Johan Linåker, Astor Nummelin Carlberg, Ciaran O'Riordan
Context: Open Source Software (OSS) is a crucial component of over 90\% of digital infrastructure underpinning industry and public digital services, facilitating collaborative software development and dissemination. Its significance in the European public sector has been emphasised through various Ministerial Declarations, highlighting its potential to accelerate digitalisation, transform businesses, and foster a digitally skilled population. Research Aim: This study aims to explore how the adoption, development, and collaboration on OSS can be enabled through organisational support functions or centres of competency, also known as Open Source Programme Offices (OSPOs) within Public Sector Organisations (PSOs) in the European Union, Norway, Liechtenstein, and Iceland. Methodology: A qualitative research approach was adopted, involving an interview survey of 18 OSPO representatives across 16 cases of public-sector OSPOs. These cases were cross-analysed and categorised into six OSPO archetypes. The findings were validated and enriched through two follow-up focus groups that included earlier interviewees and additional experts. Results: The study identified six distinct OSPO archetypes, providing insights into their organisational structures, responsibilities, and contributions to OSS adoption. The archetypes, along with policy recommendations, offer guidance on how PSOs can design their own OSPOs, taking into account their specific context, resources, and policy goals. Conclusions: The findings enhance the understanding of OSPOs as strategic endeavours aimed at promoting OSS adoption. The study offers practical guidance for PSOs and policymakers on leveraging OSS to achieve strategic objectives, foster digital sovereignty, drive economic growth, and improve the interoperability and quality of digital services.
Annette Alstadsæter, Martin Jacob, Wojciech Kopczuk
et al.
Abstract Using linked individual and firm administrative data from Norway, we look through layers of holding companies and attribute corporate profits to the ultimate personal owner as the profits accrue rather than when income is realized. We show that our accrual-based measure of income inequality changes the level and trend of income inequality over time and eliminates the sensitivity of measures of inequality and income persistence to changing payout policies in response to tax reforms. After a tax reform in 2005 that incentivized retention of earnings within businesses, the total income share of the top 0.1% more than doubled in some years, compared with ordinary realization-based income measures. We further utilize rich data to show that (1) using our accrual-based measure of personal income reduces the estimated tax elasticity of income, and (2) observed capital income in individuals’ tax returns does not proxy well for overall corporate profits, so that an imputation method based on realized dividends, which is commonly used in the literature, performs poorly. We discuss implications for top income inequality measures in other countries. We also document the importance of indirect ownership as a mechanism behind our findings and its relevance in other developed countries and discuss implications for debates on capital income and wealth taxation.
Taking a critical geopolitical approach, this article shows how Europe has been mobilized as a geographical object for thinking about and putting into practice transnational environmental conservation between Finland, Norway and Russia since the end of the Cold War. This study contributes to the literature on critical geopolitics and environmental governance by examining the intersection of conservation and international relations in post-Soviet Europe. It engages with scholarship on geopolitical imaginaries, which conceptualize how social constructed spatial entities such as Europe are mobilized for thinking about and putting into practice international relations. The article builds on existing research on transboundary conservation, European integration, and the role of environmental initiatives in shaping geopolitical narratives. By exploring how conservation efforts are used in international relations, this research adds to debates on the instrumentalization of environmental governance within broader geopolitical frameworks.Methodologically, this study employs a qualitative approach, combining documentary research with in-depth, semi-structured interviews. The documentary analysis includes official policy documents, reports, and promotional materials related to the Green Belt of Fennoscandia (GBF). The study also draws on forty interviews with policymakers, conservationists, and local stakeholders across Finland, Norway, and Russia. These interviews explore perceptions of the GBF’s role in transnational governance and its function within European-Russian relations. A thematic analysis of the collected data enables a nuanced understanding of how conservation initiatives are framed and mobilized in different political and institutional contexts.This article illustrates how the relationship between the European Union and Russia is not the work of two monolithic blocs motivated solely by their interest in power, but that it operates through diverse channels and responds to the subjectivities of the actors who make it up at local level. Firstly, it shows how the development of the green belt was linked to the idea of spreading the values of the European project in post-Soviet Russia. Secondly, it explains how transnational environmental conservation is organized and actually operates at local level through decentralized actors.
ABSTRACT The rumen microbiome plays a critical role in determining feed conversion and methane emissions in cattle, with significant implications for both agricultural productivity and environmental sustainability. In this study, we applied a hierarchical joint species distribution model to predict directional associations between biotic factors and abundances of microbial populations determined via metagenome-assembled genomes (MAGs). Our analysis revealed distinct microbial differences, including 191 MAGs significantly more abundant in animals with a higher methane yield (above 24 g/kg dry matter intake [DMI]; high-emission cattle), and 220 MAGs more abundant in low-emission cattle. Interestingly, the microbiome community of the low-methane-emission rumen exhibited higher metabolic capacity but with lower functional redundancy compared to that of high-methane-emission cattle. Our findings also suggest that microbiomes associated with low methane yields are prevalent in specific functionalities such as active fiber hydrolysis and succinate production, which may enhance their contributions to feed conversion in the host animal. This study provides an alternate genome-centric means to investigate the microbial ecology of the rumen and identify microbial and metabolic intervention targets that aim to reduce greenhouse gas emissions in livestock production systems.IMPORTANCERuminant livestock are major contributors to global methane emissions, largely through microbial fermentation in the rumen. Understanding how microbial communities vary between high- and low-methane-emitting animals is critical for identifying mitigation strategies. This study leverages a genome-centric approach to link microbial metabolic traits to methane output in cattle. By reconstructing and functionally characterizing hundreds of microbial genomes, we observe that a low-methane-emission rumen harbors well-balanced, “streamlined” microbial communities characterized by high metabolic capacity and minimal metabolic overlap across populations (low functional redundancy). Our results demonstrate the utility of genome-level functional profiling in uncovering microbial community traits tied to climate-relevant phenotypes.
We compare the performance of state-of-the-art Large Language Models on a recently released benchmarking set for automated question answering for Icelandic and compare it with performance on questions from an Icelandic trivia game. We find that the models perform worse for questions on Icelandic subjects, specifically Icelandic culture, but somewhat surprisingly do better on a trivia game for people than on the benchmark set meant for language models, built around data that the model has seen during training. We also call into question some aspects of the benchmarking set and discuss what
playing trivia games can tell us - if anything - about the capabilities of these models.
Bibliography. Library science. Information resources
Muhammad Shahzad Javed, Karin Fossheim, Paola Velasco-Herrejón
et al.
Environmental movements and climate strikes have made it apparent that youth feel excluded from the ongoing energy transformation process, highlighting the crucial need for their engagement to achieve a socially accepted transition. This interdisciplinary study focuses on the Norwegian electricity system and involves conducting educational workshops with high school students aged 15 to 16 to ascertain their perspectives towards a net-zero energy system. The workshops were structured into three segments, starting with the dissemination of common knowledge about energy and climate, followed by interactive activities designed to explore and develop an understanding of various aspects of energy transition. Three rounds of questionnaires, administered at distinct time intervals, assessed changes in students' attitudes and socio-techno-economic preferences. Our findings show that 33\% of pupils favored exclusively offshore wind as a main energy source, while 35\% opted to combine it with solar energy, indicating that over 68\% viewed offshore wind as favorable. Although 32\% supported some form of land-based wind turbines, there was strong disagreement about wind parks in agricultural, forested, and residential areas. Preferences also exhibited considerable regional variation; solar installations were favored in southern and southeastern Norway, while wind farms were suggested for central and northern regions. Pupils emphasized energy independence, showed reluctance towards demand response, prioritized reducing emissions and preserving biodiversity over minimizing electricity costs. Despite cost-minimization being core to most energy system models, youth deemed it the least important factor, highlighting a disconnect between modeling priorities and their perspectives.
Júlio O. Amando de Barros, Jakob Schwiedrzik, Falk K. Wittel
Wood's increasing role as a structural resource in sustainable materials selection demands accurate characterization of its mechanical behavior. Its performance arises from a hierarchical structure, where the dominant load-bearing component is the S2 layer of tracheid cell walls-a thick, fiber-reinforced composite of cellulose microfibrils embedded in hemicelluloses and lignin. Due to the small dimensions and anisotropic nature of the S2 layer, mechanical testing presents significant challenges, particularly in producing homogeneous stress and strain fields. In this study, we apply micropillar compression (MPC) combined with digital image correlation (DIC) to Norway spruce tracheids, enabling direct and model-free strain measurements at the cell wall scale. Micropillars were oriented at different microfibril angles (MFAs), confirming the expected dependence of stiffness and yield stress on ultrastructural alignment, with higher stiffness and yield stress at low MFAs. For these under compression fibril-aligned kink bands occurred, while shear related failure was observed at higher angles. A parameter study on the acceleration voltage of the Scanning Electron Microscope revealed that electron beam exposure significantly degrades pillar integrity, which could explain data scatter and mechanical underestimation in earlier MPC studies. By controlling imaging protocols and using DIC-based strain measurements, we report the highest direct measurements of wood cell wall stiffness to date-up to 42 GPa for MFA=0°-closer matching micromechanical model predictions compared to previous results. Findings are compared with Finite Element Method-based displacement corrections to establish a robust protocol for probing soft, anisotropic biological composites' mechanical behavior while clarifying longstanding inconsistencies in reported results of wood MPC measurements.
Music information retrieval distinguishes between low- and high-level descriptions of music. Current generative AI models rely on text descriptions that are higher level than the controls familiar to studio musicians. Pitch strength, a low-level perceptual parameter of contemporary popular music, may be one feature that could make such AI models more suited to music production. Signal and perceptual analyses suggest that pitch strength (1) varies significantly across and inside songs; (2) contributes to both small- and large-scale structure; (3) contributes to the handling of polyphonic dissonance; and (4) may be a feature of upper harmonics made audible in a perspective of perceptual richness.
Amani Thomas Mori, Amani Thomas Mori, Amani Thomas Mori
et al.
BackgroundNearly 100 million people are pushed into poverty every year due to catastrophic health expenditures (CHE). We evaluated the impact of cash support programs on healthcare utilization and CHE among households participating in a cluster-randomized controlled trial focusing on adolescent childbearing in rural Zambia.Methods and findingsThe trial recruited adolescent girls from 157 rural schools in 12 districts enrolled in grade 7 in 2016 and consisted of control, economic support, and economic support plus community dialogue arms. Economic support included 3 USD/month for the girls, 35 USD/year for their guardians, and up to 150 USD/year for school fees. Interviews were conducted with 3,870 guardians representing 4,110 girls, 1.5–2 years after the intervention period started. Utilization was defined as visits to formal health facilities, and CHE was health payments exceeding 10% of total household expenditures. The degree of inequality was measured using the Concentration Index. In the control arm, 26.1% of the households utilized inpatient care in the previous year compared to 26.7% in the economic arm (RR = 1.0; 95% CI: 0.9–1.2, p = 0.815) and 27.7% in the combined arm (RR = 1.1; 95% CI: 0.9–1.3, p = 0.586). Utilization of outpatient care in the previous 4 weeks was 40.7% in the control arm, 41.3% in the economic support (RR = 1.0; 95% CI: 0.8–1.3, p = 0.805), and 42.9% in the combined arm (RR = 1.1; 95% CI: 0.8–1.3, p = 0.378). About 10.4% of the households in the control arm experienced CHE compared to 11.6% in the economic (RR = 1.1; 95% CI: 0.8–1.5, p = 0.468) and 12.1% in the combined arm (RR = 1.1; 95% CI: 0.8–1.5, p = 0.468). Utilization of outpatient care and the risk of CHE was relatively higher among the least poor than the poorest households, however, the degree of inequality was relatively smaller in the intervention arms than in the control arm.ConclusionsEconomic support alone and in combination with community dialogue aiming to reduce early childbearing did not appear to have a substantial impact on healthcare utilization and CHE in rural Zambia. However, although cash transfer did not significantly improve healthcare utilization, it reduced the degree of inequality in outpatient healthcare utilization and CHE across wealth groups.Trial Registrationhttps://classic.clinicaltrials.gov/ct2/show/NCT02709967, ClinicalTrials.gov, identifier (NCT02709967).
Eirini Katsidoniotaki, Biao Su, Eleni Kelasidi
et al.
As the global population grows and climate change intensifies, sustainable food production is critical. Marine aquaculture offers a viable solution, providing a sustainable protein source. However, the industry's expansion requires novel technologies for remote management and autonomous operations. Digital twin technology can advance the aquaculture industry, but its adoption has been limited. Fish net cages, which are flexible floating structures, are critical yet vulnerable components of aquaculture farms. Exposed to harsh and dynamic marine environments, the cages experience significant loads and risk damage, leading to fish escapes, environmental impacts, and financial losses. We propose a multifidelity surrogate modeling framework for integration into a digital twin for real-time monitoring of aquaculture net cage structural dynamics under stochastic marine conditions. Central to this framework is the nonlinear autoregressive Gaussian process method, which learns complex, nonlinear cross-correlations between models of varying fidelity. It combines low-fidelity simulation data with a small set of high-fidelity field sensor measurements, which offer the real dynamics but are costly and spatially sparse. Validated at the SINTEF ACE fish farm in Norway, our digital twin receives online metocean data and accurately predicts net cage displacements and mooring line loads, aligning closely with field measurements. The proposed framework is beneficial where application-specific data are scarce, offering rapid predictions and real-time system representation. The developed digital twin prevents potential damages by assessing structural integrity and facilitates remote operations with unmanned underwater vehicles. Our work also compares GP and GCNs for predicting net cage deformation, highlighting the latter's effectiveness in complex structural applications.
Theresa M. Kirchner, Olivier Devineau, Marianna Chimienti
et al.
Abstract Background Monitoring the behavior of wild animals in situ can improve our understanding of how their behavior is related to their habitat and affected by disturbances and changes in their environment. Moose (Alces alces) are keystone species in their boreal habitats, where they are facing environmental changes and disturbances from human activities. How these potential stressors can impact individuals and populations is unclear, in part due to our limited knowledge of the physiology and behavior of moose and how individuals can compensate for stress and disturbances they experience. We collected data from collar-mounted fine-scale tri-axial accelerometers deployed on captive moose in combination with detailed behavioral observations to train a random forest supervised classification algorithm to classify moose accelerometer data into discrete behaviors. To investigate the generalizability of our model to collared new individuals, we quantified the variation in classification performance among individuals. Results Our machine learning model successfully classified 3-s accelerometer data intervals from 12 Alaskan moose (A. a. gigas) and two European moose (A. a. alces) into seven behaviors comprising 97.6% of the 395 h of behavioral observations conducted in summer, fall and spring. Classification performance varied among behaviors and individuals and was generally dependent on sample size. Classification performance was highest for the most common behaviors lying with the head elevated, ruminating and foraging (precision and recall across all individuals between 0.74 and 0.90) comprising 79% of our data, and lower and more variable among individuals for the four less common behaviors lying with head down or tucked, standing, walking and running (precision and recall across all individuals between 0.28 and 0.79) comprising 21% of our data. Conclusions We demonstrate the use of animal-borne accelerometer data to distinguish among seven main behaviors of captive moose and discuss generalizability of the results to individuals in the wild. Our results can support future efforts to investigate the detailed behavior of collared wild moose, for example in the context of disturbance responses, time budgets and behavior-specific habitat selection.
Lamprini Veneti, Jacob Dag Berild, Sara Viksmoen Watle
et al.
Objectives: We estimated the BNT162b2 vaccine effectiveness (VE) against any (symptomatic or not) SARS-CoV-2 Delta and Omicron infection among adolescents (aged 12-17 years) in Norway from August 2021 to January 2022. Methods: We used Cox proportional hazard models, where vaccine status was included as a time-varying covariate and models were adjusted for age, sex, comorbidities, residence county, birth country, and living conditions. Results: The VE against Delta infection peaked at 68% (95% confidence interval [CI]: 64-71%) and 62% (95% CI: 57-66%) in days 21-48 after the first dose among those aged 12-15 years and 16-17 years, respectively. Among those aged 16-17 years who received two doses, the VE against Delta infection peaked at 93% (95% CI: 90-95%) in days 35-62 and decreased to 84% (95% CI: 76-89%) in ≥63 days after vaccination. We did not observe a protective effect against Omicron infection after receiving one dose. Among those aged 16-17 years, the VE against Omicron infection peaked at 53% (95% CI: 43-62%) in 7-34 days after the second dose and decreased to 23% (95% CI: 3-40%) in ≥63 days after vaccination. Conclusion: We found a reduced protection after two BNT162b2 vaccine doses against any Omicron infection compared to Delta. Effectiveness decreased with time from vaccination for both variants. The impact of vaccination among adolescents on reducing infection and thus transmission is limited during the Omicron dominance.
Irene García-Meilán, Juan Ignacio Herrera-Muñoz, Borja Ordóñez-Grande
et al.
The effect of different main dietary compositions on growth, anticipatory digestive enzyme activities, and oxidative status was studied in the proximal intestine of juvenile European sea bass. A control diet (C, 44% protein, 17.6% lipid, and 20% starch), three diets with increasing starch levels to test protein sparing (P36S36, P40S29, and P43S24), and two diets with high lipid content (L20S13 and L22S7) were tested. After 20 weeks, growth, digestive enzyme activities, lipid peroxidation, antioxidant enzyme activities, and G6PDH activity were measured after a 24-h fast. Sea bass fed P43S24 and L20S13 maintained an oxidative status like C fish, up-regulated CAT activity, and adjusted anticipatory protease activity. Instead, the lipid peroxidation increased in the L22S7 group, although CAT activity increased, whereas anticipatory total protease activity was downregulated. P40S29 also triggered LPO and CAT activity, but G6PDH levels diminished significantly. Moreover, an up-regulation in digestive enzyme activities was found. Finally, P36S36 fish showed less antioxidant enzyme activity and G6PDH, although their LPO tended to increase and their lipase and α-amylase activities were upregulated. In conclusion, the inclusion of carbohydrates up to 24% or lipids up to 20% is possible for this species if protein requirements are met without negative effects on growth.
Christopher Page, Huiru Zheng, Haiying Wang
et al.
We sought to determine the most efficacious and cost-effective strategy to follow when developing a national screening programme by comparing and contrasting the national screening programmes of Norway, the Netherlands and the UK. Comparing the detection rates and screening profiles between the Netherlands, Norway, the UK and constituent nations (England, Northern Ireland, Scotland and Wales) it is clear that maximising the number of relatives screened per index case leads to identification of the greatest proportion of an FH population. The UK has stated targets to detect 25% of the population of England with FH across the 5 years to 2024 with the NHS Long Term Plan. However, this is grossly unrealistic and, based on pre-pandemic rates, will only be reached in the year 2096. We also modelled the efficacy and cost-effectiveness of two screening strategies: 1) Universal screening of 1-2-year-olds, 2) electronic healthcare record screening, in both cases coupled to reverse cascade screening. We found that index case detection from electronic healthcare records was 56% more efficacious than universal screening and, depending on the cascade screening rate of success, 36%-43% more cost-effective per FH case detected. The UK is currently trialling universal screening of 1–2-year-olds to contribute to national FH detection targets. Our modelling suggests that this is not the most efficacious or cost-effective strategy to follow. For countries looking to develop national FH programmes, screening of electronic healthcare records, coupled to successful cascade screening to blood relatives is likely to be a preferable strategy to follow.
Tommy Nyberg, Peter Bager, Ingrid Bech Svalgaard
et al.
Several SARS-CoV-2 variants that evolved during the COVID-19 pandemic have appeared to differ in severity, based on analyses of single-country datasets. With decreased SARS-CoV-2 testing and sequencing, international collaborative studies will become increasingly important for timely assessment of the severity of newly emerged variants. The Joint WHO Regional Office for Europe and ECDC Infection Severity Working Group was formed to produce and pilot a standardised study protocol to estimate relative variant case-severity in settings with individual-level SARS-CoV-2 testing and COVID-19 outcome data during periods when two variants were co-circulating. To assess feasibility, the study protocol and its associated statistical analysis code was applied by local investigators in Denmark, England, Luxembourg, Norway, Portugal and Scotland to assess the case-severity of Omicron BA.1 relative to Delta cases. After pooling estimates using meta-analysis methods (random effects estimates), the risk of hospital admission (adjusted hazard ratio [aHR]=0.41, 95% CI 0.31-0.54), ICU admission (aHR=0.12, 95% CI 0.05-0.27), and death (aHR=0.31, 95% CI 0.28-0.35) was lower for Omicron BA.1 compared to Delta cases. The aHRs varied by age group and vaccination status. In conclusion, this study has demonstrated the feasibility of conducting variant severity analyses in a multinational collaborative framework. The results add further evidence for the reduced severity of the Omicron BA.1 variant.
Kwaku Peprah Adjei, Robert B. O'Hara, Wouter Koch
et al.
1. Species distribution models and maps from large-scale biodiversity data are necessary for conservation management. One current issue is that biodiversity data are prone to taxonomic misclassifications. Methods to account for these misclassifications in multispecies distribution models have assumed that the classification probabilities are constant throughout the study. In reality, classification probabilities are likely to vary with several covariates. Failure to account for such heterogeneity can lead to bias in parameter estimates. 2. Here we present a general multispecies distribution model that accounts for heterogeneity in the classification process. The proposed model assumes a multinomial generalised linear model for the classification confusion matrix. We compare the performance of the heterogeneous classification model to that of the homogeneous classification model by assessing how well they estimate the parameters in the model and their predictive performance on hold-out samples. We applied the model to gull data from Norway, Denmark and Finland, obtained from GBIF. 3. Our simulation study showed that accounting for heterogeneity in the classification process increased precision by 30% and reduced accuracy and recall by 6%. Applying the model framework to the gull dataset did not improve the predictive performance between the homogeneous and heterogeneous models due to the smaller misclassified sample sizes. However, when machine learning predictive scores are used as weights to inform the species distribution models about the classification process, the precision increases by 70%. 4. We recommend multiple multinomial regression to be used to model the variation in the classification process when the data contains relatively larger misclassified samples. Machine prediction scores should be used when the data contains relatively smaller misclassified samples.
Jorge Sicacha-Parada, Diego Pavon-Jordan, Ingelin Steinsland
et al.
The constant increase in energy consumption has created the necessity of extending the energy transmission and distribution network. Placement of powerlines represent a risk for bird population. Hence, better understanding of deaths induced by powerlines, and the factors behind them are of paramount importance to reduce the impact of powerlines. To address this concern, professional surveys and citizen science data are available. While the former data type is observed in small portions of the space by experts through expensive standardized sampling protocols, the latter is opportunistically collected by citizen scientists. We set up full Bayesian spatial models that 1) fusion both professional surveys and citizen science data and 2) explicitly account for preferential sampling that affects professional surveys data and for factors that affect the quality of citizen science data. The proposed models are part of the family of latent Gaussian models as both data types are interpreted as thinned spatial point patterns and modeled as log-Gaussian Cox processes. The specification of these models assume the existence of a common latent spatial process underlying the observation of both data types. The proposed models are used both on simulated data and on real-data of powerline-induced death of birds in the Trondelag in Norway. The simulation studies clearly show increased accuracy in parameter estimates when both data types are fusioned and factors that bias their collection processes are properly accounted for. The study of powerline-induced deaths shows a clear association between the density of the powerline network and the risk that powerlines represent for bird populations. The choice of model is relevant for the conclusions from this case study as different models estimated the association between risk of powerline-induced deaths and the amount of exposed birds differently.
Olga Viberg, Mutlu Cukurova, Yael Feldman-Maggor
et al.
With growing expectations to use AI-based educational technology (AI-EdTech) to improve students' learning outcomes and enrich teaching practice, teachers play a central role in the adoption of AI-EdTech in classrooms. Teachers' willingness to accept vulnerability by integrating technology into their everyday teaching practice, that is, their trust in AI-EdTech, will depend on how much they expect it to benefit them versus how many concerns it raises for them. In this study, we surveyed 508 K-12 teachers across six countries on four continents to understand which teacher characteristics shape teachers' trust in AI-EdTech, and its proposed antecedents, perceived benefits and concerns about AI-EdTech. We examined a comprehensive set of characteristics including demographic and professional characteristics (age, gender, subject, years of experience, etc.), cultural values (Hofstede's cultural dimensions), geographic locations (Brazil, Israel, Japan, Norway, Sweden, USA), and psychological factors (self-efficacy and understanding). Using multiple regression analysis, we found that teachers with higher AI-EdTech self-efficacy and AI understanding perceive more benefits, fewer concerns, and report more trust in AI-EdTech. We also found geographic and cultural differences in teachers' trust in AI-EdTech, but no demographic differences emerged based on their age, gender, or level of education. The findings provide a comprehensive, international account of factors associated with teachers' trust in AI-EdTech. Efforts to raise teachers' understanding of, and trust in AI-EdTech, while considering their cultural values are encouraged to support its adoption in K-12 education.
Understanding and predicting interactions between predators and prey and their environment are fundamental for understanding food web structure, dynamics, and ecosystem function in both terrestrial and marine ecosystems.Thus, estimating the conditional associations between species and their environments is important for exploring connections or cooperative links in the ecosystem, which in turn can help to clarify such causal relationships. For this purpose, a relevant and practical statistical method is required to link presence/absence observations with biomass, abundance, and physical quantities obtained as continuous real values.These data are sometimes sparse in oceanic space and too short as time series data. To meet this challenge, we provide an approach based on applying categorical data analysis to present/absent observations and real-number data.This approach consists of a two-step procedure for categorical data analysis:1) finding the appropriate threshold to discretize the real-number data for applying an independent test;and 2) identifying the best conditional probability model to investigate the possible associations among the data based on a statistical information criterion.We conduct a simulation study to validate our proposed approach. Furthermore, the approach is applied to two datasets: 1) one collected during an international synoptic krill survey in the Scotia Sea west of the Antarctic Peninsula to investigate associations among krill, fin whale (Balaenoptera physalus),surface temperature, depth, slope in depth, and temperature gradient; 2) the other collected by ecosystem surveys conducted during August-September in 2014 - 2017 to investigate associations among common minke whales, the predatory fish Atlantic cod, and their main prey groups in Arctic Ocean waters to the west and north of Svalbard, Norway.
Javier de la Rosa, Rolv-Arild Braaten, Per Egil Kummervold
et al.
In this paper, we present several baselines for automatic speech recognition (ASR) models for the two official written languages in Norway: Bokmål and Nynorsk. We compare the performance of models of varying sizes and pre-training approaches on multiple Norwegian speech datasets. Additionally, we measure the performance of these models against previous state-of-the-art ASR models, as well as on out-of-domain datasets. We improve the state of the art on the Norwegian Parliamentary Speech Corpus (NPSC) from a word error rate (WER) of 17.10\% to 7.60\%, with models achieving 5.81\% for Bokmål and 11.54\% for Nynorsk. We also discuss the challenges and potential solutions for further improving ASR models for Norwegian.