mobile training system, implemented within a clinical and peer-to-peer SDL framework, facilitates and complements competency-based continuing professional development and reduces the perceived length of time healthcare professionals spend away from direct patient care. Access to clinical educators for guidance during implementation, along with support from the clinical skills centre during the planning and follow-up stages, proved essential for facilitating accessible learning. Managers play a crucial role in prioritizing continuous learning for their employees and establishing a culture of ongoing education among managers and healthcare professionals
ANISSA : Advanced Neutron Imaging for Solid-State batteries in Action
No full textThe development of high-energy density solid-state batteries is critical for the achievement of carbon neutrality goals and the advancement of clean energy. Still, the fundamental understanding of lithium transport mechanisms and degradation processes remains limited. Current characterisation methods face significant challenges in studying these complex systems, particularly due to the difficulty of detecting lithium dynamics in three-dimensional battery architectures in operando conditions. Here we present the ANISSA (Advanced Neutron Imaging for Solid-State batteries in Action) project, an integrated experimental framework combining high-resolution neutron and X-ray imaging techniques to research coupled electro-chemo-mechanical processes in lithium-based energy storage systems
A field theory approach to the critical study of global university rankings
No full textGlobal university rankings have gained immense popularity over the past two decades. Today, most actors in higher education seem to believe or at least acquiesce in the evaluations published by these benchmarking devices. At the same time, global rankings have become target of numerous critiques ranging from their underlying logic and methods to the material and symbolic reconfigurations of institutions they can be related to. My purpose in this chapter is twofold. First, by conducting an overview of the main critical literature on the topic, I will show how existing critiques of global university rankings, despite bearing relevant similarities, are characterised by a lack of a unifying theoretical framework. Then, drawing on the sociological theory of Pierre Bourdieu, I will argue for and develop an implementation of a field theory approach to connect and expand on those critiques. The adoption of this perspective will enable the uncovering of the true function served by global university rankings beyond the mere provision of performance metrics in higher education
The Quieting Storm : Longitudinal Analysis of CSF Biomarkers Following Autologous Hematopoietic Stem Cell Transplantation for Multiple Sclerosis
No full textMultiple sclerosis (MS) is a chronic inflammatory and neurodegenerative disease of the central nervous system, driven by immune-mediated damage to myelin and axons. Although several disease-modifying therapies (DMTs) can reduce inflammatory activity, some patients continue to experience relapses and accumulating disability. For these individuals, autologous hematopoietic stem cell transplantation (aHSCT) represents a potent therapeutic option capable of inducing long-term remission. The studies included in this thesis aimed to elucidate how aHSCT influences key molecular processes in MS through cerebrospinal fluid (CSF) biomarker analysis. Paper I investigated circulating cell-free mitochondrial DNA (ccf-mtDNA) as a marker of inflammation in MS. Concentrations were elevated in people with MS (pwMS) compared with healthy controls and decreased markedly after intervention with aHSCT, indicating normalization of inflammatory activity. Paper II examined microRNAs (miRNAs) biomarkers of inflammatory disease activity. Twelve dysregulated miRNAs were identified, of which a recurring four-miRNA cluster (miR-16-5p, miR-21-5p, miR-146a-5p, miR-150-5p) showed strong associations with disease activity and normalized after aHSCT, reflecting a rebalanced immune milieu. Paper III focused on the blood–brain barrier (BBB), assessing the MMP-9/TIMP-1 ratio and osteopontin (OPN) as markers of barrier integrity. Both were elevated before treatment and declined significantly after aHSCT, suggesting restoration of structural and functional BBB stability. Paper IV explored biomarkers associated with progressive disease; galectin-9 (Gal-9), growth differentiation factor (GDF-)15, and chitinase 3-like protein 1 (CHI3L1, YKL-40); all of which decreased following aHSCT, indicating reduced glial activation and mitigation of progression-related mechanisms. Collectively, these studies demonstrate that aHSCT induces profound and sustained molecular remission across inflammatory, structural, and degenerative domains of MS. The findings highlight the value of CSF biomarkers in understanding treatment mechanisms and underscore the need for integrative biomarker strategies to guide optimal timing and patient selection for aHSCT.
Optimizing implementability when developing complex interventions : Guided low-intensity behavioral activation for people with dementia
No full textThis thesis comprises three studies with the overall aim to develop and adapt a guided low-intensity behavioral activation intervention (INVOLVERA) for people with dementia and depression, while optimizing future implementability for the Swedish context. Study I was a qualitative study. Interviews and focus groups were held with healthcare professionals and community interest-holders to: (1) develop an understanding of the existing healthcare and community support in the Swedish context for people with dementia and informal caregivers (caregivers) from the perspectives of healthcare professionals and community interest-holders, and (2) identify healthcare professionals and community interest-holders’ perceived barriers and facilitators to intervention uptake informed by Normalization Process Theory (NPT). Data analysis was informed by NPT and generated ten deductive and three inductive categories related to three NPT core constructs: Coherence, Cognitive participation, and Collective action. Study II was a qualitative study. Interviews were held with people with dementia and caregivers to: (1) develop an understanding of the existing healthcare and community support in the Swedish context for people with dementia and caregivers from the perspectives of people with dementia and caregivers, and (2) identify people with dementia and caregivers’ perceived barriers and facilitators to intervention uptake informed by NPT. Data analysis was informed by NPT and generated nine deductive and three inductive categories related to three NPT core constructs: Coherence, Cognitive participation, and Collective action. Study III was a mixed-methods study aiming to explore public contributors and researchers’: (1) experiences of public contribution activities, (2) perceived impact of public contribution activities on the research and themselves, (3) perspectives on the process of public contribution, including influence on decision-making. Furthermore, with respect to public contributors, to explore: (4) ideas for the future development and expansion of the Public Advisory Group. Public contribution activities and recommendations were recorded using impact logs. Recommendations were extracted and categorized, and implementation rates calculated. Interviews were held with public contributors and research team members, and analyzed using manifest content analysis. Public contributors made 158 recommendations, of which 76% were implemented. Analysis of interviews generated three categories: Perceived impacts, Interactions and facilitators, and Future challenges and opportunities. This thesis contributes to wider literature on how to consider the future implementability of complex interventions during the intervention development phase
From disciplinary enthusiasm to soulless tasks : norms behind computing educators' emotion display
No full textComputing educators' emotion display is regulated by various norms and conventions. These (unwritten) rules affect which emotions educators feel comfortable expressing to different extents when students can perceive such displays. We draw on emotions and norms literature to investigate higher education computing educators' perceptions of what kinds of norms they have as criteria when considering which emotions they are (not) comfortable showing when teaching. Based on the qualitative content analysis of 22 interviews with computing educators from seven countries, we present various – sometimes connected or conflicting – norms that influence emotion display. These norms comprise moral, national, societal, professional, and affiliation norms that are salient in different ways for different educators. The findings contribute to a more nuanced understanding of various combinations of nested, context-dependent, and partly conflicting norms that guide educators' emotion display, an understanding of educators' work that goes beyond cognitive aspects