Publikationer från Uppsala Universitet
Not a member yet
127603 research outputs found
Sort by
Virtual sensing of temperature in a Swedish office building : Utilizing machine learning and statistical analysis
No full textClimate change mitigation demands improved energy efficiency across the building sector, where smart buildings play a pivotal role. However, widespread deployment of the underlying infrastructure to create smart buildings is expensive and complex to manage, namely, sensor installation and management. This study explores virtual sensing as a scalable alternative, aiming to reduce the required amount of sensor by leveraging machine learning models for virtual sensing. Utilized data is provided from a floor with multiple rooms in a real office building located in Sweden. This study proposes a data-driven methodology that identifies which predictor sensors are most informative for virtual sensor creation for any given target sensor. The most informative predictor sensors were determined using both a correlation analysis approach and a compute heavy machine learning approach. The model types included in this study are Linear Regression (LR), Long Short-Term Memory (LSTM) and Online Sequential-Extreme Learning Machine (OS-ELM). Subsequently, each model type was configured with memory capacity of 0-12 hours to identify if latent information in the data could be extracted. These model types were trained on one heating period and validated on the subsequent heating period to ensure temporal robustness while simulating an implementation of the virtual sensor. Results show that the non-linear model types with memory capacity, achieved the lowest Root Mean Square Error (RMSE), with 0.332 ◦C for the LSTM and 0.356 ◦C for the OS-ELM, outperforming best linear model configuration with 0.446 ◦C. Furthermore, the linear model was unable to utilize latent information in the data and performed best with low or no memory capacity. Notably, both correlation and machine learning approaches identified the single most informative sensor. However, correlation proved insufficient for selecting optimal predictor sensor combinations while the machine learning approach identified optimal predictor combinations. This study’s findings demonstrate that virtual temperature sensing can effectively replace physical temperature sensors in smart buildings, especially through deployment of non-linear models with enabled memory capacity. The investigated approaches offer an increase in the cost-efficiency in the pathway towards scalable, intelligent building management systems
General practitioners’ adoption of generative artificial intelligence in clinical practice in the UK : An updated online survey
No full textBackground: Following the launch of ChatGPT in November 2022, interest in large language model-powered chatbots has soared with increasing focus on the clinical potential of these tools. Building on a previous survey conducted in 2024, we sought to gauge general practitioners? (GPs) adoption of this new generation of chatbots to assist with any aspect of clinical practice in the UK. Methods: An online survey was disseminated in January 2025 to a stratified convenience sample of GPs registered with the clinician marketing platform Doctors.net.uk. The research was conducted as part of a scheduled monthly ?omnibus survey,? designed to achieve a fixed sample size of 1000 participants. Results: Of the 1005 respondents, 50% respondents were men, 54% were 46 years or older. 25% reported using generative artificial intelligence (GenAI) tools in clinical practice; of these, 35% reported using these tools to generate documentation after patient appointments, 27% to suggest a differential diagnosis, 24% for treatment options, and 24% for referrals. Of the 249 GPs who used generative AI for clinical tasks, 71% said that, in general, these tools reduced work burdens. In the last 12 months, 85% reported that their employer had not encouraged them to use GenAI tools, but only 3% said their employer had prohibited them from using GenAI tools in their work; 95% reported they had no professional training in using GenAI tools in their work. Conclusions: This survey suggests that doctors? use of GenAI in clinical practice may be growing in the UK. Findings suggest that UK GPs may benefit from these tools, especially for administrative tasks and clinical reasoning support, and after adopting them, most users reported a decrease in work burdens. Continued absence of reported training remains a concern
The “Savage” Within: : The Sámi People and the Archaeology of Whiteness
No full textIn this article, I will examine how the Swedish zoologist and archaeologist Sven Nilsson (1787–1883) constructed what I refer to here as the archaeology of whiteness. Whiteness denotes here a racialized and colonial epistemic venture that positioned Europeans at an “evolutionary advantage” over Indigenous and colonized peoples. Specifically, I will analyze how Nilsson’s writing denied the Sámi people, Indigenous to Scandinavia and the Kola Peninsula of Russia, any sense of temporal “coevalness” with Europeans and how they were marked as racially and temporally different from other European inhabitants. I will conclude by addressing the colonial and political implications of this denial of Sámi historical motion, particularly in the context of how narratives of Sámi “savagery” and their supposed extinction were entangled with the increased extraction of natural resources from Sámi territories in Sweden and settler colonialism
Bestående ekonomisk-historiska nätverk : Vänbok till professor Jan Ottosson
No full textThis book is a festschrift to professor Jan Ottosson in honour of his retirement from the Department of Economic History, Uppsala university at the age of 67. Jan Ottosson has been a doctoral student, researcher and university teacher sine the 1980s. Since then he has done extensive work in a number of different fields, including the role of networks, theories on economic change and continuity, infrastructure and transport history, labour markets and industrial relations, banking and finance. This anthology includes 13 articles written by 19 economic historians that have meet and collaborated with Jan Ottosson over the years
Cardiorespiratory fitness in adolescence and risk of type 2 diabetes in late adulthood in one million Swedish men : nationwide sibling controlled cohort study
No full textObjective: To examine the association between adolescent cardiorespiratory fitness and risk of type 2 diabetes in late adulthood, including the potential influence of unobserved familial confounding on the association. Design: Nationwide sibling controlled cohort study. Setting: Swedish Military Service Conscription Register, Sweden, 1972-95, with Multi-Generation Register for identifying full siblings. National Patient Register and Prescribed Drug Register for data on diagnoses of type 2 diabetes, deaths from National Cause of Death Register, and Statistics Sweden for emigration and socioeconomic data. Participants: 1 124 049 Swedish men who participated in mandatory military conscription examinations with completed standardised cardiorespiratory fitness testing. Participants were followed up until 31 December 2023. Main outcome measures: Type 2 diabetes, defined as a composite endpoint of diagnosis in inpatient or specialist outpatient care and dispensation of antidiabetic drug treatment, until 31 December 2023. Results: 1 124 049 men, including 477 453 full siblings, with a mean age of 18.3 (standard deviation 0.7) years at baseline were included. During follow-up, 115 958 men (10.3%) and 48 089 full siblings (10.1%) had a first type 2 diabetes event at a median age of 53.4 (interquartile range 47.6-59.3) years. Cardiorespiratory fitness was categorised into deciles (referred to as groups, with group 1 being the lowest fitness level and group 10 the highest). In a cohort analysis, the adjusted hazard ratio in fitness group 2 versus fitness group 1 was 0.83 (95% confidence interval (CI) 0.81 to 0.85), with a difference in the standardised cumulative incidence at age 65 years of 4.3 (95% CI 3.8 to 4.8) percentage points, decreasing to a hazard ratio of 0.38 (0.36 to 0.39; incidence difference 17.8 (17.3 to 18.3) percentage points) in fitness group 10. When comparing full siblings, and thus controlling for all unobserved shared behavioural, environmental, and genetic confounders, the association was replicated, but with a reduction in magnitude. The hazard ratio in fitness group 2 was 0.89 (95% CI 0.85 to 0.94; incidence difference 2.3 (1.3 to 3.3) percentage points) and 0.53 (0.50 to 0.57; incidence difference 10.9 (9.7 to 12.1) percentage points) in fitness group 10. Hypothetically moving all participants in fitness group 1 to fitness group 2 was estimated to prevent 7.2% (95% CI 6.4% to 8.0%) of events at age 65 years in the cohort analysis versus 4.6% (2.6% to 6.5%) in the full sibling analysis, whereas hypothetically moving all participants to fitness group 10 was estimated to prevent 35.6% (34.1% to 37.0%) versus 24.3% (20.5 to 28.0) of events. Indications of effect modification by overweight status were found, where the association was smaller in those with overweight than in those without overweight, particularly in the full sibling analysis. Conclusions: The findings indicate that adolescent cardiorespiratory fitness could be important in the development of type 2 diabetes in late adulthood, but conventional observational analysis might give biased estimates of the magnitude of the effect
Per-and polyfluoroalkyl substances exposure in hexavalent chromium exposed workers and the effects of exposure mixtures on oxidative stress and genomic instability
No full textHexavalent chromium (Cr(VI)) can induce oxidative stress, genomic instability, and epigenetic modifications. In occupational settings, Cr(VI)-exposed workers may also be exposed to other toxicants, such as elements besides Cr, and per- and polyfluoroalkyl substances (PFAS). However, research on the extent of these co-exposures and their combined effects remains limited. The objective of this study was to characterize the exposure levels of ten elements and eight PFAS in Cr(VI)-exposed workers and to assess the combined effects of these exposure mixtures on biomarkers of oxidative stress and genomic instability. This study included 138 Cr(VI)-exposed workers and 96 controls from Swedish SafeChrom and Danish SAM-Krom studies. Concentrations of elements were measured by inductively coupled plasma mass spectrometry (ICP-MS), and PFAS were measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Effect biomarkers, including 8-hydroxy-2 '-deoxyguanosine (8-OHdG), mitochondrial DNA copy number (mtDNA-cn), telomere length (TL) and O6methylguanine-DNA methyltransferase promoter (MGMT) methylation, were analyzed in blood. Bayesian Kernel Machine Regression and quantile-based g-computation models were used to evaluate the mixture effects. Exposed workers had higher concentrations of Cr, manganese, copper, zinc, lead, and perfluoroheptanoic acid (PFHpA), lower mtDNA-cn and longer TL compared to controls. In the SAM-Krom study, perfluorooctane sulfonic acid (PFOS) levels were significantly elevated among exposed workers, with the P95 reaching 2044 ng/mL. The exposure mixtures were associated with increased 8-OHdG and MGMT hypermethylation. Together, these findings highlight the complexity of multiple occupational exposures in Cr(VI)-related work environments and suggest that combined exposure may contribute to early biological alterations related to oxidative stress and DNA methylation
Assessing changes in the availability and readiness of health facilities to provide modern family planning services in Bangladesh : Insights from Bangladesh Health Facility Surveys, 2014 and 2017
No full textBackground: Modern family planning plays a vital role in reducing unintended pregnancies, a major reproductive health issue worldwide. Access to modern family planning services is essential for empowering women to have greater control over their reproductive health and rights. In Bangladesh, there remains an unmet need for modern family planning services among reproductive-aged women. Assessing the capacity of health facilities to address these unmet needs for modern family planning is crucial. Objectives: The objective of this study was to assess the changes in the availability and readiness of health facilities to provide modern family planning services in Bangladesh between 2014 and 2017, and identify factors associated with facility readiness. Methods: We performed a secondary analysis of cross-sectional data from Bangladesh Health Facility Surveys (BHFS) conducted in 2014 and 2017. Availability was determined based on whether a facility offered at least one modern family planning method, and facility readiness was measured following the Service Availability and Readiness Assessment (SARA) manual. Descriptive statistics with 95% confidence intervals (CIs) were reported, and Poisson regression models were used to identify factors associated with health facility readiness. Findings: The percentage of facilities offering modern family planning services increased significantly from approximately 81% (95% CI: 78, 85) in 2014 to 89% (95% CI: 87, 91) in 2017. The availability of oral pills, injectables, and male condoms increased over this period, while the availability of long-acting reversible contraceptives (LARCs) slightly decreased, and permanent methods (PMs) remained nearly unchanged. The overall mean readiness score of health facilities declined slightly, from about 54 (95% CI: 52, 56) in 2014 to 51 (95% CI: 50, 53) in 2017. Upazila Health Complexes and Maternal and Child Welfare Centers had significantly higher readiness compared to District Hospitals in 2017. Facilities that performed routine quality assurance activities, ensured 24-hour staff coverage, maintained a system for reviewing clients' feedback, and provided family planning services regularly demonstrated significantly higher readiness to provide modern family planning services in both 2014 and 2017. Regional disparities were also observed; facilities in rural areas had significantly lower readiness than those in urban areas, and facilities from the Rangpur division showed significantly higher readiness compared to those in Dhaka in both survey years. Conclusion: The findings indicate a significant increase in the availability of health facilities offering modern family planning services in Bangladesh; however, a slight decline has been observed in their overall mean readiness score. Ensuring an adequate provision of equipment and supplies, expanding access to LARCs and PMs, and improving staff capacity through regular training are essential. Furthermore, strengthening quality assurance activities and investing in rural facilities are required for improving the facility readiness and advancing progress toward achieving SDG 3.7 targets of universal access to modern family planning services in Bangladesh
Comparative Preclinical Evaluation of the Tumor-Targeting Properties of Radioiodine and Technetium-Labeled Designed Ankyrin Repeat Proteins for Imaging of Epidermal Growth Factor Receptor Expression in Malignant Tumors.
No full textRadionuclide molecular imaging of epidermal growth factor receptor (EGFR) expression might permit the selection of patients for EGFR-targeting therapies. Designed ankyrin repeat protein (DARPin) E01 with a high affinity to the ectodomain III of the EGFR is a possible EGFR imaging probe. The goal of this study was to evaluate the potential of radiolabeled DARPin E01 for in vivo imaging of EGFR. DARPin E01 containing the (HE)3-tag was site-specifically labeled with a residualizing 99mTc (using 99mTc]Tc(CO)3). Two methods providing non-residualizing 123I labels, direct electrophilic radioiodination and indirect radioiodination using [123I]I-para-iodobenzoate (PIB), were tested. [99mTc]Tc-(HE)3-E01 and [123I]I-(HE)3-E01-PIB preserved specific binding to EGFR-expressing cells and affinity in the single-digit nanomolar range. Direct labeling with 123I resulted in a substantial loss of binding. In vitro cellular processing studies showed that both [99mTc]Tc-(HE)3-E01 and [123I]I-(HE)3-E01-PIB had rapid binding and relatively slow internalization. Evaluation of [99mTc]Tc-(HE)3-E01 biodistribution in normal CD1 mice showed that its hepatic uptake was non-saturable, suggesting that this tracer does not bind to murine EGFR. A side-by-side comparison of biodistribution and tumor targeting of [99mTc]Tc-(HE)3-E01 and [123I]I-(HE)3-E01-PIB was performed in Nu/j mice bearing EGFR-positive A-431 and EGFR-negative Ramos human cancer xenografts. Both radiolabeled DARPins demonstrated EGFR-specific tumor uptake. However, [123I]I-(HE)3-E01-PIB had appreciably lower uptake in normal organs compared to [99mTc]Tc-(HE)3-E01, which provided significantly (p < 0.05) higher tumor-to-organ ratios. Gamma-camera imaging confirmed that [123I]I-(HE)3-E01-PIB demonstrated a higher imaging contrast in preclinical models than [99mTc]Tc-(HE)3-E01. In conclusion, DARPin (HE)3-E01 labeled using a non-residualizing [123I]I-para-iodobenzoate (PIB) label is the preferred radiotracer for in vivo imaging of EGFR expression in cancer
