37115 research outputs found
Sort by
Fault and Fracture Patterns with Linkage to Fluid Migration in the Shallow Subsurface: A 3D Seismic Study along a transect of the Western Barents Sea
Get PDFThis thesis investigates shallow subsurface deformation patterns and their relationship to fluid migration and glacial processes in the southwestern Barents Sea. The study focuses on four key areas: Lyngenfjorden, Hammerfest Basin, Bjørnøyrenna, and Storfjordrenna. Using high-resolution P-Cable 3D seismic data and interpretation tools in Petrel, the objective was to identify faults, fractures, and associated features such as pockmarks, mounds, and zones of acoustic blanking.
Seismic attributes including variance and ant tracking were applied to improve visibility of structural discontinuities. Variance proved effective for general structural outlines, while ant tracking provided directional fault detection with greater sensitivity to user-defined parameters.
The findings reveal notable differences between the study areas. Lyngenfjorden displays small pockmarks and minor subsurface anomalies, likely related to earlier fluid release or groundwater flow. Hammerfest Basin contains dense faulting beneath the Upper Regional Unconformity, with depressions aligned to faulted zones suggesting structurally controlled fluid migration. In Bjørnøyrenna, large craters and mounds point toward gas hydrate formation and collapse. Storfjordrenna features domed seafloor mounds interpreted as gas hydrate pingos, associated with post-glacial warming.
This study demonstrates that the interaction between tectonics, glacial history, and sediment dynamics has played a major role in shaping the shallow subsurface. The methodology applied here offers valuable insights into the structural and fluid-related evolution of glaciated continental margins
Compliance by Design in AI Systems – Towards Compliance with Data Protection Regulation, The AI Act, and Cybersecurity Regulations in AI System Development and Deployment
No full text"Without prejudice to" is a term frequently used in EU technology regulations. Regulation 2024/1689 (The Artificial Intelligence (AI) Act) is "without prejudice to" the General Data Protection Regulation (GDPR), and recent cybersecurity regulations such as the Network and Information Security (NIS) 2 Directive and the Cyber Resilience Act (CRA) are without prejudice to both the GDPR and the AI Act. However, in situations where these regulations and directives enter into effect simultaneously, compliance with these rule sets would entail prioritising different design choices in technological solutions. These choices would, to a varying degree, enhance or diminish the end goals of the different applicable rule sets. This thesis explores the research question of whether it is possible to develop a compliance by design approach to processing personal data in AI systems compliant with the requirements of the GDPR, the AI Act, and specific relevant cybersecurity regulations.
The thesis examines the research question through several sub-research questions. The approach examines whether compliance with the identified legal requirements could be achieved through alterations to the methodologies applied when developing and maintaining developed AI systems. Thus, the first sub-research question examines the main methods and methodologies utilised for developing AI systems and managing their lifecycle post-deployment. After such an examination, relevant legal requirements are analysed during the different phases of the lifecycle of an AI system: data collection and model training, deployment, and when finished AI models are fine-tuned on AI-as-a-service (AIaaS) platforms.
The thesis identifies that one methodology frequently applied to developing and maintaining AI systems, Machine Learning Operations (MLOps), has the potential to implement and embed several relevant legal requirements including, but not limited to, the principle of accuracy in the GDPR, the requirement for accuracy under the AI Act, the right to receive meaningful information about the logic involved in automated decision making, and the purpose limitation principle. MLOps is a concept without a clear authoritative definition. However, the methodology includes automatic, semiautomatic, and manual processes for regularly training AI models across the development and deployment phases.
Moreover, the thesis also addresses the question of how to achieve compliance with different rule sets that enter into effect at the same time. Based on the findings of the legal and technical examinations, proportionality by design, based on the principle of proportionality, is developed as a framework providing an overview, an interpretative source, and a basis for best practices for AI development and use
Weaving knowledge: writing letters with wool
Get PDFSource at https://metode.r-o-m.no/artikler/essay/weaving-knowledge-writing-letters-with-wool.We are a group of three women: an artist, a researcher, and a museum professional, who at the same time struggle to be defined as “something”. Each of us has a different relationship to wool, yet, as we have come to see, wool dissolves the type of difference that would prevent us from finding our connection. For some of us, wool is a material to work with and create with— and more. For others of us, wool is a reminder of our need as academics to create a connection to tangible materiality, of the possibility that we have sat behind our computer screens long enough. How to read our text
Bioeconomic analysis of Norwegian reindeer husbandry in the face of crowberry encroachment
Get PDFAllelopathic encroachment of Empetrum nigrum (crowberry) is impacting the Norwegian reindeer husbandry. This thesis first explores both the negative and positive impacts of crowberry, and then analyses the effect of its encroachment on the husbandry under climate change. The thesis consists of four papers, one qualitative study conducting systematic literature review of Empetrum nigrum, and three quantitative studies applying the method of bioeconomic modeling. The first paper categorizes the benefits and detriments of crowberry using three different nature assessment concepts, and finds mixed impacts on the local socioecological system, particularly eight benefits and three detriments. Focusing on the detrimental impacts of crowberry, the three quantitative papers develop various versions of the reindeer bioeconomic model, including three stocks – vegetation, crowberry, and reindeer – along with two primary adaptive measures: reindeer feeding and crowberry control. The second paper applies static optimization and concludes that controlling crowberry alongside reindeer feeding can mitigate the negative impacts of encroachment. The third paper extends the analysis to dynamic optimization, confirming the importance of controlling crowberry to maintain high quality grazing pasture and reindeer population over time. Using optimal control theory, the fourth paper explores a specific yet neglected benefit of crowberry – carbon sequestration – and demonstrates that with this beneficial value, the optimal efforts to control crowberry should be less than in scenarios without. The thesis hence underscores several policy implications. Firstly, adoption of the Nature’s Contribution to People (NCP) concept may be embraced to provide a comprehensive perspective. Secondly, the socio-ecological system of the husbandry is impacted negatively by crowberry encroachment through deteriorating grazing pastures. Thirdly, while supplementary feeding can increase reindeer numbers, it does not address the ecological issues of pasture degradation, thus should be considered only as a short-term solution. Fourthly, given governmental support, crowberry control can be considered a sustainable solution to improve pasture quality and increase reindeer population. Lastly, to effectively balance the benefits and detriments of this plant, crowberry control efforts must be carefully managed in the long-run (beyond a 10-year perspective) to achieve desired outcomes.Miljøendringer medfører allelopatisk gjengroing av Empetrum nigrum (krekling) som påvirker den norske reindriftsnæringen. Denne avhandlingen studerer både negative og positive effekter av gjengroing av krekling på reindriften. Avhandlingen består av fire artikler, én kvalitativ studie som gjennomfører en systematisk litteraturgjennomgang av Empetrum nigrum, og tre kvantitative studier som anvender bioøkonomisk modellering. Den første artikkelen kategoriserer kreklings fordeler og ulemper ved hjelp av tre forskjellige naturvurderingskonsepter og finner blandede effekter på det lokale sosioøkologiske systemet, spesifikt åtte fordeler og tre ulemper. Med fokus på de negative effektene av krekling, utvikler de tre kvantitative artiklene forskjellige versjoner av den bioøkonomiske modellen, inkludert tre bestander – vegetasjon, krekling og rein – sammen med to primære tilpasningstiltak: tilleggsfôring og kontroll av krekling. Den andre artikkelen anvender statisk optimering og konkluderer med at kontroll av krekling sammen med tilleggsfôring kan mildne de negative effektene av gjengroing. Den tredje artikkelen utvider analysen til dynamisk optimering og bekrefter viktigheten av å kontrollere krekling for å opprettholde høykvalitets beiteområder for reindriften over tid. Ved bruk av optimal kontrollteori utforsker den fjerde artikkelen en spesifikk, men neglisjert, fordel forbundet med krekling – karbonsekvestrasjon – og demonstrerer at med denne fordelaktige verdien, bør de optimale tiltakene for å kontrollere krekling være mindre enn i scenarier uten. Avhandlingen understreker dermed flere forvaltningsmessige implikasjoner. For det første kan anvendelse av NCP-rammeverket omfavnes for å gi et bredere perspektiv på naturvurdering. For det andre påvirkes det sosioøkologiske systemet i reindriftsnæringen negativt av gjengroing av krekling gjennom forverring av beiteområder. For det tredje, selv om tilleggsfôring kan øke antall rein, adresserer det ikke de økologiske problemene på beiteområdet, og bør derfor kun vurderes som en kortsiktig løsning. For det fjerde, med statlig støtte, kan kontroll av krekling vurderes som en bærekraftig løsning for å forbedre beitekvaliteten og øke reinflokken. Til slutt, for å effektivt balansere fordelene og ulempene forbundet med krekling, må innsatsen for å kontrollere denne arten håndteres nøye på lang sikt (utover et 10-års perspektiv) for å oppnå ønskede resultater
Why do researchers decline reviewer invitations? Response to the editorial ‘The Peer Review Process: Growing Problem in Recruiting Qualified Reviewers’
Get PDFIn the editorial ‘The Peer Review Process: Growing Problem in Recruiting Qualified Reviewers’, Editor-in-Chief of Acta Odontologica Scandinavica, Palle Holmstrup, addresses the increasing difficulty in recruiting qualified reviewers for the publishing process. He cites statistics indicating that over the past decade, editors at his journal have had to double the average number of reviewer invitations to secure enough acceptances. This situation results in additional work for handling editors, prolongs the peer-review process, and causes frustration for authors awaiting the conclusion of the editorial process. Holmstrup appeals to researchers’ conscience by emphasizing the principle of reciprocity in the reviewing process
Suggestion for national monitoring of seabed litter in Norway - M-2897|2024
No full textSource at https://www.hi.no/hi.Flere typer utstyr kan brukes for å overvåke søppel på havbunnen, inkludert droppkameraer, ROV-er (Remotely Operated Vehicles) og AUV-er (Autonomous Underwater Vehicles). Droppkamera og videorigger er kostnadseffektive og kan dekke store områder, men har begrenset fleksibilitet for nærmere undersøkelser. ROV-er gir større frihet og kan stoppe for nærmere inspeksjon, men er dyrere og vær- og strømavhengige. AUV-er er effektive for å dekke store områder raskt og kan ta høyoppløselige bilder, men er kostbare og teknisk krevende.
Søppel registreres i felt med kvalitetssjekk av usikre bestemmelser senere etter tokt. Objektdeteksjon ved hjelp av kunstig intelligens (KI) vil være effektivt når algoritmer for objektdeteksjon er utviklet. Med de lave tetthetene av søppel som forekommer i norske farvann, anses i dag den mest kostnadseffektive metoden å være direkte registrering i felt, noe som krever at trent personell utfører feltarbeidet. Når relevant KI-teknologi er utviklet, vil søppel kunne registreres automatisk med kvalitetskontroll av trent personell.
Søppel på havbunnen er generelt mest utbredt nær kysten på 200–300 meters dyp med plast, og da særlig fiskerirelatert plast (som tau og garn), som dominerende kategori. I tillegg er det på visse steder oppsamling av søppel på dypt vann så som i marine daler, trau og marine gjel. Mareano-programmet har kartlagt søppel siden 2006, og observasjoner viser at plast utgjør 64 % av alt søppel, og at tettheten er høyest mellom 400–1000 meters dyp. Elver bidrar betydelig til plastforurensning, spesielt plast fra jordbruk som plastfolie rundt høyballer, som i stor grad havner nær elvemunningene. Gjennom det årlige økotoktet registrerer Havforskningsinstituttet også søppel som bifangst i bunntrål i Barentshavet og Nordsjøen. Dette søppelet utgjøres hovedsakelig av fiskerirelatert avfall, som garn og nylontau. Det er knyttet usikkerhet til omfanget av tapte fiskeredskaper, "spøkelsesfiske", i kystnære områder, men data fra Fiskeridirektoratet og Havforskningsinstituttet tilsier at omfanget er betydelig.
Styrkeanalyser av data fra Mareano-programmet viser at for å oppdage en 30 % endring i søppelmengde med 80 % sikkerhet, kreves det ved overvåking hvert 5. år av mellom 47 og 132 transekt per havområde, avhengig av terskelverdier og søppeltetthet. Det anbefales å overvåke med videotransekt hvert femte år på et tilstrekkelig antall stasjoner (≥3 søppelbiter) i hvert område (totalt 55 transekt per område). For å kunne påvise endringer i søppelmengde, anbefales det å gjenbesøke transekt fra Mareano-programmet som allerede har registrert tre eller flere søppelenheter. I områder hvor tidligere kartlegging av søppel mangler, anbefales det å supplere med stasjoner i områder med høy sannsynlighet for søppelansamling. En ny styrkeanalyse bør gjennomføres etter første datainnsamling for justering av overvåkingsopplegg. For hvert forslag må det beregnes noe høyere kostnader første år, da man må forvente større innsats for å identifisere transekt med ≥3 søppelbiter. I forslaget til overvåkinsprogram er kostnad estimert for en runde med undersøkelser. Vi anbefaler 5 års mellomrom mellom hver runde.Several types of equipment can be used to monitor litter on the seabed, including drop cameras, ROVs (Remotely
Operated Vehicles), and AUVs (Autonomous Underwater Vehicles). Drop cameras and video rigs are cost-effective
and can cover large areas but have limited flexibility for detailed inspections. ROVs offer greater freedom and can
stop for closer examination but are more expensive and dependent on weather and currents. AUVs are efficient for
covering large areas quickly and can capture high-resolution images, but they are costly and technically demanding. Litter is recorded in the field with quality checks of uncertain identifications conducted later after field surveys. Object
detection using artificial intelligence (AI) will become efficient once detection algorithms are developed. Given the
low densities of litter in Norwegian waters, the most cost-effective method today is direct field recording, requiring
trained personnel to perform the work. When relevant AI technology is developed, litter could be automatically
detected with quality control by trained personnel. Seabed litter is generally most prevalent near the coast and at intermediate depths, with plastic, particularly fishingrelated plastic (such as ropes and nets), being the dominant category. Other accumulation areas include marine
valleys, troughs, and canyons. The Mareano program has mapped litter since 2006, showing that plastic accounts
for 64% of all litter, with the highest density at depths between 400–1000 meters. Rivers significantly contribute to
plastic pollution, especially plastic from agriculture, such as silage wrap, which often accumulates near river mouths. Through annual ecosystem surveys, the Institute of Marine Research also records litter as bycatch in bottom trawls
in the Barents Sea and North Sea. This litter mainly consists of fishing-related waste, such as nets and nylon ropes. The extent of lost fishing gear and "ghost fishing" in coastal areas is uncertain, but data from the Directorate of
Fisheries and the Institute of Marine Research indicate it is significant. Power analyses of data from the Mareano program show that detecting a 30% change in litter quantities with 80%certainty requires 47–132 transects per sea area annually, depending on threshold values and litter density. It is
recommended to monitor with video transects every five years at a sufficient number of stations (≥3 litter items) in
each area (a total of 55 transects per area). To detect changes in litter quantities, transects from the Mareano
program that have already recorded three or more litter items should be revisited. In areas where previous litter
mapping is lacking, additional stations should be added in locations with a high likelihood of litter accumulation. Anew power analysis should be conducted after the initial data collection to adjust the monitoring approach. For each
proposal, slightly higher costs are expected in the first year due to the effort needed to identify transects with ≥3 litter
items. For the estimated cost of monitoring program, we have given estimates per round of investigation. We
recommend new investigation every 5 years
Investigating the quantile-varying impact of energy policy uncertainty on mineral-driven renewable energy production: Evidence from Germany
Get PDFThe recent climate crisis has solemnized renewable energy for emission reduction and halting an environmental
disaster globally. However, renewable energy technologies are severely dependent on the import of critical
minerals and elements that can be severely impacted by energy policy uncertainties (ERPU). Germany highly
relies on renewables for powering economic activities and is the largest renewable energy producer in the European Union (EU) countries. However, German renewable energy generation heavily depends on imported
critical minerals and rare earths. Therefore, the primary objective of this study is to examine how ERPU drives
mineral-dependent renewable energy production in Germany from 1996M1–2018M12 by utilizing the Quantile-on-Quantile (QQR) regression method as a baseline estimator. Quantile regression (QR) is also used to check for
robustness. Empirical outcomes denote that ERPU can pose significant risks for cobalt-graphite-, copper-, nickel and rare earths-driven renewable energy production. However, the harmful impact of ERPU on nickel-driven
renewable energy production is relatively limited. In addition, QR results majorly confirm the QQR results.
Therefore, policymakers in Germany should focus on developing alternative energy policies to ease the undesired
impacts of rising ERPU on renewable energy production
Innovative nanofiber systems as antimicrobial wound dressings: The impact of biopolymers, liposomes and varying nanofiber complexities on chloramphenicol delivery
Get PDFChronic wounds are a worldwide burden, affecting both patients and healthcare systems and therefore novel treatment options are needed. The development of novel wound dressings with an antimicrobial effect can help to address the problem of chronic wound healing and chronic wound infections. We propose the use of nanofibers, which are scaffolds with a fibrous structure fabricated via electrospinning from polymers and present a promising approach as platforms for antibiotic delivery. We aimed to use natural polymers with biological activity as material for nanofiber fabrication to manufacture a multifunctional dressing with properties that can aid the wound healing process. We chose the natural polymers beta-glucan, chitosan and pectin, which have the immunomodulating, antimicrobial and high swelling abilities, respectively. In this work, we developed chloramphenicol (CAM) containing nanofibers from the chosen natural polymers using electrospinning. To achieve a tailored delivery system of the chosen model antibiotic CAM, we incorporated CAM-loaded liposomes into the nanofibers. For fabrication of the nanofibers we used to different electrospinning methods, wire- and coaxial electrospinning, which utilized environmentally friendly solvents. We showed successful fabrication of nanofibers containing beta-glucan, chitosan together with co-polymers and CAM and successful electrospinning of CAM-loaded liposomes into pectin-nanofibers. Both nanofibers delivered CAM, maintained the antimicrobial activity of CAM against Escherichia coli and Staphylococcus aureus, and showed a high swelling index and anti-inflammatory activity in murine macrophages, all of which are desirable characteristics for a wound dressing. The addition of CAM-loaded liposomes resulted in a slower CAM release, one of the main aims in this work. We furthermore fabricated nanofibers containing all ingredients within one nanofiber by fabricating core-shell nanofibers. This dressing combined the advantages of the two single previous dressings: a high swelling index, antimicrobial properties and a slower CAM release, showing potential as antimicrobial wound dressing for delivery of antibiotics.Kroniske sår er et stort problem som påvirker pasienter og helsevesener over hele verden, og skaper derfor et stort behov for nye behandlingsmetoder. Utvikling av nye sårbandasjer med en antimikrobiell effekt kan fremme sårtilheling og motvirke eller bekjempe infeksjoner i kroniske sår. Vi foreslår bruk av nanofiber, et nett med en fiberstruktur laget ved hjelp av fiberspinning av polymerer, som er lovende for spesifikk levering av antibiotika, som sårbandasje. Vi hadde som mål å bruke naturlige polymerer med biologiske aktiviteter som materialer for fremstilling av nanofiberne for å oppnå en multifunskjonell sårbandasje som kan forbedre sårhtilhelingsprosessen. Vi valgte de naturlige polymerene betaglukan, som øker immunforsvaret, kitosan, som er antimikrobielt, og pektin, som har en høy svellekapasitet og kan ta opp sårvæske. Under dette arbeidet utviklet vi nanofibre med kloramfenikol (CAM) ved bruk av de utvalgte polymerene og fiberspinning. For å oppnå et spesifikt leveringssystem av CAM, inkorporerte vi CAM-liposomer i nanofibrene. Under framstilling av nanofibre brukte vi to ulike metoder: tråd-fiberspinning og koaksial-fiberspinning sammen med miljøvennlige løsemidler. Vi framstilte nanofiber som innehold betaglukan og kitosan sammen med kopolymerer og CAM og i tillegg klarte vi å lage nanofiber med CAM-liposomer i pektin ved hjelp av fiberspinning. Begge nanofibrene fungerte som leveringssystemer for CAM, opprettholdt den antimikrobielle aktiviteten mot Escherichia coli og Staphylococcus aureus, viste en høy absorpsjonskapasitet av simulert sårvæske og en anti-inflammatorisk effekt. Alle disse er ønskelige egenskaper for en sårbandasje. Ved å tilsette CAM-liposomer til nanofibre klarte vi å oppnå en mer langsom frigjøring av CAM, et av hovedmålene i prosjektet. For å utnytte de gode egenskapene funnet i de tillagede leveringssystemene, produserte vi en formulering som inneholdt alle ingrediensene i en og samme nanofiber, som da hadde en kjerne-mantel struktur. Denne sårbandasjen kombinerte de ønskelige fordelene av de to enkelte sårbandasjene, nemlig høy svellingskapasitet, antimikrobielle egenskaper og en langsom CAM-frigjøring, noe som gjør disse nanofibrene lovende som antimikrobielle sårbandasjer for levering av antibiotika
MaxCutPool: differentiable feature-aware Maxcut for pooling in graph neural networks
Get PDFSubmitted to the International Conference on Learning Representations (ICLR), april 2025.We propose a novel approach to compute the MAXCUT in attributed graphs, i.e., graphs with features associated with nodes and edges. Our approach works well on any kind of graph topology and can find solutions that jointly optimize the MAXCUT along with other objectives. Based on the obtained MAXCUT partition, we implement a hierarchical graph pooling layer for Graph Neural Networks, which is sparse, trainable end-to-end, and particularly suitable for downstream tasks on heterophilic graphs
Importance of substrate type and its constituents on overall performance of microbial fuel cells
Get PDFMicrobial Fuel Cells (MFC) have emerged as a potential wastewater treatment technology that utilizes metabolic processes of microorganisms present in the wastewater to disintegrate organic substrates and harness direct electricity. This paper reviews the potential of different wastewater types as a suitable substrate for microbial activities in MFCs. Substrate composition (carbon source, nutrient content and inhibitory compounds) directly affects the microbial growth, wastewater treatment potential, electron transfer rate and power harvested. Readily biodegradable substrates such as acetate and glucose promote microbial metabolism and electron transport, thus resulting in enhanced power generation. Substrates such as municipal or agricultural wastewater that constitute both simple and complex organic matter require longer breakdown durations but can provide MFCs with long term operational stability. On the other hand, substrates such as leachate from landfills, mining wastewater etc. are generally unsuitable for regular operations. The paper discusses the challenges such as suitability of various wastewaters, unpredictability of composition, presence of inhibitory chemicals like heavy metals or toxic organics that can subdue the microbial activity and reduce efficiency of a MFC reactor. It aims to identify the relationship between the substrate characteristics and MFC performance in order to select the substrate for achieving optimal output from MFC technology. The suitable substrates that exhibit optimal performance in terms of current and power output, coulombic efficiency, and reduction in chemical oxygen demand are identified