Digitala Vetenskapliga Arkivet - Academic Archive On-line
Not a member yet
734120 research outputs found
Sort by
Sustainable solid-state polymer electrolyte based on PEO-Xanthan gum blend for enhanced lithium-metal batteries
No full textDespite their widespread adoption, LIBs are still facing several challenges, mainly related with safety concerns of conventional electrolytes that are currently limiting their practical use. Solid-state polymer electrolytes (SPEs) represent a promising alternative to produce safer devices, as they offer higher flexibility and energy density, easy processability, non-flammability and improved mechanical strength, even if they often suffer from low ionic conductivity at room temperature. To address the latter issue, the development of novel SPEs based on a blend of polyethylene oxide (PEO) and Xanthan gum (XG), a natural polysaccharide with notable mechanical and rheological properties, is proposed. In this study, the thermal, physical, and electrochemical properties of the PEO-XG blends were investigated, aiming to assess their potential as electrolyte for all solid-state lithium metal batteries. Improved ionic conductivity, electrochemical stability window and cycling stability are achieved, confirming the effectiveness of XG incorporation. The most promising electrolyte formulation was studied using self-diffusion 7Li pulse-field-gradient (PFG) NMR measurements and different temperatures and further evaluated in full-cell configurations employing two olivine-type cathodes (LFP and LMFP). When paired with a lithium manganese iron phosphate (LMFP) cathode and cycled over an extended voltage window of 2.5–4.5 V, the cell demonstrates high specific capacity and excellent capacity retention, maintaining stable performance for at least 750 cycles.Validerad;2025;Nivå 2;2025-11-25 (u5);Full text license: CC BY 4.0;</p
Knowledge ecosystem emergence: Organizing participation, identity, and actorhood to bridge the governance void
No full textMulti-organizational collaborations involving the industry, academia, and government have become prevalent in developing knowledge to address complex societal problems. These fluid and loosely coupled forms of collaboration, known as knowledge ecosystems, provide the necessary organizing elements for the search and creation of new knowledge. While the literature acknowledges the prevalence of knowledge ecosystems, it remains relatively silent on how their organization develops over time. This lacuna in our understanding is problematic, given the challenge of governance voids for cross-sectoral knowledge collaborations, which lead to difficulties in mobilizing action, securing resources, and ultimately achieving the knowledge-related goals of these collaborations. To address this gap, we theorize knowledge ecosystems as meta-organizations, examining how they gradually develop organizing elements that bridge the initial governance void. Empirically, we draw on an in-depth field case study of the High-Capacity Transport ecosystem in Sweden, demonstrating how three interrelated organizational elements—participation, identity, and actorhood—emerge through an iterative, yet broadly sequential process to resolve governance void challenges in resourcing, belonging, and collective action. Furthermore, we identify discursive and performative meta-organizational practices that enable the actors to collectively enact the aforementioned organizational elements and to engage in knowledge search. We further demonstrate how the organization of knowledge ecosystems is never ‘complete’ and the governance void remains only partially resolved, given the underdefined nature of new knowledge as the ecosystem's shared goal. Overall, our process model contributes to the theory, practice, and policy of knowledge ecosystem emergence and organizing. Validerad;2025;Nivå 2;2025-11-25 (u2);Full text: CC BY license;</p
Few-shot learning and explainable AI for colon cancer histopathology : A prototypical network with multi-technique interpretability
No full textBackground: Colon cancer diagnosis from histopathology is challenging due to limited annotated data and the lack of interpretability in deep models. Objective: We present a data-efficient framework combining few-shot learning and explainable AI for accurate and transparent diagnosis. Methods: A Prototypical Network with a ConvNeXt-Tiny backbone was trained on small colon-tissue image sets. Explanations from Grad-CAM and LIME were validated by a pathologist, and generalization was tested on an external dataset. Results: The model achieved 98.5 % accuracy in-domain and 90 % on the EBHI dataset, showing strong generalization. Conclusions: This few-shot and explainable model performs well with minimal data and generates clinically interpretable visual outputs, supporting its potential for reliable colon cancer diagnosis.
Percutaneous and surgical management of aortic stenosis in the SWEDEHEART registry (2013–2023) : a nationwide observational study
No full textBackground: Management of severe aortic stenosis (AS) has evolved over the past decade, driven by the widespread adoption of transcatheter aortic valve implantation (TAVI). This study aims to assess trends in procedural volumes, patient characteristics, and outcomes for patients undergoing TAVI or surgical aortic valve replacement (SAVR) in Sweden. Methods: This was a descriptive, non-comparative, nationwide cohort study using the SWEDEHEART registry. We included 21,383 patients who underwent TAVI or SAVR between 2013 and 2023 (11,366 TAVI and 10,017 SAVR). Trends in patient characteristics, preoperative risk, complications and mortality were examined. Findings: TAVI procedures increased from 307 (26.1%, n = 307/1174) in 2013 to 1851 (71.2%, n = 1851/2601) in 2023, while SAVR volumes declined from ∼1000 annually before 2018 to roughly 750 procedures annually. Median age of TAVI patients were 81 (IQR 77, 85) years and 71 (IQR 65, 76) years for SAVR patients. The median EuroSCORE II for TAVI decreased from 5.6 (IQR 3.3, 10.2) to 2.7 (IQR 1.7, 4.6) (p = 0.002), and STS-PROM from 3.3 (IQR 1.9, 4.1) to 1.6 (IQR 1.1, 2.8) (p = 0.0021). Among SAVR patients, EuroSCORE II decreased from 1.5 (IQR 1.0, 2.3) to 1.3 (IQR 0.9, 2.1) (p = 0.022) and STS-PROM from 1.8 (IQR 1.2, 3.0) to 1.6 (IQR 1.1, 2.6) (p = 0.0082). Any in-hospital complications declined significantly for TAVI (29.2%, n = 210/719 to 13.2%, n = 244/1851), while SAVR complication rates increased slightly (18.4%, n = 354/1921 to 18.7%, n = 140/750). In-hospital mortality for TAVI declined from 3.6% (n = 26/719) to 1.0% (n = 18/1851), and 1-year mortality from 11.1% to 6.9% (p = 0.019). SAVR in-hospital all-cause death decreased from 1.6% to 0.4% (n = 3/750) and 5.0% to 2.2% for 1-year mortality (p = 0.013). Interpretation: TAVI has become the predominant treatment strategy for AS in Sweden expanding access within the treated cohort. Despite this, current 2023 SAVR results demonstrate similar in-hospital complication rates compared to TAVI (18.7% vs 13.2%), but lower in-hospital (0.4% vs 1.0%) and 1-year mortality rates (2.2% vs 6.9%). Funding: This study was supported by ALF and national research funding bodies
Short rainbow cycles for families of small edge sets
No full textIn 2019, Aharoni proposed a conjecture generalizing the Caccetta-Häggkvist conjecture: if an n -vertex graph G admits an edge coloring (not necessarily proper) with n colors such that each color class has size at least r , then G contains a rainbow cycle of length at most ⌈ n / r ⌉ . Recent works (Aharoni and Guo, 2023 [1] ; Aharoni et al., 2023 [3] ; Guo, 2025 [8] ) have shown that if a constant fraction of the color classes are non-star, then the rainbow girth is O ( log n ) . In this note, we extend these results, and show that even a small fraction of non-star color classes suffices to ensure logarithmic rainbow girth. We also prove that the logarithmic bound is of the right order of magnitude. Moreover, we determine the threshold fraction between the types of color classes at which the rainbow girth transitions from linear to logarithmic
Designing Synthetic Active Learning for model refinement in manufacturing parts detection
No full textThis paper introduces Synthetic Active Learning (SAL), a fully automatic model refinement strategy for manufacturing parts detection using only synthetic data actively generated with domain randomization for training. SAL iteratively updates the detection model by identifying its weaknesses, such as in specific categories, materials, or object sizes, using custom evaluators, and generating targeted synthetic data to address them; it selectively synthesizes new useful data with respect to active learning, where traditionally humans in the loop select data to label. During each iteration, model training and data generation occur simultaneously to improve efficiency. Evaluated on four use cases from two industrial datasets, SAL achieved mAP@50 improvements of 2 to 6% percentage points over static learning, which refers to training on a fixed, pre-generated dataset. It also showed notable gains in underperforming categories, leading to more balanced performance across classes. Another benefit is that it uses a consistent configuration across multiple use cases, avoiding the need for extensive hyperparameter tuning common in prior domain randomization studies. Given its encouraging performance across diverse scenarios, we believe that SAL can scale to broader industrial applications where training can be fully or mostly based on synthetic data. CC BY 4.0Correspondence Address: X. Zhu; KTH Royal Institute of Technology, Stockholm, Sweden; email: [email protected]; CODEN: JMSYEThis work was partially supported by the Wallenberg AI, Autonomous Systems and Software Program (WASP) funded by the Knut and Alice Wallenberg Foundation, Sweden. The computations were enabled by the Berzelius resource provided by the Knut and Alice Wallenberg Foundation at the National Supercomputer Centre. We gratefully acknowledge colleagues at the Production Oskarshamn, Production Zwolle, Transmission Assembly, Engine Assembly, Academy, and Smart Factory Lab Departments at Scania CV AB for providing the CAD models and use cases. We also extend our thanks to Prof. Joakim Lindblad at the Department of Information Technology, Uppsala University, for his valuable insights and constructive feedback on this study.</p
Self-Management Therapies for Temporomandibular Disorders : Evidence From Systematic Reviews
No full textObjectives: Chronic pain in the orofacial region affects 10%–15% of the general population. It is most often related to temporomandibular disorders (TMD): pain in the masticatory muscles and the temporomandibular joints. Managing chronic pain conditions requires a biopsychosocial, evidence‐based and patient‐centred approach. Accordingly, self‐management has been advocated for TMD. This umbrella review aimed to evaluate existing evidence in systematic reviews (SRs) on self‐management therapy for TMD. Methods: The review was carried out in accordance with the PRISMA and PRIOR guidelines, with a pre‐registered protocol (PROSPERO CRD42021276856). PubMed, Scopus, LILACS and the Cochrane Library were searched until December 1st 2023 to identify SRs that evaluated self‐management for TMD. Two independent reviewers screened titles and abstracts, performed full‐text assessments, extracted all data and performed risk of bias assessment with AMSTAR 2. Results: A total of 1740 studies were identified. After title and abstract screening, 399 studies underwent full‐text assessment and 11 SRs comprising 49 unique primary studies were included. Of these, six SRs reported favorable results for self‐management for TMD, whereas five reported insufficient evidence either for or against the use of self‐management compared to other interventions. The overlap of primary studies between the SRs was 53%, and the main evidence gaps were related to quality of life and adverse effects outcomes. The methodological shortcomings of the SRs primarily stem from insufficient primary‐study design criteria or not providing references for excluded studies. Conclusions: Existing evidence generally suggests beneficial effects from self‐management strategies such as patient education, behavioural therapy and jaw exercises
Measured service reliability and customer satisfaction in public transport
No full textHow travellers perceive the quality of public transport services has profound influence on their satisfaction and thus of their inclination to choose this mode of transport. Punctuality, which may be interpreted as timetable adherence or overall travel time predictability, is often pointed out as the most important constituent of overall service reliability of public transport systems. However, in order for operators to be able to improve passengers’ perceptions of service reliability, an empirically derived linkage between perceptions and actual service performance indicators must be established. This paper has addressed this issue by using six-year longitudinal dataset on half-year resolution level from the regional public transport system of Scania, Sweden, to relate various indices of actual service reliability, obtained from service performance indicators, to stated satisfaction with punctuality for trains and buses. This was done by means of linear regression models using a stepwise approach to eliminate redundant variables, based on the level of independence of each variable. The results indicate that average arrival delay and frequency of early departures in relation to schedule are the most important factors of service reliability to explain travellerś perceptions of punctuality. Thus, for instance, average arrival delay and schedule adherence are indicated to have three times larger impact on satisfaction than frequency of early departures. But we also find that other service quality factors such as cleanliness, safety, and seat availability are also indicated to have an impact on perceptions of service reliability. It is emphasized that appropriate service performance data, and their relation to perceptions of passengers, constitutes crucial information for commercial and public providers in their ambition to improve levels of passenger satisfaction with operations but that the optimal level of punctuality is likely to differ between different passenger categories, and thus services. Research funding provided by K2 - Swedish Knowledge Center for Collective Mobility.</p
Results from a Swedish model-based analysis of the cost-effectiveness of AI-assisted digital mammography
No full textObjectiveTo evaluate the cost-effectiveness of AI-assisted digital mammography (AI-DM) compared to conventional biennial breast cancer digital mammography screening (cDM) with double reading of screening mammograms, and to investigate the change in cost-effectiveness based on four different sub-strategies of AI-DM.Materials and methodsA decision-analytic state-transition Markov model was used to analyse the decision of whether to use cDM or AI-DM in breast cancer screening. In this Markov model, one-year cycles were used, and the analysis was performed from a healthcare perspective with a lifetime horizon. In the model, we analysed 1000 hypothetical individuals attending mammography screenings assessed with AI-DM compared with 1000 hypothetical individuals assessed with cDM.ResultsThe total costs, including both screening-related costs and breast cancer-related costs, were &lt;euro&gt;3,468,967 and &lt;euro&gt;3,528,288 for AI-DM and cDM, respectively. AI-DM resulted in a cost saving of &lt;euro&gt;59,320 compared to cDM. Per 1000 individuals, AI-DM gained 10.8 quality-adjusted life years (QALYs) compared to cDM. Gained QALYs at a lower cost means that the AI-DM screening strategy was dominant compared to cDM. Break-even occurred at the second screening at age 42 years.ConclusionThis analysis showed that AI-assisted mammography for biennial breast cancer screening in a Swedish population of women aged 40-74 years is a cost-saving strategy compared to a conventional strategy using double human screen reading. Further clinical studies are needed, as scenario analyses showed that other strategies, more dependent on AI, are also cost-saving.Key PointsQuestionTo evaluate the cost-effectiveness of AI-DM in comparison to conventional biennial breast cDM screening.FindingsAI-DM is cost-effective, and the break-even point occurred at the second screening at age 42 years.Clinical relevanceThe implementation of AI is clearly cost-effective as it reduces the total cost for the healthcare system and simultaneously results in a gain in QALYs.Key PointsQuestionTo evaluate the cost-effectiveness of AI-DM in comparison to conventional biennial breast cDM screening.FindingsAI-DM is cost-effective, and the break-even point occurred at the second screening at age 42 years.Clinical relevanceThe implementation of AI is clearly cost-effective as it reduces the total cost for the healthcare system and simultaneously results in a gain in QALYs.Key PointsQuestionTo evaluate the cost-effectiveness of AI-DM in comparison to conventional biennial breast cDM screening.FindingsAI-DM is cost-effective, and the break-even point occurred at the second screening at age 42 years.Clinical relevanceThe implementation of AI is clearly cost-effective as it reduces the total cost for the healthcare system and simultaneously results in a gain in QALYs.Funding Agencies|Region Ostergotland and Linkoping University; Linkoping University</p
Propagation of Opinions on Signed Graphs : Stability and Collective Wisdom
No full textThis thesis investigates how collective opinions evolve in social networks, focusing on two crucial, real-world factors: individual stubbornness and antagonistic inter-actions. Classical models of opinion dynamics typically assume a cooperative environment, where all individuals collaborate to reach an agreement, causing opinions to naturally draw closer to each other. However, many real-world scenarios—from political debates to online discussions—are defined by stubbornness, where individuals resist changing their stance, and by rivalry or distrust, which can actively push opinions apart. We address these phenomena through the framework of signed networks, which explicitly encodes antagonistic relationships. Unlike models that rely on structural balance—a clean division into two opposing camps—we adopt a repelling interaction mechanism, treating negative ties as additive repulsive forces. This approach offers a more flexible and realistic representation of complex, unbalanced social structures. The first contribution is a thorough analysis of the signed Friedkin–Johnsen (SFJ) model, which combines stubborn attachment to initial opinions with antagonistic interactions. This combination fundamentally changes system behavior: while the classical FJ model is always stable, the SFJ model can diverge. Therefore, we establish sufficient stability conditions and show that even when the model is stable, antagonism allows opinions to escape the the convex hull of the initial opinions, meaning that agents can adopt positions more extreme than anyone held at the start. The thesis then examines repeated discussions through a concatenated SFJ model, reflecting scenarios such as a series of negotiations. Here, stability in an individual discussion does not imply stability for the repeated process. Two distinct behaviors emerge when the individual discussion rounds are stable: opinions may temporarily move apart before reconverging to consensus inside the convex hull, or they may drift farther apart with each round, leading to divergence. The first case corresponds to a transient amplification, which is a known consequence of non-normal matrices and is often referred to as reactivity. Given these possibilities, we provide sufficient stability conditions for the concatenated SFJ model. This analysis is extended to explore the dynamics in more complex scenarios, including time-varying interaction structures and bipartite (two-camp) networks. Next, we address multidimensional opinion dynamics models, where agents debate several interdependent topics (for example, climate policy is inherently linked to economic policy). We show that a set of topics that would be stable if discussed independently can become unstable and diverge when discussed together. The analysis of repeated (concatenated) multidimensional interactions reveal diverse outcomes, from full or partial or bipartite consensus to divergence. Finally, we shift the focus from opinion convergence to the quality of resulting collective judgement and study the wisdom of crowds problem, i.e., analyze the conditions under which a group’s aggregated opinion becomes more accurate due to collective interactions. The improvement of wisdom depends entirely on the allocation of the social power vector—the centrality measure that each individual has in the group. In classical cooperative systems, the social power vector is positive, which confines the concentration region (the set of allocations that improve wisdom) to a simplex. In contrast, signed networks allow for negative social powers, which expands the concentration region beyond the simplex, to a hyperplane. The analysis on signed networks also leads to a critical insight: agreement does not imply accuracy. The group can become confidently wrong—converging with high certainty to a false truth. We also analyze the case where the agents’ initial opinions are correlated, and characterize the new concentration regions and their properties under these conditions