Alle publikasjoner ved avdelingen

• Njå, Ola; Nielsen, Arne Eirik; Holone, Harald; Joshi, Suhas Govind & Herstad, Jo (2026). Closing Time: A Theatre of Augmentation. I Havinen, Laura & Hägglund, Susanne (Red.), NordiCHI '26: Proceedings of the 14th Nordic Conference on Human-Computer Interaction. Association for Computing Machinery (ACM). ISSN 9798400723735. s. 1–10. doi: 10.1145/3821402.3830169. Fulltekst i vitenarkiv
• Haugen, Øystein; Andersen, Martin Arthur; Klikovits, Stefan; Bordeleau, Francis; Mertens, Joost & Denil, Joachim (2026). Evolving the DarTwin evolution language. Innovations in Systems and Software Engineering. ISSN 1614-5046. 23(1). doi: 10.1007/s11334-026-00647-3. Fulltekst i vitenarkiv
• Herrebrøden, Marte Karoline; Marentakis, Georgios & Nordahl-Hansen, Anders (2026). Educational professionals' perspectives on hearing protection devices for autistic students. Research in Autism Spectrum Disorders. ISSN 1750-9467. 139. Fulltekst i vitenarkiv Vis sammendrag

  Background Autistic students who are sensitive to auditory stimuli often experience sensory overload in school due to insufficient adaptation to auditory barriers. Hearing protection devices (HPD) may serve as a valuable support by reducing auditory input and facilitating participation in learning activities and overall well-being. This study aims to explore the practices, perceptions, and competence of employees in school and the Educational and Psychological Service regarding the use of HPD for autistic students across different educational settings. Furthermore, the study examines the need for comprehensive guidelines and the implications for strengthening inclusive educational practices. Method Data were collected through a digital questionnaire completed by 608 school professionals and employees of the Educational and Psychological Service, all working with autistic students. Descriptive and inferential statistical analyses were used to examine patterns and predictors of HPD use and related outcome measures. Results HPD are widely used in educational settings, and educational professionals perceive their use as potentially supporting autistic students' participation and well-being across school contexts. The findings also reveal potential barriers, including a lack of standardized routines and practices, as well as variability in awareness and competence related to HPD use among educational professionals. Conclusions This study highlights the considerable potential of HPD to support autistic students in everyday school contexts. However, a gap remains between their perceived benefits and their practical implementation in inclusive educational settings. This discrepancy is likely related to variability in professional competence, as well as the absence of consistent routines and clearly established guidelines.

• Nizamani, Abdul Ghafoor; Vu, Dung Thi; Nizamani, Rafeeque Ahmed; Aliyari, Mostafa & Li, Cunrong (2026). Physics Guided AI for Urban Heat Mitigation Planning. Research Square. ISSN 2693-5015. doi: 10.21203/rs.3.rs-10949253/v1. Fulltekst i vitenarkiv Vis sammendrag

  Abstract Urban heat-mitigation planning requires ways to compare intervention locations when only part of the candidate set can be treated. We developed a time-dependent physics-guided model of land surface temperature for Oslo using Landsat, ERA5, and OpenStreetMap data, and used the fitted model to screen green-roof surface-property scenarios. Across the three fitted Landsat scenes, the Physics informed neural network achieved RMSE values of 0.2549–0.3959 °C, MAE values of 0.1961–0.3043 °C, and R 2 values of 0.9638–0.9676, although some data-driven models achieved lower fitted errors. Prescribing higher NDVI and emissivity at hotspot building pixels produced maximum local modelled LST reductions of 3.49–4.28 °C, while most affected pixels reported smaller responses. Under partial deployment, mean response depended on acquisition date, spatial group, and prioritisation strategy; no single proximity-based rule produced the more negative mean response across all conditions. The analysis supports comparative thermal screening rather than prediction of complete green-roof performance.

• Qassi, Abdelaziz; Hallac, Ibrahim Riza & Ogul, Hasan (2026). Investigating the Relevance of NLP Benchmarks for Automated Essay Grading: A Multi-trait Evaluation of Pre-trained Transformers. Lecture Notes in Networks and Systems. ISSN 2367-3370. 1754, s. 137–151. doi: 10.1007/978-3-032-13153-9_11. Fulltekst i vitenarkiv
• Tran, Thi Thuy Chau; Nguyen, Hoa Thi; Magnusson, Lars Vidar & Olsson, Roland (2026). How Well Do Automatically Evolved Neurons Perform? A Comparative Study in Human Activity Recognition. ESREL 2026. doi: https:/cmsweb.com.sg/rps2prod/esrel2026/epro/pdf/esrel26-p26301.pdf. Fulltekst i vitenarkiv Vis sammendrag

  Neuroevolution, a subfield of AI leveraging evolutionary algorithms, offers an effective method for automatically generating data-adaptive equations that optimize fit for specific datasets while minimizing human design bias. This study investigates the efficacy of the WISDM-ARN, a specialized Adaptive Recurrent Neuron (ARN) synthesized via the Automatic Design of Algorithms through Evolution (ADATE) system, for Human Activity Recognition (HAR). In safety-critical applications, such as healthcare monitoring and industrial safety, the reliability of locomotion classification is paramount to prevent system failures and ensure user protection. We benchmark the evolved WISDM-ARN against three established architectures: Simple RNN, Long Short-Term Memory (LSTM), and Transformers. Experimental results demonstrate that the WISDM-ARN successfully adapts to the unique temporal characteristics of gyroscope data, achieving a testing accuracy of 97\%, which is comparable to the baseline models with resulting 70\%, 95\%, and 92\%, respectively. These findings highlight the potential of neuroevolutionary synthesis as a dependable alternative to manually engineered architectures. By integrating neuroevolution with modern deep learning, the ADATE-generated neuron represents a significant step toward automated reliable neural modeling for human activity recognition tasks and safety-enhanced time-series analysis.

• Tooranian, Abdolrahim & Sturm, Bob L. T. (2026). Audio-Side Reinforcement Learning for Symbolic Music Style Adaptation: Extending SMART with Composite Rewards. I Coelho, Guilherme (Red.), Proceedings of International Conference on AI and Musical Creativity. AI Music Creativity (AIMC). doi: 10.5281/ZENODO.22254724. Fulltekst i vitenarkiv
• Safhi, Amine el Mahdi; Mejjad, Nezha; Ghribi, Mounir; Pilehvar, Shima & Kioumarsi, Mahdi (2026). Benchmarking sustainable blue economy in Canada and Norway: A comparative KPI-based assessment. Marine Policy. ISSN 0308-597X. 195. doi: 10.1016/j.marpol.2026.107298. Fulltekst i vitenarkiv
• Stigberg, Susanne Koch; Carcani, Klaudia; Joshi, Suhas Govind & Bratteteig, Tone (2026). Participatory Design Meets Artificial Intelligence. Co-Imagining Mutual Learning of AI Technologies and Designing with AI Tools. ID&A Interaction design & architecture(s). ISSN 1826-9745. s. 5–14. doi: 10.55612/s-5002-068-001psi. Fulltekst i vitenarkiv Vis sammendrag

  This Special Issue (N. 68) emerged as a step toward broadening the discussion about the relationship between Participatory Design (PD) and Artificial Intelligence (AI). The original PD projects were motivated by the introduction of computer technologies into industrial workplaces, where they threatened to impoverish or take over jobs. Just like computers were a threat to work and workers in the 1970s and 80s, AI is seen as a threat to workers—and societies—today. Building on a workshop held at NordiCHI 2024 [1], we wanted to invite a wider community of researchers and practitioners to critically reflect on how PD can question, shape, and reimagine AI technologies and practices. The notion of AI was used for the first time in 1955 referring to a machine that can simulate human intelligence [2]. Webster [3] argues that current AI systems that are based on Machine Learning (ML) and Large Language Models (LLMs) can “be said to fulfill the original 1955 definition of AI for the first time” (p. 8). Today, AI systems are widespread and deployed rapidly across domains of considerable social significance— in healthcare, education, employment, criminal justice, and many others—without appropriate safeguards or accountability structures in place [4,5]. AI can replace human performance of tasks and “cooperate” with humans [6]. Similar to the expert systems of the second wave of AI [3], the current third wave of AI advocates AI that improves human capabilities rather than replacing them. The initiative for Human-Centered AI (HCAI) argues that AI systems should be explainable, comprehensible, useful and usable [7]. To achieve this Shneiderman [8] argues that user-centered and participatory methods engaging stakeholders in design is needed. A study of research on HCAI by Capel and Brereton [9] found that almost half the research papers could be categorized as human-centered approaches to design of AI, but very few of these were based on PD [10]. This Special Issue aims to change this.

• Linnes, Cathrine; Ronzoni, Giulio; Agrusa, Jerome; George, Babu & Lema, Joseph (2026). From Word-of-Mouth to Word-of-Machine: Examining AI-Generated Travel Recommendations Through the Word-of-Machine (WOMa) Perspective. Tourism and Hospitality. 7(9), s. 1–25. doi: 10.3390/tourhosp7090293. Fulltekst i vitenarkiv Vis sammendrag

  Generative artificial intelligence (AI) is increasingly used in travel planning, yet limited research has examined how travelers evaluate AI-generated travel recommendations as an information source. This study examines travelers’ evaluations and behavioral intentions toward such recommendations, extending and contextualizing the Word-of-Machine (WOMa) perspective within tourism using an Information Adoption Model (IAM)-informed framework. An online survey of 485 U.S. respondents with prior experience using AI for travel planning measured trust, perceived usefulness, perceived ease of use, information presentation, and behavioral intention. Respondents generally evaluated AI-generated travel recommendations favorably, and all four predictors were positively associated with behavioral intention. The regression model explained 84.5% of the variance in behavioral intention (R2 = 0.845), while demographic differences were observed, particularly across household-income groups. The study contributes to tourism research by extending the WOMa perspective to AI-generated travel recommendations and examining the factors associated with travelers’ willingness to use them. Practically, the findings highlight the importance of useful, trustworthy, clearly presented, and easy-to-use AI recommendations, while maintaining opportunities for verification and human judgment.

• Linnes, Cathrine; Ronzoni, Giulio; Lema, Joseph; George, Babu & Agrusa, Jerome (2026). Public Trust in Generative AI: Risk Perceptions, Regulatory Safeguards, and the Acceptance of Deepfake Technology. Information. 17(8), s. 1–23. doi: 10.3390/info17080757. Fulltekst i vitenarkiv Vis sammendrag

  Generative artificial intelligence (GenAI) now produces synthetic text, images, audio, and video at a quality and cost that place convincing synthetic fabrication within reach of non-specialist users. Deepfakes, synthetic media that alter a person’s appearance, voice, or behavior through machine learning (ML), are one of the most contested applications of this capability, yet public willingness to accept them under regulation remains less understood. This study examines how perceived benefits, perceived risks, privacy concerns, and demographic characteristics relate to trust in deepfake technology under regulatory safeguards. Survey data from 924 respondents from several countries were analyzed using descriptive statistics, independent-samples t-tests, analysis of variance, multiple regression, and thematic analysis of open-ended responses. Respondents recognized the potential benefits of deepfake technology for digital content creation and education while expressing widespread concern about misinformation, privacy violations, and criminal misuse. When respondents evaluated deepfake technology under an assumed privacy-protecting regulatory scenario, perceived risks did not independently predict trust, while perceived benefits were the strongest predictors. The findings indicate that institutional confidence may contribute to public acceptance of beneficial applications of generative AI, although the cross-sectional design does not establish a causal effect of regulation.

• Linnes, Cathrine; Ronzoni, Giulio; Moghadam, Reza Saneei; Lema, Joseph; George, Babu & Agrusa, Jerome (2026). Forecasting Visitor Activity in a Historic Urban District: A Machine-Learning Framework for Destination Management. Tourism and Hospitality. 7(8), s. 1–27. doi: 10.3390/tourhosp7080234. Fulltekst i vitenarkiv Vis sammendrag

  Historic urban districts face growing challenges in balancing visitor activity with the needs of residents, local businesses, accessibility, and mobility. This study investigates whether short-term visitor forecasts can support destination management and planning in these complex environments. Unlike most visitor-forecasting studies, which focus on large cities and major tourist destinations, this study examines forecasting in a small historic urban district where visitor activity is more variable and management resources are limited. Using hourly pedestrian counts, weather data, and temporal data from the historic urban district of Fredrikstad, Norway, this research forecasts pedestrian activity up to 24 h in advance. The sensors record all pedestrians and do not distinguish tourists from residents or other users. Consequently, pedestrian counts are treated as an operational proxy for overall visitor activity at this heritage destination. Forecasting performance was evaluated using statistical, machine-learning, deep-learning, and benchmark forecasting models. Random Forest achieved the strongest overall forecasting performance (MAE = 209.52; RMSE = 357.18), outperforming the seasonal naïve benchmark (MAE = 247.44; RMSE = 421.70). Random Forest and GRU both outperformed the seasonal naïve benchmark on MAE, demonstrating the value of incorporating weather, temporal, and sensor variables into short-term forecasting. The findings suggest that short-term visitor forecasts may support operational planning, mobility management, service coordination, and other day-to-day destination management decisions. By anticipating periods of increased visitor activity, destination managers, local authorities, and government officials will be better able to allocate resources and coordinate services. This study demonstrates the potential of predictive analytics and sensor technology to support visitor management and operational planning in historic urban districts.

• Alazawi, Amer Abdullah; Alchalabi, Oday Qusay Abdulqader; Alhafody, Ashraf Ibrahim & Nizamani, Abdul Ghafoor (2026). Aesthetics of Interruption: Professional Disconnection and Façade Transformation in Post-2017 Mosul Residential Design. Architecture. 6(3), s. 103–103. doi: 10.3390/architecture6030103. Fulltekst i vitenarkiv Vis sammendrag

  Post-conflict reconstruction research has examined façade materiality and symbolism, yet the process conditions under which aesthetic specifications are systematically overridden during construction remain neglected. This study investigates why designed architectural aesthetics fail to survive implementation in post-2017 Mosul, Iraq. A mixed-methods design combined formal visual analysis of 12 recently completed residential façades with a structured survey of 45 practicing architects. Survey data reveal that designers are excluded from construction supervision in 76% of projects and that clients intervene in material and color selection in 70% of cases. Visual analysis identifies a sophisticated design language—orthogonal massing articulated through contrasting materials—that is rarely realized in built form. Where designers retain supervisory authority, projects most consistently achieve material–form coherence. The study advances the concept of an aesthetics of interruption (the systematic degradation of designed form–material relationships through the fragmentation of professional authority during construction). Exclusion produces four distinct pathologies: material substitution, execution degradation, language override, and ornamental hollowing. The findings demonstrate that aesthetic degradation in post-conflict reconstruction stems not from design incapacity but from broken process structures. Safeguarding architectural quality requires contractual frameworks mandating designer supervision and material-substitution protocols that protect design intent.

• Nizamani, Abdul Ghafoor; Torgersen, Geir & Edvardsen, Magdalena (2026). Building-Integrated Green Walls in Cold Climate: Operational Requirements and Maintainability. SSRN Electronic Journal. doi: 10.2139/ssrn.7518015. Fulltekst i vitenarkiv Vis sammendrag

  Green walls are increasingly integrated into building envelopes, but their long-term operation depends on maintaining both living and technical components. This study examines maintenance practices, operating-cost evidence and post-installation responsibility for building-integrated green-wall systems in Norway. An exploratory multiple-case study used six built green-wall projects, four semi-structured interviews with professionals in maintenance, supply, municipal practice and research, and case-specific documentation supplemented by relevant standards, certification and policy material. Operational requirements varied substantially between systems and settings. Exposed outdoor systems followed a marked seasonal cycle involving irrigation shutdown, spring recommissioning, vegetation maintenance and system inspection, whereas the indoor case operated year-round. Across the documented cases and interview examples, maintenance frequencies ranged from approximately every two weeks to one or two scheduled visits per year. Wall-specific cost data were limited and heterogeneous, and operating figures could not be compared reliably per square metre because service scope, access requirements, system complexity, scale and seasonal workload differed between projects. Interviews further indicated that practical operating knowledge often remained with suppliers and specialist providers while longer-term responsibility shifted towards owners and building operators. Maintainable access to technical components, usable operation and maintenance documentation, monitoring arrangements and post-installation service were therefore important during the transition to routine operation. The findings indicate that maintainability is partly determined by decisions made during design and procurement. Irrigation design, seasonal operation, maintenance access, expected operating requirements and transfer of technical knowledge should therefore be considered alongside installation requirements.

• Emblem, Anne Wenche & Borgersen, Trond Arne (2026). Differentiated New Housing: Implications for Market Equilibrium and the Effect of Shocks. Nordic Journal of Surveying and Real Estate Research. ISSN 1459-5877. 19(2), s. 7–29. doi: 10.30672/njsr.177940. Fulltekst i vitenarkiv Vis sammendrag

  It is widely established that housing supply is differentiated across multiple dimensions, and a substantial body of empirical research has explored how various housing and locational characteristics influence prices. While the responsiveness of housing supply to price changes is of both theoretical and practical importance, the supply side of the housing market remains relatively underexplored. This paper revisits core economic insights on product differentiation and examines their implications for housing prices and supply elasticity. We question the common assumption that increased housing supply necessarily dampens prices. Instead, we propose that higher supply, when accompanied by increased differentiation, may indeed reduce the elasticity of supply, making supply less responsive to rising demand. By means of a simple theoretical model, we show that when housing supply is more differentiated, demand shocks tend to produce sharper price increases and more limited changes in housing quantity, compared to markets with more uniform supply. These insights are highly relevant to housing market debates and should be considered, despite the methodological challenges of estimating price effects from unobservable housing attributes.

• Fedushko, Solomiia; Syerov, Yuriy; Mochurad, Bohdan; Kravets, Ruslan & Roddvik, Irina Nikolskaja (2026). Sentiment Analysis of Music Discourse in Social Media Networks: A Case Study of Reddit Posts. I Starchon, Peter; Sedov, Yuriy; Gregus, Michal & Ivanochko, Iryna (Red.), Data-centric Business and Applications. Springer Cham. ISSN 9783032164285. Fulltekst i vitenarkiv
• Sierov, Yuriy; Fedushko, Solomiia; Kravets, Ruslan; Karpov, Viktor & Roddvik, Irina Nikolskaja (2026). Artificial Intelligence Approaches for Automated and Optimized Ad Design: A Comparative and Multimodal Perspective. I Starchon, Peter; Sedov, Yuriy; Gregus, Michal & Ivanochko, Iryna (Red.), Data-centric Business and Applications. Springer Cham. ISSN 9783032164285. Fulltekst i vitenarkiv
• Fedushko, Solomiia; Mykhalchyshyn, Natalia; Hoshovska,, Oksana; Trevoho, Olena & Roddvik, Irina Nikolskaja (2026). Comparative Governance: Power, Success, and Leadership in Public Administration. I Starchon, Peter; Sedov, Yuriy; Gregus, Michal & Ivanochko, Iryna (Red.), Data-centric Business and Applications. Springer Cham. ISSN 9783032164285. Fulltekst i vitenarkiv
• Baranyuk, Andriy; Mykhalchyshyna, Natalia; Syerov, Yuriy; Roddvik, Irina Nikolskaja & Korzh, Halyna (2026). Causal Modellimg of socio-economic Relations: Interdepenceses between Needs, Resources, Technology, and Institutional Transformations. I Starchon, Peter; Sedov, Yuriy; Gregus, Michal & Ivanochko, Iryna (Red.), Data-centric Business and Applications. Springer Cham. ISSN 9783032164285. Fulltekst i vitenarkiv
• George, Babu; Lema, Joseph; Linnes, Cathrine & Agrusa, Jerome (2026). Beyond the Departure Gate: De-Anchoring Tourism Through Strategic Foresight. Tourism and Hospitality. 7(8), s. 1–12. doi: 10.3390/tourhosp7080242. Fulltekst i vitenarkiv
• Sarangi, Rachita R.; Ray, Prakash K.; Kishor, Nand; Mohanty, Soumya R. & Mohanty, Asit (2026). A Noise-Aware Framework for Reliable Detection of High-Impedance Faults in DC Microgrids. IEEE Transactions on Instrumentation and Measurement. ISSN 0018-9456. doi: 10.1109/TIM.2026.3720848. Fulltekst i vitenarkiv
• Nizamani, Abdul Ghafoor; Pregnolato, Maria & Torgersen, Geir (2026). Operational stormwater performance of building integrated green walls | Research Square. Research Square. ISSN 2693-5015. doi: 10.21203/rs.3.rs-10841067/v1. Fulltekst i vitenarkiv Vis sammendrag

  Green walls offer a building-integrated option for managing roof runoff in dense urban areas, but their operational retention depends on wall architecture, wall-to-roof ratio and pre-storm storage. We combined full-scale experiments on tank-type and porous-substrate walls with a parsimonious two-...

• Ramstad, Ola Huse; Sandvig, Axel; Nichele, Stefano & Sandvig, Ioanna (2026). Influence of neural network bursts on functional development. bioRxiv. ISSN 2692-8205. doi: 10.1101/2025.01.21.633559. Fulltekst i vitenarkiv Vis sammendrag

  Network bursts or synchronized burst events are a typical activity seen in most in vitro neural networks. Network bursts arise early in development and as networks mature, activity becomes dominated by bursts propagating across the entirety of the network. The reason for this developmental plateau in vitro is unknown, but to bypass it would confer a significant advantage in the use of in vitro networks for computation. As most neurons in a network participate in network bursts, burst onset supersedes any ongoing activity thereby placing a limit on short term computations equal to the recovery period between bursts. By assessing 521 multielectrode array recordings from day 4-39 in vitro we find that network bursts influence the connectivity, but this change is only weakly associated with the origin of the network bursts. The impacts of bursts on functional integration and segregation in neural networks are discussed along with approaches to mitigate the development and propagation of network bursts in vitro. Additionally, we hypothesize that burst initiation zones or pacemakers are viable targets for stimulation for computation in the context of control and reservoir.

• Driouich, Anas; Goudarzi, Ahmad; Santoro, Francesco; Dalconi, Maria Chiara; Dau, Massimo & Bettanini, Carlo [Vis alle 8 forfattere av denne artikkelen] (2026). Role of amorphous phase, temperature fluctuations, and vacuum, on the performance of alkali-activated lunar regolith simulants. Materials and Structures. ISSN 1359-5997. 59(7). doi: 10.1617/s11527-026-03243-w. Fulltekst i vitenarkiv
• Pedro, Kringen,; Alessio, Gallucci,; Elisabeth, Hildt,; Johansen, Frode Ramstad; Magnus, Westerlund, & Christa, Tigerstedt, (2026). Using LLMs without trust: the adoption–trust paradox as institutional governance failure in academic ecosystems | AI & SOCIETY | Springer Nature Link. AI & Society: Knowledge, Culture and Communication. ISSN 0951-5666. doi: 10.1007/s00146-026-03323-z. Fulltekst i vitenarkiv Vis sammendrag

  Large language models have rapidly made their way into higher education. Still, the integration of these tools remains contested. This study presents the results from a survey of 80 higher education professionals, most of them in Nordic countries, examining LLM adoption, trust, academic integrity, and institutional preparedness. The picture that came back is contradictory. In this self-selected sample, 82.5% already use LLMs in their teaching work. At the same time, they do not really trust what these tools produce. Trust in LLM outputs scored only 2.45 out of 5. Concerns about privacy scored considerably higher (3.82/5) and so did concerns about academic integrity (3.42/5). The respondents are fairly confident that students use LLMs regularly (4.19/5), but they have little confidence in telling student work from LLM-generated output (2.58/5). The lowest score across all areas (mean 1.99/5) was found for use of LLMs as a tool for grading students. When asked about challenges, 47.5% named academic integrity, indicating vulnerability within institutional assessment practices rather than individual misconduct. The data are consistent with a self-reinforcing cycle. Educators in this sample report clear benefits and assume widespread student use, while trusting the outputs little, and continued use does not appear to build that trust. We call this the adoption-trust paradox. The survey did not measure institutional governance directly, but Nordic policy studies report institutional responses ranging from comprehensive AI guidelines to little more than references to existing integrity norms. Read against that background, the same pattern repeated itself across every section of the survey. High adoption, low trust, and educators left to draw the lines themselves. The sample is modest and weighted toward the social sciences and humanities, but the consistency of the pattern makes it hard to dismiss, and it supports the case for an institutional response.

• Ganesan, Manikandan; Chokkalingam, Bharatiraja; Mihet, Nicolae Lucian; Verma, Rajesh & Mihet, Nicolae Lucian (2026). A Hierarchical Real-Time Object Detection and Fine-Grained Classification framework for Autonomous Driving. IEEE Access. doi: 10.1109/ACCESS.2026.3711217. Fulltekst i vitenarkiv
• Ramya, K.; Gopalakrishnan, Janaki; Verma, Rajesh; Chokkalingam, Bharatiraja & Mihet, Nicolae Lucian (2026). Experimental investigation of a π–L EMI filter for conducted noise suppression in a dual active bridge converter. Results in Engineering (RINENG). 32. doi: 10.1016/j.rineng.2026.111795. Fulltekst i vitenarkiv
• Šobot, Aleksandar; Starc, Jasmina; Hodžić, Nezmir; Adeyemi, Idris Babatunde; Colarič-Jakše, Lea Marija & Bilić-Šobot, Diana [Vis alle 7 forfattere av denne artikkelen] (2026). Analysis of Air Quality in Three Slovenian Municipalities During the New Year Holiday Period. Applied Sciences. 16(13). doi: 10.3390/app16136597. Fulltekst i vitenarkiv
• Yadav, Pankaj Kumar; Biswal, Monalisa; Vemuganti, Hari Priya & Kishor, Nand (2026). Flexible future grid integration possibilities of electric vehicles: A review. Utilities Policy. ISSN 0957-1787. 102. doi: 10.1016/j.jup.2026.102281. Fulltekst i vitenarkiv
• Kozadajevs, Jevgenijs; Mutule, Anna; Jamal, Abdullah; Zalitis, Ivars; Petrichenko, Lubova & Kishor, Nand (2026). Space heating retrofits with rooftop photovoltaics in Norwegian buildings: tailoring financial support levels to drive localised investments in green technologies. Frontiers in Energy Research. 14. doi: 10.3389/fenrg.2026.1898478. Fulltekst i vitenarkiv Vis sammendrag

  Globally, the energy supply systems (ESSs) of buildings are increasingly integrating clean energy technologies to reduce energy costs while supporting the transition to a net-zero-emission future. In this regard, the role of energy use in buildings in the European Union (EU) has already been established. Norway also has ambitious climate goals, with improvements to energy use in buildings representing a significant part of the effort. There is consensus among Norwegian energy experts, consumers, and prosumers regarding the need for greater consumer involvement in ESSs and financial incentives to support it. However, the necessary support levels and supported actions have not been explored deeply in regard to sustainability of energy supply and flexibility, both of which are recognised as important factors for future energy systems. Therefore, we analysed the impacts of financial support relative to the living area on the sustainability and flexibility achieved by retrofitting ESSs in buildings and by considering different support priorities. The genetic algorithm (GA) was utilised to derive an optimal equipment set for each sub-scenario and perform scheduling with a hybrid (optimisation and rule-based) control for three buildings in Norway. To better evaluate the flexibility of energy use, we devised a new parameter called the energy cost flexibility improvement factor (ECFIF) . The results indicate that support focused primarily toward improvements in sustainability provide the most benefits overall, with indications that a support strategy that does not enforce equipment selection could be even more successful. For the studied Norwegian buildings, a support of 10–40 EUR/m 2 represents the highest increase in profitability while a support of 20–35 EUR/m 2 achieves the highest impact on ESS performance. However, separate determination of the support level may be necessary for high-cost and high-impact projects, such as rebuilding commonly used electric-based space heating to hot-water-circulation-based systems, to achieve the envisioned climate goals.

• Tellnes, Lars Gunnar Furelid; Svanes, Erik; Pamies, Ramon Francisco Porras & Kjøniksen, Anna-Lena (2026). Environmental sustainability of protein rich food from tunicates. Case study on Ciona aquaculture in Europe. Future Foods. 14. doi: 10.1016/j.fufo.2026.101133. Fulltekst i vitenarkiv
• Redi-Tränk, Cathrine Rismo; Dechsling, Anders; Marentakis, Georgios & Nordahl-Hansen, Anders (2026). Mapping the landscape of tests employed in diagnosing developmental auditory processing disorder in adults: A scoping review. Acta Psychologica. ISSN 0001-6918. 269. doi: 10.1016/j.actpsy.2026.107637. Fulltekst i vitenarkiv Vis sammendrag

  Purpose Auditory Processing Disorder (APD) is a common but often overlooked neurodevelopmental disorder affecting auditory cognition with psychological, physical, and social consequences. Despite affecting all ages, research on APD diagnosis primarily focuses on children, leaving adult diagnostic criteria and evidence-based tests and guidelines poorly defined. This scoping review aims to map tests used to diagnose developmental APD in adults and identify gaps in the literature. Method A systematic search was conducted across ten databases (ERIC, CINAHL, Scopus, PubPsych, PsycINFO, PubMed, Cochrane, MEDLINE, Web of Science, and EMBASE). Studies including adults with normal peripheral hearing and reporting diagnostic auditory tests for APD, without cooccurring cognitive, developmental, congenital, or neurological disorder, were included. After deduplication, 1787 records were identified, title and abstract screening yielded 163 full-text articles. Results Six studies were included, two experimental studies (n = 59 adults), and four surveys (n = 603 clinicians). Substantial inconsistency in APD test use, heterogeneity in participant characteristics, and referral criteria were revealed. Conclusion There is currently no consensus on diagnostic tests or criteria for APD in adults. The low number of studies, along with methodological variations, highlights a need for research to provide evidence for what an appropriate test battery would be. Addressing these gaps should lead to improvements in diagnostic reliability and intervention outcomes for adults with APD.

• Nizamani, Abdul Ghafoor; Pregnolato, Maria & Torgersen, Geir (2026). Assessing the Stormwater Retention Capacity of Green Walls: lab experiments and modelling. SSRN Electronic Journal. doi: 10.2139/ssrn.7268562. Fulltekst i vitenarkiv Vis sammendrag

  Vertical green infrastructure could help manage roof runoff in dense cities, but reported retention values from individual storms are difficult to transfer because wall architecture, antecedent wetness and measurement methods are rarely considered together. This study combines full-scale green wall experiments with a parsimonious event model and design-scenario analysis for roof-runoff-fed systems. A two-zone bucket model was parameterised and evaluated against observations from the Green Sponge Buildings Laboratory for two contrasting modular wall architectures: a tank-type wall (GW1) and a porous-substrate wall (GW2). The model represents wall-specific retention capacity, drainage onset and detention routing. It reproduced retained volumes with Nash–Sutcliffe efficiency ≈ 0.99, R² ≈ 0.99 and percent bias ≈ -0.9% for GW1, and Nash–Sutcliffe efficiency ≈ 0.97, R² ≈ 0.98 and percent bias ≈ +2.1% for GW2. Using Oslo 60-min intensity–duration–frequency design storms as a demonstration case, dry-start retention was governed mainly by wall storage capacity and wall-to-roof ratio. Latin Hypercube sampling (N = 500) identified the derived pre-storm storage fraction as the dominant explanatory state variable, with single-predictor R² values of 0.967 for GW1 and 0.968 for GW2. For the 10-year design storm at a wall-to-roof ratio of λ = 1.0, dry-start assumptions overstated median retention by 89 percentage points for GW1 and 30 percentage points for GW2. Green walls can therefore contribute to urban stormwater management, but their operational effectiveness depends on wall architecture, wall-to-roof ratio and antecedent wetness. Green wall performance claims and design guidance should therefore explicitly report wall architecture, wall-to-roof ratio and antecedent storage condition.

• Aliyari, Mostafa; Droguett, Enrique Lopez; Barabady, Javad & Ayele, Yonas Zewdu (2026). Impact of Input Sequence Length on TransformerBased Time Series Forecasting: Extreme Weather Sensitivity. International Conference on Natural Hazards and Infrastructure. Fulltekst i vitenarkiv
• Palanichamy, Manikandan & Lin, Gwendolyn Claire (2026). From Automation to Adaptation: The Quad Intelligence Framework for Climate-Resilient Agriculture. I Qureshi, Kashif Naseer; Ghafoor, Kayhan Zrar & Omer, Saman Muhammad (Red.), AI-Driven Farming Systems With Minimal Human Intervention. IGI Global. ISSN 9798337392974. s. 31–60. doi: 10.4018/979-8-3373-9295-0.ch002. Fulltekst i vitenarkiv Vis sammendrag

  Digital transformation in agriculture prioritized efficiency over climate adaptation, and such systems struggle when conditions shift beyond historical patterns. Current research remains fragmented while perceiving AI as pure automation and overlooking human cognition, collective networks, and ethics. This chapter introduces the Quad Intelligence Framework: Computational Intelligence (AI optimization), Human Cognitive Intelligence (adaptive judgment), Collective Intelligence (farmer networks), and Ethical Intelligence (accountability). Ten global implementations were analyzed across monitoring, early warning, and crop management, finding they favor efficiency over adaptation. Computational intelligence alone delivers short-term productivity but cannot adapt to climate uncertainty. Strategic integration of all four dimensions transforms efficiency-based systems into climate-resilient instruments by incorporating farmer's climate knowledge, collective networks, and sustainability thinking. The chapter provides roadmaps for transitioning to resilience-oriented digital agriculture.

• Nichele, Stefano; Peters, Carl-Hendrik & Reimers, Felix Simon (2026). Benchmarking the State of Networks with a Low-Cost Method Based on Reservoir Computing. I Beushausen, Hans; Kioumarsi, Mahdi & Shafei, Behrouz (Red.), Proceedings of the 2nd International Conference on Net-Zero Built Environment. Springer. ISSN 9783032276988. s. 992–1004. doi: 10.1007/978-3-032-27698-8_87. Fulltekst i vitenarkiv Vis sammendrag

  Using data from mobile network utilization in Norway, we showcase the possibility of monitoring the state of communication and mobility networks with a non-invasive, low-cost method. This method transforms the network data into a model within the framework of...

• Ihlebæk, Hanna Marie; Ihlebæk, Camilla; Knez, Rajna; Busso, Lilliana Andrea Del; Prøitz, Lin & Femdal, Ingrid [Vis alle 7 forfattere av denne artikkelen] (2026). What to do with school refusal? A scoping review of digital and non-digital interventions. Frontiers in Education. 11. doi: 10.3389/feduc.2026.1862436. Fulltekst i vitenarkiv Vis sammendrag

  Introduction: School refusal poses significant challenges to children's education and wellbeing. Despite increased public, political and scholarly attention, there remains limited systematic knowledge about available interventions, in particular digital interventions. This scoping review therefore aims to identify and describe the range and nature of treatment interventions for school refusal, with a specific focus on digital approaches. Methods: A scoping review was conducted in accordance with PRISMA-ScR guidelines. Eligible studies were empirical, peer-reviewed articles published in 1993 or later that examined interventions for school refusal among children and adolescents. Screening and data extraction were completed independently in pairs. The data was analyzed descriptively, utilizing an adapted TIDieR framework applied to intervention reporting. Results: Thirty-one studies published between 1998 and 2025 met the inclusion criteria. Research activity increased notably from 2018 onward and was dominated by quantitative designs conducted primarily in Western, high-income contexts. Most interventions were cognitive-behavioral therapy (CBT)-based, delivered individually, in groups, or as multimodal programs, sometimes combined with pharmacological treatment. CBT interventions generally reported improved school attendance, although effect sizes and durability varied. Only six studies (19%) incorporated digital components, including remote therapeutic delivery, virtual reality exposure, serious games, and online parent coaching. These interventions were feasible and acceptable but showed limited evidence of direct effects on school attendance. Outcome measures were heterogeneous, though school attendance and anxiety symptoms were most consistently reported. Conclusion: The evidence-base on interventions for school refusal is growing but remains fragmented, with an underrepresentation of digital approaches. Existing digital interventions predominantly target the individual or family rather than school environments or the interaction between the individual, their family and the school environment. Future research should explore how digital tools can support integrated and systemic school-refusal approaches that address individual, family and school-related factors over time.

• Ramya, K.; i Gopalakrishnan, Janaki; Verma, Rajesh; a Chokkalingam, Bharatiraja & Mihet-Popa, Lucian (2026). RedirectingExperimental investigation of a π–L EMI filter for conducted noise suppression in a dual active bridge converter. Results in Engineering (RINENG). Fulltekst i vitenarkiv Vis sammendrag

  High-frequency dual active bridge (DAB) converters used in electric vehicle (EV) charging and bidirectional energy conversion systems often generate severe conducted electromagnetic interference (EMI), including common-mode (CM) and differential-mode (DM) noise, which complicates compliance with automotive EMI standards. This paper proposes an impedance-based EMI filter design methodology for a high-frequency voltage-mode controlled DAB converter. A passive π–L filter topology is analytically designed to simultaneously suppress CM and DM noise while considering converter–filter interaction and stability constraints. Unlike conventional trial-and-error approaches, the proposed method enables predictable broadband attenuation through impedance-oriented analysis. A 500 W hardware prototype operating between a 60 V source and a 48 V battery is developed to validate the proposed approach under single-phase-shift control, achieving a peak efficiency of 97.6%. Conducted emissions are evaluated using a line impedance stabilization network (LISN) in accordance with CISPR 25 Class 5 standards over the frequency range of 150 kHz–30 MHz. Experimental results show that the optimized filter achieves full EMI compliance with significant attenuation margins, while maintaining high conversion efficiency. The close agreement between analytical and experimental results demonstrates the effectiveness and practical applicability of the proposed EMI-aware DAB design methodology for high-density Grid-to-Battery and Battery-to-Grid systems.

• Sánchez-Gordón, Mary; Colomo-Palacios, Ricardo; Gama, Kiev & Hubin, Aliaksandr (2026). Demystifying Knowledge Hiding in Software Testing: Insights from Practitioners. IEEE/ACM International Conference on Software Engineering: Software Engineering in Society Track (ICSE-SEIS). ISSN 2832-7608. 2026, s. 148–153. doi: 10.1145/3786581.3786937. Fulltekst i vitenarkiv Vis sammendrag

  Software development is a knowledge-intensive and collaborativeactivity. However, practitioners may sometimes deliberately conceal or withhold information, ideas, or expertise relevant to softwaretasks, even when a teammate explicitly requests such knowledge.This is known as knowledge hiding behaviour (KHBs). This study explores what knowledge is hidden in software testing. A survey was designed based on measures of work conditions, KHBs, andknowledge related to software testing. Responses from 102 software practitioners were collected and analysed. Although only13 participants reported KHBs, these should not be overlooked. Findings highlight the contextual nature of KHBs, suggesting that exploring KHBs in software teams is a promising avenue for future research.

• Chaaban, Omar; Sánchez-Gordón, Mary & Demi, Selina (2026). Perceptions of Undesirable Software Development Tasks among Computer Science Students. Applied Human Factors and Ergonomics International (AHFE International). 231, s. 88–99. doi: 10.54941/ahfe1007999. Fulltekst i vitenarkiv Vis sammendrag

  Software development projects include both technical and non-technical tasks, some of which may be perceived as undesirable by developers. While prior research has examined how developers perceive and manage such tasks, students’ perspectives remain unexplored.Therefore, this study aims to explore task undesirability for students. A survey-based approach was adopted to gather insights from 30 participants enrolled in IT-related programs at Østfold University College. The preliminary results revealed that back-end related tasks, bug fixing and writing documentation were among the most undesirable tasks, leading to negative emotional responses, such as frustration and demotivation among students. In addition, strategies such as “Just Do It” and working in groups were commonly employed to cope with these tasks. Although further research is needed to better understand support mechanisms that enhance students’ engagement and well-being, findings suggest that coping strategies arise when students’ experiences are not sufficiently supported. This provides an opportunity to design courses that enhance learning experiences by not only mitigating the negative impact of undesirable tasks but also making these tasks more desirable for students.

• Palanichamy, Manikandan (2026). SpaceX Starlink: Revolutionizing Satellite Internet Through LEO Innovation, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 299–320. doi: 10.1002/9781394359462.ch15. Fulltekst i vitenarkiv Vis sammendrag

  This chapter presents a case study on SpaceX's Starlink project and its revolutionary approach to overcoming traditional satellite communication limitations, particularly latency issues. The study explores the strategic transition from geostationary satellites to Low Earth Orbit (LEO) constellations, and how Starlink's unique engineering concepts have revolutionized satellite communication. Specifically, technologies such as inter‐satellite links and phased‐array antennas have been implemented to enhance performance. Thus, their innovative technologies have successfully compressed latency from 500–600 ms down to 20–40 ms. This work also discusses how SpaceX's cost‐effective satellite deployment via reused rockets becomes the enabler, although there are also fears about space debris and orbital congestion. The chapter also considers the implications of Starlink technology for the access to global internet and how it could help (or not) to close the digital divide.

• Palanichamy, Manikandan (2026). Huawei's Autonomous Networks: AI Innovation Amid Global Tensions, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 277–298. doi: 10.1002/9781394359462.ch14. Fulltekst i vitenarkiv Vis sammendrag

  This chapter examines Huawei's pioneering role in developing AI‐driven telecom network optimization systems, and their innovative Autonomous Driving Network (ADN) concept. It breaks down how Huawei has evolved the old school rule‐based network management into an intelligent self‐optimizing system, and able to handle various data in real time. The study takes readers from traditional systems to AI‐driven solutions that predict network failures down to an accuracy of 90%+. It also discusses numerous cybersecurity, scalability, and energy efficiency challenges. This chapter also explores the political and regulatory consequences surrounding Huawei's involvement in global telecommunications infrastructure. Security concerns and international relations issues have significantly impacted global telecommunication network development. These factors have created complex challenges due to Huawei's market dominance in the sector.

• Palanichamy, Manikandan (2026). Ericsson's 5G Infrastructure: Mobile to Connected Future, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 251–276. doi: 10.1002/9781394359462.ch13. Fulltekst i vitenarkiv Vis sammendrag

  From 1G to 5G, Ericsson has been at the heart of telecommunications evolution, shaping both the technologies and architectures that define each generation. Particularly, they played a central role in developing and deploying 5G infrastructure. This chapter explores Ericsson's approach to overcoming critical 5G deployment hurdles, and how key challenges were addressed including spectrum use optimization, infrastructure investment strategies, and cybersecurity protocols. The study highlights their innovative solutions such as network slicing technology, energy‐efficient base stations, and Massive MIMO implementations. This uncovers 5G technical challenges and outlines the way Ericsson has strategically positioned itself to face regulatory and infrastructure development difficulties. The chapter also touches on the future vision of the company to 6G technology. Their strategic positioning establishes them as a critical enabler of next‐generation telecommunications infrastructure that could revolutionize sectors ranging from autonomous vehicles to smart city ecosystems.

• Palanichamy, Manikandan (2026). Silicon Carbide Breakthrough: STMicroelectronics’ EV Leadership, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 231–250. doi: 10.1002/9781394359462.ch12. Fulltekst i vitenarkiv Vis sammendrag

  How did an accidental crystal discovery in 1824, became the cornerstone of today's electric vehicle revolution? The answer is silicon carbide (SiC), a remarkable semiconductor that is changing the face of power electronics. This case study unveils STMicroelectronics’ strategy to reach the top of the SiC market between 2019–2024. STMicroelectronics went from being a standard semiconductor company to becoming a vertically integrated powerhouse. Its state‐of‐the‐art facility in Catania, Italy, together with a game‐changing partnership with Tesla, has pushed the boundaries of what's possible in electric vehicle efficiency. The fact that they achieved 99% efficiency in third generation SiC MOSFETs is a significant technical achievement. But most importantly, it speaks not only to the strategic vision but also of how material innovation can speed up the move toward sustainable transportation.

• Palanichamy, Manikandan (2026). AMD−NVIDIA GPU Revolution: Power and Performance Dominance, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 215–230. doi: 10.1002/9781394359462.ch11. Fulltekst i vitenarkiv Vis sammendrag

  This chapter explores the intense competition between AMD and NVIDIA in developing power‐efficient GPUs, tracing the evolution from basic graphics accelerators to sophisticated computing powerhouses. It explores how these titans of the industry have pivoted to satisfy increasing pressure for energy‐efficient solutions in gaming, data centers, AI, and cryptocurrency mining. The study shows the journey of the evolution in GPU technology. It is going from Ivan Sutherland's Sketchpad to programmable architecture today, with high importance put on pivotal inventions and their influence in industry. Second, this chapter examines the strategic battle between these two tech giants, AMD and NVIDIA. It explores the motivations driving each company and the engineering breakthroughs that have shaped what is arguably the most dominant rivalry of the last decade. And further, it recognizes the leading‐edge emerging competitors and GPU centered markets in the worldwide GPU landscape.

• Palanichamy, Manikandan (2026). Intel's 3D Transistor Stacking: Vertical Scaling Breakthrough, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 199–214. doi: 10.1002/9781394359462.ch10. Fulltekst i vitenarkiv Vis sammendrag

  When Moore's Law was running up against its physical and economic boundaries, Intel redefined the roadmap of semiconductor design from 2D scaling to vertical 3D stacking and state‐of‐the‐art transistor architectures. This chapter explores how Intel pioneered FinFET and RibbonFET, two advanced transistor technologies that address critical challenges in power leakage, heat dissipation, and economic viability as semiconductor nodes continue to shrink. Much like the construction of skyscrapers, this innovation represents a shift in architectural design philosophy. The new approach achieves higher transistor density and lower power consumption. These improvements ensure sustained performance scaling for high‐performance computing, artificial intelligence, and data‐centric applications for the future. Viewed through historical glasses, this chapter observes how the industry went from planar MOSFETs to modern tech, a new era milestone in chip design. It provides a comprehensive real‐world example of Intel's innovation journey. The case study covers key milestones, the limitations of traditional scaling throughput, and the industry's broader shift toward 3D chip integration. It is culminating in the next generation of computing with a concluding chapter on how this shift affects the future of computation.

• Palanichamy, Manikandan (2026). Nordic Semiconductor's Ultra‐Low‐Power Revolution: IoT Energy Trilemma, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 183–198. doi: 10.1002/9781394359462.ch09. Fulltekst i vitenarkiv Vis sammendrag

  The explosive growth of the Internet of Things (IoT) has brought with it a persistent challenge: power consumption. In this chapter, we trace the journey of ultra‐low‐power wireless technologies with dynamic power management and FinFET chips design. Using an in‐depth examination of Nordic Semiconductor, reveals how BLE (Bluetooth Low Energy) remade wearables, smart homes, and medical devices. Dynamic voltage scaling, duty cycling and low‐power transistors are described in depth over selected technical strategies using examples from industry. The necessary trade‐offs between power, performance and range in a landscape with an emphasis made from the different ways Nordic have solved this trilemma. This chapter provides a balanced assessment of how power efficiency has emerged as the lifeblood of IoT innovation with the help of a well‐balanced blend of technical‐depth and historical context.

• Palanichamy, Manikandan (2026). TSMC's N3 Breakthrough: Beyond Moore's Law, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 167–182. doi: 10.1002/9781394359462.ch08. Fulltekst i vitenarkiv Vis sammendrag

  As the semiconductor industry edges closer to the physical limits of Moore's Law, Taiwan Semiconductor Manufacturing Company's (TSMC's) 3 nm (N3) process technology sets a new benchmark with approximately 300 million transistors per square millimeter, delivering up to 15% performance gains and 30% power‐efficiency improvements over its 5 nm predecessor. This chapter explores the technological marvel and engineering challenges behind the development of TSMC's N3 process technology. Key innovations such as FinFlex™, Gate‐All‐Around (GAA) transistor architectures, and extreme ultraviolet (EUV) lithography have enabled these advancements despite daunting manufacturing complexities, power leakage, and quantum effects. Using practical analogies and real‐world examples like the A17 Pro chip in the iPhone 15 Pro, the chapter illustrates how TSMC continues to lead in nanoscale semiconductor innovation while navigating the complexities of defect control, yield optimization, and cost‐performance trade‐offs in the post‐Moore era.

• Palanichamy, Manikandan (2026). Semiconductors, Circuits, and Connectivity, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 165–166. doi: 10.1002/9781394359462.part02. Fulltekst i vitenarkiv Vis sammendrag

  This section examines transformative innovations reshaping semiconductor technology and global connectivity infrastructure. It explores TSMC's N3 process node advancement, Nordic Semiconductor's ultra-low-power IoT solutions, and Intel's pioneering 3D transistor stacking techniques. The chapters analyze AMD and NVIDIA's GPU architectures driving computational performance, STMicroelectronics' silicon carbide breakthroughs enabling efficient electric vehicle power systems, and Ericsson's 5G infrastructure deployment strategies. Additionally, it investigates Huawei's AI-driven autonomous network technologies developed amid geopolitical challenges and SpaceX Starlink's low Earth orbit satellite constellation revolutionizing global internet accessibility. These case studies demonstrate how industry leaders navigate technical constraints, market demands, and regulatory landscapes while pushing the boundaries of Moore's Law, energy efficiency, and wireless communication capabilities in an increasingly interconnected world.

• Palanichamy, Manikandan (2026). Tesla's Approach to Grid‐Scale Battery Storage, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 139–164. doi: 10.1002/9781394359462.ch07. Fulltekst i vitenarkiv Vis sammendrag

  This chapter explores the cutting edge of grid‐scale battery energy storage systems (BESS) as laid out by Tesla. It is addressing the essential problem of how to integrate intermittent renewable energy sources into existing power grids. Using the Hornsdale Power Reserve project in South Australia as an example, this discussion highlights how state‐of‐the‐art BESS can solve grid stability challenges and unlock the potential for greater renewable energy adoption. In this study, the impact of Tesla's battery systems is analyzed in detail, along with measures used worldwide to evaluate battery performance. Challenges such as heat dissipation, mechanical fatigue, and grid interconnection complexity are examined. The chapter also explores advancements and implications for scaling BESS solutions, emphasizing the critical role of energy storage in supporting the global transition to renewable energy.

• Palanichamy, Manikandan (2026). London Array: NEXANS’ Subsea Cable Breakthrough, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 117–138. doi: 10.1002/9781394359462.ch06. Fulltekst i vitenarkiv Vis sammendrag

  This chapter examines the London Array Offshore Wind Farm project as a landmark achievement in renewable energy engineering. The study focuses on the technical challenges of harnessing offshore wind power, and how NEXANS overcame the big issues of design and installation of subsea cables. It furthermore elaborates the important technical field of power systems, cable optimization, and electromagnetic interference control. This study also highlights how this project contributes to further sustainability in the future energy infrastructure development and supporting the European climate goals.

• Palanichamy, Manikandan (2026). Blacksmith to Smart Wind: Vestas’ AI Transformation, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 93–116. doi: 10.1002/9781394359462.ch05. Fulltekst i vitenarkiv Vis sammendrag

  The remarkable transformation of Vestas as a global wind energy leader perfectly illustrates how traditional industries can evolve through innovation and smart technologies to become pioneers in renewable energy solutions. In this chapter we investigate the combination of artificial intelligence, IoT sensors, and big data analytics in Vestas turbine systems that illustrate the intra‐operability pioneer projects within the company. It features predictive maintenance, aerodynamic research, and energy hybrids irrelevant to financial meltdowns, infrastructure struggles, and cybersecurity attacks. Vestas continue to develop even further turbine efficiencies with hybridization. A step in the evolution of Vestas, the wind turbine is very similar to the evolution of wind energy from primitive early designs to modular AI controlled giants. In smart wind turbines and smart grid battle, Vestas show their strength as an energy supplier through agile technologies.

• Palanichamy, Manikandan (2026). First Solar's Cadmium Telluride Revolution: Disrupting Silicon Dominance, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 75–92. doi: 10.1002/9781394359462.ch04. Fulltekst i vitenarkiv Vis sammendrag

  In this chapter we trace the radical emergence of cadmium telluride (CdTe) solar panel technology through the First Solar company. It begins by examining the limitations of traditional silicon‐based photovoltaic cells and the increasing global demand to develop scalable, efficient, and affordable solar solutions. CdTe panels became a disruptive alternative with essentially no performance trade‐off across climates and shorter energy payback time draw on this historical survey of the development of solar power from limited discoveries, and the photovoltaic effect‐up until our current paradigm primarily dominated by silicon‐based photovoltaics and finally entering in the thin‐film technology era. Centrality is attributed to First Solar's innovations, strategy and relentless research and development on CdTe. It brought technology from lab proof of concept into a commercial product. It explores an example of new technologies potentially shaking the foundations of renewable energy and disrupting the deeply rooted practices within this sector.

• Palanichamy, Manikandan (2026). DONG to Ørsted: Viking Longship Green Hydrogen Strategy, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 53–74. doi: 10.1002/9781394359462.ch03. Fulltekst i vitenarkiv Vis sammendrag

  Ørsted, formerly Denmark's dominant oil and gas company, DONG Energy, exemplifies one of the most profound corporate energy transitions in recent history. Having divested its fossil fuel assets and flipped gears to offshore wind and green hydrogen it will now be seen as a global sustainable energy leader. This chapter discusses Ørsted's strategy shift to lead in the field of large‐scale green hydrogen through land‐based projects in Ørsted Land, SeaH2Land, Gigastack and Green Fuels for Denmark. This chapter focuses on the innovative “Viking Longship Strategy” of Viking Longship ‒ an all‐in‐one coordination of technology developers, blue dynamos and industrial clusters for efficient scale‐up of hydrogen production. It emphasized synergy policy navigator and infrastructure integration.

• Palanichamy, Manikandan (2026). GreenHyScale Denmark: Large‐Scale Green Hydrogen Engineering, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 31–52. doi: 10.1002/9781394359462.ch02. Fulltekst i vitenarkiv Vis sammendrag

  Transitioning from fossil‐based hydrogen to green hydrogen is like converting a diesel ship to an electric mid‐voyage ‒ complex, but essential. That journey is the subject of this chapter, GreenHyScale project in Denmark ‒ a green belt hydrogen production from renewable electricity on a larger scale. This covers the progression through green hydrogen systems and renewable energy, predominantly focusing on large‐scale electrolysis technologies (alkaline, PEM and SOEC) development. This chapter lays out the importance of green hydrogen for decarbonization and describes key technical, economic, and policy hurdles that will determine its path forward. Through extensive use of engineering solutions, collaborative activities and strategic frameworks the book demonstrates how green hydrogen can turn energy storage, industrial processes, and power systems upside down. At the end of the chapter, it stresses the disruptive ability of engineering to deliver a long‐term European hydrogen economy across Europe and worldwide.

• Palanichamy, Manikandan (2026). BorWin2: HVDC Breakthrough for Offshore Wind, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 15–30. doi: 10.1002/9781394359462.ch01. Fulltekst i vitenarkiv Vis sammendrag

  Designed to test the harshest marine conditions, BorWin2 presented an upgrade from a high‐voltage direct current transmission, HVDC Light® to HVDC Plus technology allowing for the engineering of long‐distance power transfer with minimal losses. Examining Siemens challenges and solutions provides valuable insights within Germany's energy transformation context. Siemens faced technology, environmental, and strategic challenges. Their experience reflects Europe's broader green energy shift. It demonstrates effective methods for connecting massive offshore renewable projects to national power grids. This chapter explores the groundbreaking BorWin2 project, designed by Siemens to transport offshore wind power from the North Sea to the German mainland.

• Palanichamy, Manikandan (2026). Renewable Energies and Power Infrastructure introduction, Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISSN 9781394359431. s. 1–14. doi: 10.1002/9781394359462.cintro. Fulltekst i vitenarkiv Vis sammendrag

  This provides comprehensive explorations of fundamental engineering principles, theories, and methodologies essential for understanding renewable energies and power infrastructure. Through carefully selected case studies, it offers practical discussions on the limitations and boundaries of each application, clearly identifying contexts where specific approaches may not be suitable. By bridging the gap between theoretical frameworks and real-world implementation, this resource serves as an invaluable tool for engineers, researchers, and practitioners working in the renewable energy sector. The book emphasizes critical thinking by not only presenting successful solutions but also analyzing constraints, challenges, and lessons learned, enabling readers to make informed decisions when designing and deploying sustainable power systems in diverse operational environments.

• Kumar, Vivek; Mohanty, Soumya R.; Kishor, Nand & Prasad, Udit (2026). Robust Global Adaptive Asymptotic Performance in Power Systems with Indefinite Control Parameters. Journal of the Institution of Electronics and Telecommunication Engineers. ISSN 0377-2063. doi: 10.1080/03772063.2026.2683561. Fulltekst i vitenarkiv
• Nizamani, Abdul Ghafoor; Vu, Dung Thi; Nizamani, Rafeeque Ahmed & Li, Cunrong (2026). Physics-informed neural network calibration and application for urban cooling. Sustainable Cities and Society (SCS). ISSN 2210-6707. 147, s. 107525–107525. doi: 10.1016/j.scs.2026.107525. Fulltekst i vitenarkiv
• Herrebrøden, Marte Karoline; Marentakis, Georgios; Dechsling, Anders; Esposito, Gianluca; Dimitriou, Dagmara & Nordahl-Hansen, Anders (2026). Advantages and disadvantages of noise-cancelling headphones for autistic students in educational settings. Research in Developmental Disabilities. ISSN 0891-4222. 173. doi: 10.1016/j.ridd.2026.105301. Fulltekst i vitenarkiv Vis sammendrag

  This perspective article explores the various advantages and disadvantages of using noise-cancelling headphones in educational settings for autistic students with auditory sensory differences. Although noise-cancelling headphones can have several advantages for autistic students in educational settings, it is important that use of such assistive technological tools should not be implemented without a plan and tailoring at an individual level.

• Jamal, Abdullah; Kishor, Nand & Mutale, Anna (2026). Disaggregation of Space-Heating load in buildings from hourly Smart-Meter data using a hybrid ICA–PLR approach. Energy and Buildings. ISSN 0378-7788. 364. doi: 10.1016/j.enbuild.2026.117618. Fulltekst i vitenarkiv Vis sammendrag

  Space-heating dominates electricity consumption in Norwegian public buildings, yet hourly space-heating load (SHL) is rarely sub-metered, limiting monitoring and demand response (DR). The objective of this study is to develop and evaluate a physics-guided, scalable method to disaggregate hourly SHL from whole-building hourly electric load using only weather and calendar information, without SHL sub-metering. Specifically, this study evaluates the performance of a hybrid ICA-PLR estimator for hourly SHL disaggregation against PLR, RFR, and FastICA benchmarks. This study presents a hybrid disaggregation framework that couples Independent Component Analysis (ICA) with a heating-degree-hour piecewise linear regression (HDH-PLR) under a temperature- and schedule-aware cold-hour mask. ICA extracts a heating-like latent component from cold-hour residuals across multiple buildings; HDH-PLR re-anchors scale and slope to a physics-consistent temperature signature; and mask gating with lag, seasonal, and weekend factors suppresses warm-weather false activation. The proposed framework requires only hourly smart-meter, weather, and calendar data. The method is applied to a portfolio of 22 public buildings (schools, kindergartens, offices, and institutions). All quantitative validation is performed on a sub-metered validation subset of 5 buildings using hourly SHL as ground truth. Across the validation subset, results indicate that hourly accuracy reaches R2 ≈ 0.91 with nMAE ≈ 13 % to 14 %, while estimated hourly SHL remains near zero during summer periods. At the building level in the validation subset, the method tracks morning recovery, night setbacks, and cold-spell amplitudes with fewer mild-weather false positives than the benchmark techniques. The approach yields interpretable parameters (balance temperature, lag, seasonal/weekend factors) that support weather-normalized monitoring and flexibility operations in public building portfolios. This work demonstrates that physics-guided disaggregation can extend operational visibility of public-building heating systems using only utility-grade data and support weather-normalized monitoring and flexibility planning across public-building portfolios.

• Dey, Rakesh; Palaiahnakote, Shivakumar; Bhattacharya, Saumik; Pal, Umapada & Chanda, Sukalpa (2026). SFTT: A Spatial-Frequency-Temporal-Based End-to-End Transformer for Heart Rate Estimation. IEEE Transactions on Emerging Topics in Computational Intelligence (TETCI). 10(1), s. 454–467. doi: 10.1109/TETCI.2025.3582841. Fulltekst i vitenarkiv
• Fatima, Maham; Khan, Shanza Rauf; Ali, Sarmad; Jamil, Saba; Munawar, Khurram Shahzad & Aslam, Rimsha [Vis alle 8 forfattere av denne artikkelen] (2026). Synergistic Ag-Mn<inf>3</inf>O<inf>4</inf>/SA/CNBs Nanocatalyst: a solution for organic pollutants reduction in wastewater. Journal of chemical technology and biotechnology (1986). ISSN 0268-2575. 101(4), s. 740–748. doi: 10.1002/jctb.70138. Fulltekst i vitenarkiv
• Prøitz, Lin (2026). Affect in the classroom: Self-reflections towards professional growth. I Muthanna, Abdulghani; Tandberg, Håkon Naasen & Khine, Myint Swe (Red.), Pedagogical Competence Development in Higher Education Volume 2: Navigating Reflective Practices and Transformative Pedagogies. Routledge. ISSN 9781041153979. doi: 10.4324/9781003679233-5. Fulltekst i vitenarkiv Vis sammendrag

  This chapter reflects on two teaching episodes, separated by 15 years, from courses on Diversity and Intercultural Communication. The first vignette recounts an early-career classroom collapse, explored through three phases of self-reflexive processing: initial defensiveness, followed by deeper emotional inquiry using feminist theories on affect and perspectives from psychoanalytic theory and education. The second case, shaped by years of pedagogical training, presents a cooperative learning session integrating theory and student-centred activities. The two cases illustrate how teaching, as a dynamic practice, can never be planned in detail as it involves moments of unique and unforeseen contact between bodies, emotions and individuals who are moved and engaged by each other in various and sometimes unexpected ways.

• Yadav, S.; Kishor, Nand; Purwar, S. & Rajkumar, V. S. (2026). State Partition-Particle Filter Detection for Cyber-Physical Attacks. IEEE Transactions on Information Forensics and Security. ISSN 1556-6013. doi: 10.1109/TIFS.2026.3671018. Fulltekst i vitenarkiv
• Ajeeb, Rayan; Madi, Mervat; Aaker, Olav; Ntantis, Efstratios L.; Srayeddin, Rayyan & Ghader, Hilal Bou (2026). Computational investigation of amorphous core materials in electromagnetic induction systems using power transformers. Proceedings of the Institution of mechanical engineers. Part C, journal of mechanical engineering science. ISSN 0954-4062. doi: 10.1177/09544062261423243. Fulltekst i vitenarkiv
• Deepak, M.; Chokkalingam, Bharatiraja; Verma, Rajesh; Sathiyasekar, K. & Mihet, Nicolae Lucian (2026). Sensorless Model Predictive Direct Torque Regulation with Modified Switching Table for Double‑Pole Salient Machines to Improve the Performance. IEEE Access. doi: 10.1109/ACCESS.2026.3675882. Fulltekst i vitenarkiv
• Stigberg, Henrik & Stigberg, Susanne Koch (2026). Unpacking Teachers’ Transformative Agency Following Digital Fabrication Workshops on Making Mathematical Manipulatives. Digital Experiences in Mathematics Education. ISSN 2199-3246. 12, s. 393–415. doi: 10.1007/s40751-026-00194-3. Fulltekst i vitenarkiv Vis sammendrag

  Digital fabrication (DF) presents innovative opportunities for reimagining the use of manipulatives in mathematics education. Manipulatives are fundamental for helping students grasp abstract mathematical concepts. However, utilizing DF to make mathematical manipulatives requires more than just access to technology; it demands teachers’ transformative agency. This article presents the experiences of seven teachers who participated in a series of eight full-day workshops over a span of one year, focusing on making mathematical manipulatives using DF. Through interviews, the teachers offered insights into their transformative agency, detailing their sense of purpose, competency, autonomy, and reflexivity. Our results highlight the complexities, opportunities, and challenges encountered by these teachers as they navigated the integration of DF into their teaching practice. Our study led to three key implications: Utilizing DF enables teachers to make manipulatives that are pedagogically relevant; Making manipulatives with DF supports changes in teaching practices, showcasing teachers’ transformative agency; but also, a critical need for supportive infrastructure to facilitate teachers becoming makers in education. The article underscores the potential of utilizing DF to transform mathematics education while emphasizing the importance of providing sustained support to realize its potential fully.

• Sadagopan, A.; Udhayakumar, K.; Verma, Rajesh; Chokkalingam, Bharatiraja & Mihet, Nicolae Lucian (2026). Predictive Control Techniques for Switched Reluctance Motor: An Overview. IEEE Access. Fulltekst i vitenarkiv Vis sammendrag

  The Switched Reluctance Motor (SRM) is a developing electric drive innovation that outperforms the Induction Motor (IM) and the Permanent Magnet Synchronous Motor (PMSM) in numerous industrial uses, including Electric Vehicles, HVAC Systems, and applications that demand high speeds and reliable fault tolerance. Nonetheless, it has many drawbacks, such as Torque Ripple, Acoustic Noise, and Non-Linear Torque Characteristics, which may restrict its use in certain industrial settings. This article gives an overview of Predictive Control (PC) techniques for controlling current-torque and Torque-Speed characteristics, thereby eliminating main issues such as torque ripple and current ripple, and compensating for Nonlinear Torque characteristics. The paper analyzes Model-based Predictive Control (MPC) techniques and Model-free Predictive Control techniques from a modeling perspective, including switching vector optimization methods, cost function choices, and control efficiency. Finally, the existing difficulties and future advancements in implementing MPC schemes in SRM are discussed.

• Svellingen, Werner; Torgersen, Geir; Bruland, Oddbjørn & Muthanna, Tone Merete (2026). Scalable pluvial flood risk assessment: A data-driven framework integrating machine learning (ML) and discrete global grid systems (DGGS H3). International Journal of Disaster Risk Reduction. 137, s. 106091–106091. doi: 10.1016/j.ijdrr.2026.106091. Fulltekst i vitenarkiv
• Mainsah, Henry; Prøitz, Lin & Busso, Lilliana Andrea Del (2026). Young people, social media, and critical health media literacies. Qualitative Health Communication (QHC). 5(1), s. 97–111. doi: https:/tidsskrift.dk/qhc/article/view/158616. Fulltekst i vitenarkiv Vis sammendrag

  Background: Social media and other digital platforms are becoming significant sources of health, fitness, and nutrition information for young people. This underlines the need to examine the potential consequences of such information practices for young people’s health knowledge, their relationship to their bodies, and their wellbeing. Aim: Using a sociocultural approach to literacy, this article examines how young people understand, experience, and engage with health information mediated through digital technologies. Methods: Five activity-based focus group interviews were organized with 14–17-year-olds as part of a pilot project informing a broader study about health information and critical literacies. Results: Content about fitness, nutrition, and body image constituted a dominant part of the health-related information the participants encountered on social media. The participants were aware of algorithmic processes that produced health information and affirmed their ability to evaluate the quality of content on social media. They were critical of but accepted being the object of surveillance and data capture. Participants were critical of health narratives on social media by highlighting the unhealthy practices of influencers and questioning perfect body ideals. However, they still aspired to these “perfect” masculine and feminine body ideals. Discussion and conclusion: The study highlights the affective, situated, and relational nature of young people’s health literacy practices. This reflects the tensions, contradictions and ambivalences that are an integral part of the wider youth digital culture.

• Heiney, Kristine; Józsa, Mónika; Rule, Michael E.; Sprekeler, Henning; Nichele, Stefano & O’Leary, Timothy (2026). Information theoretic measures of neural and behavioural coupling predict representational drift. PLoS Computational Biology. ISSN 1553-734X. 22(2), s. 1013130–1013130. doi: 10.1371/journal.pcbi.1013130. Fulltekst i vitenarkiv Vis sammendrag

  In many parts of the brain, population tuning to stimuli and behaviour gradually changes over the course of days to weeks in a phenomenon known as representational drift. The tuning stability of individual cells varies over the population, and it remains unclear what drives this heterogeneity. We investigate how a neuron’s tuning stability relates to its shared variability with other neurons in the population using two published datasets from posterior parietal cortex and visual cortex. We quantified the contribution of pairwise interactions to behaviour or stimulus encoding by partial information decomposition, which breaks down the mutual information between the pairwise neural activity and the external variable into components uniquely provided by each neuron and by their interactions. Information shared by the two neurons is termed ‘redundant’, and information requiring knowledge of the state of both neurons is termed ‘synergistic’. We found that a neuron’s tuning stability is positively correlated with the strength of its average pairwise redundancy with the population. We hypothesize that subpopulations of neurons show greater stability because they are tuned to salient features common across multiple tasks. Regardless of the mechanistic implications of our work, the stability–redundancy relationship may support improved longitudinal neural decoding in technology that has to track population dynamics over time, such as brain–machine interfaces.

• Nizamani, Abdul Ghafoor; Vu, Dung Thi; Nizamani, Rafeeque Ahmed & Torgersen, Geir (2026). Urban resilience through nature-based solutions in an era of environmental extremes: A global systematic analysis. Sustainable Cities and Society (SCS). ISSN 2210-6707. 139. doi: 10.1016/j.scs.2026.107192. Fulltekst i vitenarkiv
• Oruma, Samson Ogheneovo; Colomo-Palacios, Ricardo & Gkioulos, Vasileios (2026). From Framework to Reliable Practice: End-User Perspectives on Social Robots in Public Spaces. Systems. 14(2). doi: https:/www.mdpi.com/2079-8954/14/2/137. Fulltekst i vitenarkiv Vis sammendrag

  As social robots increasingly enter public environments, their acceptance depends not only on technical robustness but also on ethical integrity, accessibility, transparency, and consistent system behaviour across diverse users. This paper reports an in situ pilot deployment of an ARI social robot functioning as a university receptionist, designed and implemented in alignment with the SecuRoPS framework for secure, ethical, and reliable social robot deployment. Thirty-five students and staff interacted with the robot in a real public setting and provided structured feedback on safety, privacy, usability, accessibility, ethical transparency, and perceived reliability. The results indicate strong user confidence in physical safety, data protection, and regulatory compliance while revealing persistent challenges related to accessibility and interaction dynamics. These findings show that reliability in public-facing robotic systems extends beyond fault-free operation to include equitable and consistent user experience across contexts. Beyond reporting empirical outcomes, the study contributes in three key ways. First, it demonstrates a reproducible method for operationalising lifecycle governance frameworks in real-world deployments. Second, it provides new empirical insights into how trust, accessibility, and transparency are experienced by end users in public spaces. Third, it delivers a publicly available, open-source GitHubrepository containing reusable templates for ARI robot applications developed using the PAL Robotics ARI SDK (v23.12), lowering technical entry barriers and supporting reproducibility. By integrating empirical evaluation with practical system artefacts, this work advances research on reliable intelligent environments and provides actionable guidance for the responsible deployment of social robots in public spaces.

• Manimaran, S; Chokkalingam, Bharatiraja; Verma, Rajesh; Datta, Shubhabrata & Mihet-Popa, Lucian (2026). Design and Development of Multi-Utility Electric Vehicle Powertrain System. IEEE Access. 14, s. 28701–28716. doi: 10.1109/access.2026.3657296. Fulltekst i vitenarkiv
• Andersson, Gunnar; Hauge, Bjørn Gitle; Johansen, Frode Ramstad; Leister, Eivind Andre; Valter, Per & Lynch, Matthew [Vis alle 7 forfattere av denne artikkelen] (2026). Fostering Resilience and Innovation in Island Communities – Identifying Systemic Challenges and Opportunities. THE JOURNAL OF DEVELOPMENT STUDIES. ISSN 2720-8672. 6(1), s. 44–56. doi: 10.63410/jds2025/04. Fulltekst i vitenarkiv Vis sammendrag

  This paper explores how stakeholders in island communities perceive and address systemic challenges to resilience and innovation in the context of the green transition and broader societal change. Drawing on insights from the Interreg FREIIA project – a multi-country, participatory initiative across six European islands—this study employs a design thinking framework and grounded theory methodology to explore how local stakeholders perceive and respond to development challenges, highlighting the voices and lived experiences of island residents and aiming to understand the realities on the ground and the multidimensional conditions shaping resilience and innovation in their communities. The findings highlight recurring development challenges, including youth outmigration, housing shortages, seasonal economic dependency, and limited institutional capacity. The paper argues that sustainable development in island contexts requires participatory governance, inclusive innovation ecosystems, and place-based strategies that reflect local realities. By situating island development within broader debates on peripheralization, productive capabilities, and participatory transformation, the study contributes to the literature on systemic development in island communities.

• Jana, Teremranova,; Jamal, Abdullah; Lubov, Petrichenko,; Marina, Arias,; Fernando, Borges, & Dragos, Cirstea, Stefan [Vis alle 10 forfattere av denne artikkelen] (2026). Electricity customer satisfaction with renewable energy support schemes: a case study for Latvia, Romania, Spain, Portugal, Norway. Frontiers in Energy Research. 13. doi: 10.3389/fenrg.2025.1736680. Fulltekst i vitenarkiv Vis sammendrag

  The success of the energy transition increasingly depends on engaging household electricity customers as active customers. The paper investigates how active ...

• Sánchez-Gordón, Mary; Colomo-Palacios, Ricardo; Gordon, Alex Sanchez & Hubin, Aliaksandr (2026). The role of voice and silence behaviors in software development: a structural equation modeling analysis. Information and Software Technology. ISSN 0950-5849. 190. doi: 10.1016/j.infsof.2025.107972. Fulltekst i vitenarkiv Vis sammendrag

  Context Most software companies strive to have high-performing teams and mitigate withdrawal behaviors like being present but unproductive. In this context, psychological safety and developers’ perceived impact are suggested as potential drivers of voice and silence behaviors. However, understanding these social aspects of software development entails the incorporation of social science theories. Objective This study aims to empirically explore the relationships among software professionals’ perceived impact, psychological safety, voice and silence behaviors, burnout particularly withdrawal and performance using a theoretical model. Method A survey questionnaire was conducted, resulting in 158 valid responses from software development teams. Then, we analyzed the responses using structural equation modeling (SEM) and a novel semi-confirmatory SEM. All variables were measured using pre-validated instruments. Results The findings supported the theoretical model, showing that psychological safety was more related to silence than voice, whereas perceived impact showed a stronger relationship with voice than with silence. Furthermore, silence contributed to higher burnout, whereas voice alleviated it. In contrast, silence showed a negative association with performance, whereas voice was positively associated with it. Conclusions This study examines the direct effects of perceived impact and psychological safety on voice and silence behaviors. Additionally, it examines the direct effects of these behaviors on burnout and task performance among software professionals. Consequently, the findings offer both theoretical and practical insights.

• Aliyari, Mostafa; Safhi, Amine el Mahdi; Pilehvar, Shima; Nehdi, Moncef L. & Kioumarsi, Mahdi (2026). High-fidelity data-driven multi-objective design of class F fly ash–geopolymer concrete. Powder Technology. ISSN 0032-5910. 469(2). doi: 10.1016/j.powtec.2025.121900. Fulltekst i vitenarkiv Vis sammendrag

  Class F fly-ash geopolymers (FA-GPCs) can slash the CO₂ burden of structural concrete, yet their mix design still relies on costly trial-and-error. The first geometry-normalized database of FA-only geopolymer concretes was assembled—799 mixes from 67 publications, every reported strength converted to the 150 Ø 300 mm cylinder equivalent—and interrogate it with a three-scenario, data-availability workflow. Eighteen machine-, deep- and hybrid-learning models are benchmarked; a Voting ensemble predicts 7–28 days compressive strength with an R2 of 0.962, while an LSTM + XGBoost hybrid forecasts mix-specific CO₂ footprints at an R2 of 0.996. SHAP analysis reveals that the Si/Al ratio and initial curing temperature dominate early-age strength, whereas total Na₂SiO₃/NaOH and Na₂O/binder ratios control embodied carbon, underscoring an inherent strength–footprint trade-off. These surrogate models feed a multi-objective optimizer (NSGA-II, confirmed superior to NSGA-III and MOEA/D) that explores the 1st–99th-percentile bounds of 11 mix variables. The resulting Pareto fronts deliver eco-efficient recipes reaching 65 MPa at ≤30 kg CO₂/m3 (at 90-d curing age) and extend to 87 MPa under relaxed carbon targets. Compared with state-of-the-art optimization that rely on mixed precursors or un-normalized data, the proposed pipeline widens the design space, lifts strength by 10–20 MPa and trims CO₂ by up to 70 %. The framework— in this order: data curation, geometry correction, scenario-specific ML, evolutionary optimization—cuts trial batching efforts and offers practitioners ready-to-deploy, class F FA-GPC mixtures while providing a transparent template for other alkali-activated binders which was confirmed by an experimental validation.

• Fedushko, Solomia; Mykhalchyshyn, Natalia; Trevoho, Olena; Oksana, Hoshovska & Roddvik, Irina Nikolskaja (2026). Comparative Governance: Power, Success, And Leadership In Public Administration. Social Sciences & Humanities Open. Social Sciences & Humanities Open (SSHO). Fulltekst i vitenarkiv
• Naeem, Aiman; Ali, Sarmed; Jamil, Saba; Bibi, Shamsa; Noreen, Saima & Latif, Muhammad Jamshed [Vis alle 9 forfattere av denne artikkelen] (2025). Ionotropic gelation synthesis of Cu⁰-Fe⁰mono- and bi-metallic SA/g-CNBs for the reduction of single and binary dye solutions: A comparative study. Materials Today Communications. 49. doi: 10.1016/j.mtcomm.2025.113827. Fulltekst i vitenarkiv
• Islam, Fayzul; Ahmed, Ibrahim Abbas Elamen & Mihet, Nicolae Lucian (2025). Demand-Side Management in Residential Buildings: Load Shifting and Cost Savings via a Cloud-Based IoT-BEMS. I 2025, IECON (Red.), IECON 2025 – 51st Annual Conference of the IEEE Industrial Electronics Society. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331596811. s. 1–7. doi: 10.1109/IECON58223.2025.11221160. Fulltekst i vitenarkiv
• Palanichamy, Manikandan (2025). Artificial Intelligence Applications in Green Hydrogen Production Technologies. IEEE East-West Design & Test Symposium (EWDTS). ISSN 2472-761X. doi: 10.1109/ewdts67441.2025.11303674. Fulltekst i vitenarkiv Vis sammendrag

  This paper explores how artificial intelligence, and computational methods like ML can transform green hydrogen while overviewing green hydrogen production technologies. AI has already revolutionized manufacturing industries by turning complex data into useful insights. Now, it offers similar potential for hydrogen technology. This work shows various AI techniques used in green hydrogen production, including optimization strategies, predictive maintenance systems, and intelligent control systems. The review highlights AI-driven digital twins and hybrid architectures in electrolysis systems as particularly promising developments. Machine learning applications demonstrate significant improvements in production efficiency and cost reduction. This study emphasizes key challenges including data quality, standardization, and regulatory frameworks, while identifying future research directions in this fast-evolving field for cost-effective, and sustainable hydrogen production.

• Pontes-Filho, Sidney; Nichele, Stefano & Lepperød, Mikkel Elle (2025). Reservoir Computing with Evolved Critical Neural Cellular Automata. I Witkowski, Olaf; Adams, Alyssa M.; Sinapayen, Lana; Baltieri, Manuel & Khosravy, Mahdi (Red.), ALIFE 2025: Ciphers of Life: Proceedings of the Artificial Life Conference 2025. MIT Press. doi: 10.1162/isal.a.832. Fulltekst i vitenarkiv Vis sammendrag

  Criticality is a behavioral state in dynamical systems that is known to present the highest computation capabilities, i.e., information transmission, storage, and modification. Therefore, such systems are ideal candidates as a substrate for reservoir computing, a subfield in artificial intelligence. Our choice of a substrate is a cellular automaton (CA) governed by an artificial neural network, also known as neural cellular automaton (NCA). We apply evolution strategy to optimize the NCA to achieve criticality, demonstrated by power law distributions in structures called avalanches. With an evolved critical NCA, the substrate is tested for reservoir computing. Our evaluation of the substrate is performed with two benchmarks, 5-bit memory task and image classification of handwritten digits. The result of the 5-bit memory task achieved a perfect score and the system managed to remember all 5 bits. The result for the image classification task matched and sometimes surpassed the performance of the best elementary CA for this task. Moreover, the proposed critical NCA may operate as a self-organized critical system, due to its robustness to extreme initial conditions.

• Mieruch, Sebastian; Kreps, Gastón; Chouai, Mohamed; Reimers, Felix Simon; Vredenborg, Myriel & Rabe, Benjamin [Vis alle 8 forfattere av denne artikkelen] (2025). SalaciaML-2-Arctic — a deep learning quality control algorithm for Arctic Ocean temperature and salinity data. Frontiers in Marine Science. 12. doi: 10.3389/fmars.2025.1661208. Fulltekst i vitenarkiv
• Leseanu, Anda; Hallac, Ibrahim Riza; Ogul, Burcin Buket & Ogul, Hasan (2025). Revisiting Automatic Essay Assessment: A Relative Approach. Lecture Notes in Computer Science (LNCS). ISSN 0302-9743. s. 41–52. doi: 10.1007/978-3-032-08462-0_4. Fulltekst i vitenarkiv Vis sammendrag

  Automatic essay assessment has remained a challenging problem in natural language processing (NLP) for over fifty years. Traditionally framed as a regression task that predicts examiner-assigned scores, this work instead adopts a relative assessment perspective. The task is redefined as pairwise ranking, where essays are compared using standard assessment criteria, but their exact grades are assumed to be unknown and ultimately unimportant. Three core contributions are presented. First, an architecture is proposed for pairwise essay ranking using labeled quality criteria, with two fusion strategies explored. Second, an experimental setup for relative ranking is redefined based on a benchmark dataset originally intended for absolute scoring. Third, empirical comparisons are reported across architectural components and design choices. Results indicate that data-level fusion using LLaMA-based embeddings and a gradient-boosted classifier achieves the highest accuracy. A model-level fusion approach with a Siamese neural network and probabilistic loss also yields competitive performance, highlighting its potential for further research. This relative assessment framework offers practical value for student progress tracking and norm-referenced grading in educational contexts.

• Pal, Kim Kristoffer; Eriksen, Aleksander Vanberg & Dinh, Thi Thuy Nga (2025). XGBoost Feature Selection for Multi-Class and Binary Classification on UNSW-NB15 Dataset. I IEEE (Red.), 2025 IEEE International Conference on Consumer Electronics (ICCE). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331521165. doi: 10.1109/icce63647.2025.10930023. Fulltekst i vitenarkiv Vis sammendrag

  Network intrusion detection systems (NIDS) are a critical part of ensuring the integrity and reliability of a network, which is essential in secure communication. To achieve this, an efficient NIDS with as low complexity as possible is preferred. Therefore, this paper presents a feature selection (FS) method that utilizes extreme gradient boosting (XGBoost) to train binary and multi-class machine learning (ML) algorithms on the 19 most important features in the UNSW-NB15 dataset. The FS method is evaluated on seven ML algorithms namely Artificial Neural Network (ANN), Logistic Regression (LR), k-Nearest-Neighbors (kNN), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and XGBoost. Results from both the multi-class and binary classifications showed that RF and XGBoost had the highest accuracy and Fl-score at (77.14%, 78%), (87.445%, 87%), (78.156%, 79%) and (87.527%, 87%), respectively. Compared to existing works, most ML algorithms demonstrate better or similar performance on the UNSW-NB15 dataset. In addition, the proposed FS method significantly reduces the running time for XGBoost and RF by approximately 80% and 48%, respectively in both binary and multi-class classifications.

• Lampropoulos, Georgios; Mukta, Boishakhi Ghosh & Anastasiadis, Theofylaktos (2025). Gamification in Learning Management Systems: A Systematic Literature Review. Information. 16(12). doi: 10.3390/info16121094. Fulltekst i vitenarkiv
• Stokseth, Øyvind; Borgebund, Harald & Sætra, Henrik Skaug (2025). For us, by us? I Zaidan, Esmat; Ibrahim, Imad Antoine & Azar, Elie (Red.), Smart Cities to Smart Societies: Moving Beyond Technology. Routledge. ISSN 9781003439325. s. 11–28. doi: 10.4324/9781003439325-3. Fulltekst i vitenarkiv Vis sammendrag

  Smart city initiatives promise to revolutionize urban living, offering technological solutions to improve everything from transportation to public safety. Yet, the success of these initiatives depends not just on the technology but also on the preferences, attitudes, and competencies of the people who use it. In this chapter, we take a closer look at the role of citizens, who are often seen only as end users but are, in fact, pivotal players in shaping their cities. We explore various interpretations of what it means to be a ‘smart’ city and map out the different stakeholders involved, placing a special focus on citizens. Throughout the chapter, we question and discuss how citizens can and should be involved in the planning, development, and governance of smart cities. Our goal is to explain how cities can better involve and engage their residents, not just as passive recipients of smart technologies but as active contributors to the smart cities being developed. In closing, we outline practical ways to enhance citizen involvement, ensuring that smart cities fulfill their promise to improve life for all.

• Biedacha, Maciej; Tellnes, Lars G. F. & Kjøniksen, Anna-Lena (2025). Mapping Recent Developments in Bio-Based Insulation Materials: A review of trends and environmental impacts. IOP Conference Series: Earth and Environmental Science (EES). ISSN 1755-1307. 1568(1). doi: 10.1088/1755-1315/1568/1/012020. Fulltekst i vitenarkiv Vis sammendrag

  Abstract European legislation continuously increases requirements for carbon footprint accounting for buildings. The energy efficiency and embodied carbon of buildings are regulated through the energy efficiency directive with targets for 2030. While construction production regulation makes it mandatory to disclose carbon footprints of key materials such as insulation by 2027. Improving building products to lower their embodied CO2 emissions and to include additional carbon storage capabilities is seen as one of the most promising ways to mitigate the climate impact of the construction sector. Technological advancements and increased use of bio-based materials are vital. However, strategies to prolong the lifespan and maximize effective carbon storage of these materials are equally important. This paper aims to explore trends in the development of new bio-based materials (with focus on insulation materials) as alternatives to existing building materials supporting the global shift towards carbon neutrality. The study objectives are to investigate and map the main areas and types of new bio-based building materials, with equal interest in strategies for reuse and recycling to extend stored carbon and reduce emissions. The results show that bio-based material could have additional benefits of energy storage, carbon sink and recycling at end-of-life. However, most assessments of these materials focus on a narrow cradle-to-gate scope and thermal conductivity as only insulation material property. Hence, to further advance on bio-based insulation materials, there is a need for further development and harmonization of life cycle assessment methods that include these aspects and can be implemented in standards and regulations.

• Haugen, Øystein; Klikovits, Stefan; Andersen, Martin Arthur; Beaulieu, Jonathan; Bordeleau, Francis & Denil, Joachim [Vis alle 7 forfattere av denne artikkelen] (2025). DarTwin made precise by SysMLv2 – An Experiment. I Kessentini, Marouane; Sahraoui, Houari & Ali, Shaukat (Red.), 2025 ACM/IEEE 28th International Conference on Model Driven Engineering Languages and Systems Companion (MODELS-C). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331579906. s. 322–332. doi: 10.1109/models-c68889.2025.00051. Fulltekst i vitenarkiv Vis sammendrag

  The new SysMLv2 adds mechanisms for the built-in specification of domain-specific concepts and language extensions. This feature promises to facilitate the creation of Domain-Specific Languages (DSLs) and interfacing with existing system descriptions and technical designs. In this paper, we review these features and evaluate SysMLv2’s capabilities using concrete use cases. We develop DarTwin DSL, a DSL that formalizes the existing DarTwin notation for Digital Twin (DT) evolution, through SysMLv2, thereby supposedly enabling the wide application of DarTwin’s evolution templates using any SysMLv2 tool. We demonstrate DarTwin DSL, but also point out limitations in the currently available tooling of SysMLv2 in terms of graphical notation capabilities. This work contributes to the growing field of Model-Driven Engineering (MDE) for DTs and combines it with the release of SysMLv2, thus integrating a systematic approach with DT evolution management in systems engineering.

• Alskaif, Tarek; Babur, Önder; Bordeleau, Francis; Cleophas, Loek; Combemale, Benoit & Denil, Joachim [Vis alle 12 forfattere av denne artikkelen] (2025). Evolution at the Core of Digital Twin Engineering. I Kessentini, Marouane; Sahraoui, Houari & Ali, Shaukat (Red.), 2025 ACM/IEEE 28th International Conference on Model Driven Engineering Languages and Systems Companion (MODELS-C). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331579906. s. 210–216. doi: 10.1109/models-c68889.2025.00119. Fulltekst i vitenarkiv Vis sammendrag

  Engineering Digital Twins (EDT) presents a multifaceted challenge that extends beyond managing the lifecycle of a Digital Twin (DT) to include its continuous, dynamic interaction with the lifecycle of the actual object, system, or process it represents, referred to as the Actual Twin (AT). The relationship between the lifecycles of DT and AT necessitates a rethinking of the software development lifecycle of DTs. This vision paper examines the deeply intertwined lifecycles of DT and AT, arguing that effective methods for EDT must embrace the mutual and adaptive evolution of both over time. We propose placing evolution at the core of EDT. We identify key triggers of DT evolution, examine the engineering dimensions involved, and explore how the best practices, technologies, and tools of DevOps can support this evolution. Finally, we discuss current challenges and opportunities in the field. This paper serves as a call to action for the EDT community to adopt evolution as a crucial factor and core principle in EDT.

• Ahmad, Shahzad; Mukherjee, Surajit; Chanda, Sukalpa; Palaiahnakote, Shivakumara; Pal, Umapada & Pedersen, Marius (2025). PHASE: Physiological Dynamics-Based Attention for SpO$$_2$$ Estimation. I Petersen, Jens & Dahl, Vedrana Andersen (Red.), Image Analysis: 23rd Scandinavian Conference, SCIA 2025, Reykjavik, Iceland, June 23–25, 2025, Proceedings, Part I. Springer Nature Switzerland. ISSN 9783031959103. s. 30–44. doi: 10.1007/978-3-031-95911-0_3. Fulltekst i vitenarkiv Vis sammendrag

  The two-volume set 15725 and 15726 constitutes the refereed proceedings of the 23rd Scandinavian Conference on Image Analysis, SCIA 2025, held in Reykjavik, Iceland, during June 23–25, 2025. The 60 revised full papers presented in these proceedings were carefully reviewed and selected from 98 submissions. Part I: Machine learning and deep learning; Segmentation, Grouping, and Shape; Vision for robotics and autonomous vehicles; Biometrics, faces, body gestures and pose; 3D vision from multiview and other sensors; and Vision applications and systems.

• Gurav, Aniket Anand; Chanda, Sukalpa & Pedersen, Marius (2025). ResNet-TPP: A Parallel PHOC-PHOS Framework for Zero-Shot Handwritten Word Recognition in Low-Resource Scripts. I Jin, Lianwen; Zanibbi,, Richard & Eglin, Veronique (Red.), Document Analysis and Recognition – ICDAR 2025 Workshops: Wuhan, China, September 20–21, 2025, Proceedings, Part I. Springer Nature. ISSN 9783032046147. s. 245–259. doi: 10.1007/978-3-032-09368-4_15. Fulltekst i vitenarkiv
• Larsen, Elin Strand & Ihlen, Øyvind (2025). Political Crisis Communication and the Rhetorical Arena: An Analysis of President Zelensky's War Addresses and Social Media Use. I Schwarz, Andreas; Seeger, Matthew & Kim, Sora (Red.), The Handbook of International Crisis and Risk Communication Research. Wiley-Blackwell. ISSN 9781394180837. doi: https:/onlinelibrary.wiley.com/doi/10.1002/9781394180844.ch15. Fulltekst i vitenarkiv Vis sammendrag

  International political crises are often immensely complicated, involving many different actors communicating in different arenas and employing different formats. Legacy media and social media can fuel the crises, but crisis communicators seeking to manage them can also take advantage of such media. In this chapter, we discuss international political crisis and crisis communication using the so-called rhetorical arena model and draw on empirical examples from the war addresses and social media use of the Ukrainian President in the ongoing war with Russia.

• Lindell, Trym; Ramstad, Ola Huse; Sandvig, Ioanna; Sandvig, Axel & Nichele, Stefano (2025). Information Encoding and Decoding in In Vitro Neural Networks on Micro Electrode Arrays through Stimulation Timing. International Journal of Unconventional Computing. ISSN 1548-7199. 20(3), s. 155–195. doi: https:/www.oldcitypublishing.com/wp-content/uploads/2025/09/IJUCv20n3p155-195.pdf. Fulltekst i vitenarkiv Vis sammendrag

  A primary challenge in utilizing in vitro biological neural networks for computations is finding good encoding and decoding schemes for inputting and decoding data to and from the networks. Furthermore, identifying the optimal parameter settings for a given combination of encoding and decoding schemes adds additional complexity to this challenge. In this study we explore stimulation timing as an encoding method, i.e. we encode information as the delay between stimulation pulses and identify the bounds and acuity of stimulation timings which produce linearly separable spike responses. We also examine the optimal readout parameters for a linear decoder in the form of epoch length, time bin size and epoch offset. Our results suggest that stimulation timings between 36 and 436ms may be optimal for encoding and that different combinations of readout parameters may be optimal at different parts of the evoked spike response.

• Demi, Selina; Guttormsen, Bjørnar; Reinaas, Kevin Forberg; Hansen, Mats & Sand, Sofie (2025). Advancing Cybersecurity Education in the Era of Generative AI. NIKT: Norsk IKT-konferanse for forskning og utdanning. ISSN 1892-0713. 37(4). doi: 10.5324/42gawg86. Fulltekst i vitenarkiv Vis sammendrag

  Artificial intelligence advancements have recently received atten-tion in both academic and industrial settings, demonstrating disruptive potential in many domains, including education. Previous studies have shown that AI-driven tools can transform conventional learning methods. This paper focuses on identifying how generative AI (GAI) tools are leveraged by cybersecurity stu-dents. A survey has been designed based on cognitive process dimension of Bloom’s taxonomy and distributed to students studying cybersecurity in Norway. The results of the survey were illustrated with a set of use cases that demonstrated how Microsoft Copilot can enable students with skills ranging from lower-order knowledge-based skills to higher-order competence-based skills.The applied use cases presented in this paper indicate that GAI tools can en-hance lower- and higher-order thinking skills of cybersecurity students and con-tribute to bridging the gap between theoretical knowledge and competence-based skills needed to face real-world cybersecurity challenges. However, the results of the survey indicated that this potential has not yet been fully utilized by students, as they use such tools mainly for tasks that require lower-order thinking skills, such as remembering and understanding cybersecurity concepts, and less for more advanced cognitive tasks, such as creating role-playing scenarios. These preliminary findings can be useful to trigger discussions on how to improve fu-ture cybersecurity curricula to include applied GAI use cases and align them with learning outcomes. Ultimately, future efforts can be dedicated to developing a roadmap as a collaborative effort among policy-makers, cybersecurity academics and practitioners on the trustworthy incorporation of applied GAI in cybersecu-rity education.

• Peters, Carl-Hendrik & Sánchez-Gordón, Mary (2025). Reservoir Computing as a Promising Approach for False Data Injection Attack Detection in Smart Grids. NIKT: Norsk IKT-konferanse for forskning og utdanning. ISSN 1892-0713. 37(3). doi: 10.5324/hyp73c69. Fulltekst i vitenarkiv Vis sammendrag

  Smart grids are highly digitalized electricity networks that are increasingly at risk from False Data Injection Attacks (FDIAs), which threaten grid stability. Although traditional machine learning techniques are well‑established for detecting such anomalies, a neuromorphic computing approach like Reservoir Computing (RC) offers a promising alternative. Therefore, this study aims to explore RC for FDIA detection in Smart grids by conducting a Systematic Mapping Study to identify the current research trends and a Benchmark Evaluation of seven models across 21 simulated attack scenarios. The evaluation included the following three metrics: accuracy, robustness, and training‑time efficiency. Findings show that the two traditional approaches included in this evaluation lead with up to 99 % accuracy and minimal training time. Among the five RC approaches, a Delayed Feedback Reservoir with Latency Encoding and Multi-Layer Perceptrons readout achieved ~93 % accuracy (albeit with longer training), while pairing that reservoir with a Logistic Regression readout delivered ~86 % accuracy in under 0.02 s. These findings suggest that appropriately coded and read‑out RC models can serve as robust, resource‑efficient solutions for real‑time FDIA detection.

• Aliyari, Mostafa; Droguett, Enrique Lopez; Barabady, Javad & Ayele, Yonas Zewdu (2025). Ant colony optimization of hyper-parameters in multi-head attention layer for time series forecasting. Applied Soft Computing. ISSN 1568-4946. 183. doi: 10.1016/j.asoc.2025.113562. Fulltekst i vitenarkiv Vis sammendrag

  Transformer-based models excel in capturing complex temporal dependencies in time-series forecasting, yet their permutation-invariant self-attention mechanism overlooks the critical role of sequence order. To address this limitation, we propose ‘‘ACOFormer’’, a novel multi-head attention layer optimized through Ant Colony Optimization (ACO) for time-series prediction. Tuning Transformer hyper-parameters for power load forecasting is challenging due to the vast configuration space, which grows exponentially with the number of tunable parameters, rendering exhaustive searches impractical and resource-intensive. ACOFormer overcomes this challenge by integrating a dual-phase ACO algorithm with K-means clustering, achieving an 20.59% reduction in Mean Absolute Error (MAE) compared to a baseline Transformer model. Facing a configuration space exceeding 82 million permutations, ACOFormer employs a dual-phase iterative approach. Initially, cluster-based exploration leverages local pheromone updates to guide probabilistic hyper-parameter selection. Subsequently, global pheromone updates expand the search across the most promising hyperparameter regions, informed by aggregated insights. This balanced approach enhances tuning efficiency and optimizes computational resource utilization, enabling ACOFormer to capture temporal nuances effectively. Empirical results demonstrate ACOFormer’s superior performance, with a 12.62% MAE reduction and a 10.54% improvement in Mean Squared Error (MSE) compared to Informer, alongside 27.33%–29.4% MAE reductions and 38.58%–47.51% MSE improvements against MICN, Reformer, and Autoformer over a two-hour forecast horizon. These results underscore ACOFormer’s effectiveness in improving the precision of Transformer-based models for time-series forecasting and highlight its potential for resource-efficient applications across diverse industries.

Se alle arbeider i NVA

• Nehaniv, Chrystopher L.; Lewis, Peter R.; Nichele, Stefano; Čejková, Jitka; Salge, Christoph & Khan, Imran Qayym [Vis alle 7 forfattere av denne artikkelen] (2026). ALIFE 2026: Proceedings of the 2026 Artificial Life Conference. MIT Press. doi: https:/direct.mit.edu/isal/isal2026/volume/38. Fulltekst i vitenarkiv
• Palanichamy, Manikandan (2026). Real-World Electrical Engineering Case Studies. John Wiley & Sons. ISBN 9781394359431. 352 s. doi: 10.1002/9781394359462. Fulltekst i vitenarkiv Vis sammendrag

  This textbook describes real world engineering case studies included within follow a common, intuitively organized format. Each case begins with a brief synopsis outlining the example’s subdiscipline, industry, and stakeholders. It then moves on to discuss the case’s background and technical details, followed by a succinct problem statement, a detailed analysis and solution, the real-world outcome and an evaluation of the attempted solution, and key takeaways and teaching notes for instructors. Perfect for professional engineers and engineering consultants, Real-World Electrical Engineering Case Studies will also benefit science journalists and communicators, researchers, academics, and entrepreneurs.

• Karlsen, Kristine Høeg; Karlsen, Joakim & Leatherman, Devin (2026). Museet utenfor museet: oppsøkende formidling i skolen. C.B. Myrvold, L.Skregelid & R. Tronstad (Red.), Den kulturelle skolesekken: Forskning møter praksis (Kap. 20, s. 533–560). Cappelen Damm Akademisk. ISBN 9788202917814. doi: https:/cdforskning.no/cdf/catalog/book/286. Fulltekst i vitenarkiv Vis sammendrag

  In this chapter, we present findings from six regional museums in Norway. The study is anchored in a comparative investigation conducted during a research stay in Seattle, USA, where we identified a unique and well-developed model for museum outreach to schools and local communities—the Burke model. The study asks how Norwegian museums can take a more active role in students’ learning beyond the museums’ physical premises. Using the Burke model as a comparative frame, it seeks to highlight, describe, and document practices that are widespread but not well described in Norwegian museum research. The results provide insight into how museum outreach in Norway can be further developed, with DKS as a central actor. This involves bringing museum knowledge and resources out of traditional museum spaces, for example through city walks, thematic resource boxes with artifacts, or visits to archaeological sites. In the chapter, we offer a deeper understanding of collaboration models and identify potential areas for further development of museums.

• Esnoul, Coralie; Jee, Eunkyoung; Simensen, John Eidar; Ryan, Ita; Sánchez-Gordón, Mary & Lau, Nathan (2025). 2025 IEEE/ACM 6th International Workshop on Engineering and Cybersecurity of Critical Systems (EnCyCriS). IEEE (Institute of Electrical and Electronics Engineers). ISBN 9798331538101. doi: https:/ieeexplore.ieee.org/xpl/conhome/11029521/proceeding. Fulltekst i vitenarkiv

Se alle arbeider i NVA

• Lindahl, Sondre (2026). Fakta under press: Desinformasjon og dømmekraft. doi: https:/www.hiof.no/om/hms-sikkerhet-beredskap/sikkerhet-og-beredskap/sikkerhetsmaneden/arrangementer/seminarer/gjenkjenne-desinformasjon.html. Fulltekst i vitenarkiv
• Prøitz, Lin (2026). Ny studie: – For noen er det å være skeiv mer krevende enn å være adoptert, for andre er det motsatt. [Journal]. BLIKK. doi: https:/www.blikk.no/adopsjon-forskning-lin-proitz/ny-studie-for-noen-er-det-a-vaere-skeiv-mer-krevende-enn-a-vaere-adoptert-for-andre-er-det-motsatt/298439. Fulltekst i vitenarkiv Vis sammendrag

  Forskning på skeive utenlandsadopterte presentertes på Pride House tirsdag, hvor det også blir mulighet for erfaringsdeling.

• Kjøniksen, Anna-Lena (2026). The pathway to interdisciplinary research. Fulltekst i vitenarkiv
• Biedacha, Maciej; Tellnes, Lars G. F. & Kjøniksen, Anna-Lena (2026). pLCA Scenario Based Production Pathways of Mycelia-Wood Bio-Composite. doi: https:/wsenetwork.org/plca-scenario-based-production-pathways-of-mycelia-wood-bio-composite/. Fulltekst i vitenarkiv Vis sammendrag

  Novel bio-based construction materials and phase-change materials (PCMs) have attracted increasing attention due to their potential to reduce operational energy demand and improve building sustainability, and the potential to enlarge anthropogenic carbon sinks (Livne et al., 2022, Hawkins et al., 2023). However, current studies mainly assess their environmental performance through conventional life cycle assessment (LCA) based on laboratory-scale data. This approach has important limitations and may overlook environmental impacts associated with full-scale production (Alaux et al., 2024, Sacchi et al., 2022). To address this gap, this study applies a prospective, ex-ante LCA framework to evaluate the environmental performance of a novel bio-based composite.

• Tellnes, Lars G. F.; Biedacha, Maciej & Kjøniksen, Anna-Lena (2026). Valuation of carbon storage in public procurement. doi: https:/wsenetwork.org/valuation-of-carbon-storage-in-public-procurement/. Fulltekst i vitenarkiv
• Andersson, Gunnar (2026). Fra planstyring til lærende strategiledelse. Fulltekst i vitenarkiv Vis sammendrag

  Presentasjonen viser hvordan Hoshin Kanri kan brukes som en modell for lærende strategiledelse der strategi utvikles gjennom dialog, involvering og kontinuerlig læring. Gjennom Catchball-prosessen, X-matrisen og PDSA-syklusen knyttes langsiktige mål til daglig forbedringsarbeid på en måte som skaper både retning, forankring og læring i organisasjonen.

• Andersson, Gunnar (2026). Lean og Obeya i internasjonale utviklingsprosjekter. doi: https:/flowit.no/events/webinar-lean-utenfor-fabrikken-forbedring-i-kontor-tjenester-og-administrasjon/. Fulltekst i vitenarkiv Vis sammendrag

  Presentasjonen viser hvordan Lean og Obeya kan brukes for å skape struktur, læring og samarbeid i komplekse internasjonale utviklingsprosjekter. Med utgangspunkt i Interreg-prosjektet FREIIA illustreres hvordan visuell prosjektstyring, felles problemløsning og kontinuerlige læringsprosesser kan bidra til å håndtere utfordringer på tvers av land, fagmiljøer og interessenter. Erfaringene peker mot videre utvikling av Obeya som verktøy for styring av hele prosjektporteføljer.

• Andersson, Gunnar (2026). Samtaler på sosial innovasjon i innovasjonssnekka. doi: https:/www.visitfredrikstadhvaler.com/no/arrangement/kulturnatt-2026-hogskolen-i-ostfold-presentasjon-av-studentprosjekter-stillelesing-i-biblioteket-og-elvetur-med-innovasjonssnekke?tlp=8589263. Fulltekst i vitenarkiv Vis sammendrag

  Innovasjonssnekke. Vi putrer langs elva fra campus til Gamlebyen hvor Sten Helberg fra Trebåtkollektivet innleder om hvordan dem kopler kystkultur og Fartøyvern med sosial innovasjon. Prof. Gunnar Andersson fra HiØ og IPL er ordstyrer for praten. Avgang fra Smia (HiØ).

• Andersson, Gunnar (2026). Kan vi forske Oslofjorden ren? doi: https:/www.hiof.no/forskning/forskningsdagene/program/Program%202026/forskerne-kommer.html. Fulltekst i vitenarkiv Vis sammendrag

  I forbindelse med Forskningsdagene 2026 reiser ett knippe forskere fra HiØ ut i fylket for å møte ungdom i videregående skole med korte, engasjerende formidlingsopplegg om egen forskning. Kan forskning bidra til en renere Oslofjord? Professor Gunnar Andersson ved Innovasjon og Prosjektledelse - Høgskolen i Østfold diskuterte spørsmålet med elever i Askim basert på det pågående forskningsprosjektet i samarbeid med Hvaler. Prosjektet er støttet av Miljødirektoratet med 5,27 millioner kroner. Prosjektet kombinerer profesjonell opprydding av spøkelsesredskap og marin forsøpling i ytre Oslofjord med registrering, sortering og sirkulær bruk av materialene. Studenter deltar aktivt sammen med forskere og eksterne partnere. Under forskningsdagene diskuterte vi problemene i Oslofjorden. Problemer som har mange årsaker og berører mange aktører. Slike utfordringer kalles ofte wicked problems. De har ingen enkel løsning. De har ingen enkel oppskrift. Derfor må forskere, studenter, fiskere, skoler og offentlige aktører arbeide sammen. for å bygge gode løsninger over tid. Spørsmålet åpnet også for en samtale om hva det vil si å være forsker. Forskere stiller spørsmål, samler kunnskap og prøver ut nye løsninger. I eksternfinansierte forskningsprosjekter skjer dette ofte i tett samarbeid med arbeidslivet og samfunnet rundt oss.

• Seddighi, Gilda; Vrieler, Tina; Radlick, Rebecca Lynn; Corneliussen, Hilde G; Nessa, Benedicte & Prøitz, Lin (2026). Invisible labour in GenAI use among young adults with migration backgrounds outside work and education. Fulltekst i vitenarkiv Vis sammendrag

  In increasingly digitalised societies, artificial intelligence (AI) and generative AI (GenAI) in particular, play a growing role in structuring everyday practices, from information seeking to emotional and social navigation. While much scholarship frames GenAI as an empowering tool, emphasising issues of literacy and skills (Sabherwal & Grover, 2024), other research suggests that its use may also have negative implications for well-being (Hashem et al., 2025). Young adults in NEET situations (not in education, employment, or training), who often face intersecting challenges such as low completion rates of secondary education, socioeconomic disadvantage, and weak institutional and social support, are particularly vulnerable (Fyhn et al., 2021). Building on research on socio-digital inequalities, which demonstrates that digital outcomes are shaped not only by technical skills but also by access to supportive social relations and by the capacity to interpret (Schellewald, 2022), negotiate, and navigate algorithmic systems (Helsper, 2021; Livingstone, 2025), this paper illuminates how such inequalities may be reproduced through everyday uses of GenAI. This study explores how young adults with migrant backgrounds in NEET situations engage with GenAI in their everyday lives, focusing on how such use can be understood as a form of affective labour. Drawing on science and technology studies, particularly Susan Leigh Star’s work on invisible work and marginality, the paper illuminates GenAI use as embedded in everyday forms of affective labour, coping with isolation, and managing uncertainty in precarious life situations. Rather than focusing on access or skills alone, drawing on qualitative interviews with 25 young people aged 16–29 in Neet situations, this paper highlights how inequalities rooted in non-supportive social ties are reproduced through everyday GenAI practices of coping, care, and emotional regulation.

• Herrebrøden, Marte Karoline; Skafle, Ingjerd; Marentakis, Georgios & Nordahl-Hansen, Anders (2026). Unlocking the Potential of Hearing Protection Devices in Supporting Inclusive School Environments for Autistic Students. Fulltekst i vitenarkiv
• Lindahl, Sondre (2026). – Eg er meir bekymra, mykje meir bekymra enn for fem år sidan - hallingdolen.no. [Journal]. Internett. Fulltekst i vitenarkiv Vis sammendrag

  Terrorforskar Sondre Lindahl var uroleg då Hallingdølen intervjua han ti år etter terroren. Fem år seinare er bodskapen endå mørkare.

• Korseberg, Karl-Arne Næss & Holone, Harald (2026). Bør teknologi­studenter skikkethets­vurderes? Fulltekst i vitenarkiv Vis sammendrag

  Skikkethetsvurdering skal beskytte sårbare grupper mot uforsvarlig profesjonsutøvelse. Men hvem har ansvaret når vurderinger og beslutninger i økende grad formes av teknologi og automatiserte systemer utviklet av personer som selv ikke omfattes av slike ordninger?

• Robles-Bykbaev, Vladimir; Luján-Mora, Sergio; López-Nores, Martín; Tortosa, Salvador Otón; Sánchez-Gordón, Mary & Mendoza-González, Ricardo (2026). Artificial Intelligence and emerging technologies for inclusive and innovative education. Frontiers in Computer Science. 8, s. 1–3. doi: 10.3389/fcomp.2026.1892105. Fulltekst i vitenarkiv
• Prøitz, Lin (2026). 'You should be slim, smooth, preferably shorter, and a passive receiver’: A qualitative study of international LGBTQ+ adoptees’ embodied experiences of sexualised racialisation in Norway. Fulltekst i vitenarkiv Vis sammendrag

  Based on qualitative life-biographical interviews with international LGBTQ+ adoptees in Norway, this study examines participants’ embodied and subjective experiences of racialisation in general, and sexualised racialisation in particular. Drawing on affect theory, gender studies, critical race theory, and skin studies, we ask: How do international adoptees who challenge norms of gender and sexuality experience discrimination in general and sexualised racialisation in particular? And what subjectivities become possible for adoptees through the ways adoption, gender, sexuality, race, and ethnicity are experienced and narrated in both online and offline publics? To address these questions, we ‘follow the feelings’ in the data, focusing on how participants talk about skin and hair.

• Johansen, Frode Ramstad (2026). Skole og utdanning, Høgskolen i Østfold | Bedriftene vet at de kan få ubehagelige sannheter presentert. Fulltekst i vitenarkiv Vis sammendrag

  Debattinnlegget beskriver hvordan studenter ved Høgskolen i Østfold analyserer lokale bedrifter for å avdekke utfordringer i arbeidslivet. Gjennom intervjuer og UNIC-metoden fant de særlig problemer med informasjonsflyt, forbedringssystemer som ikke brukes i praksis, og mangelfull oppfølging av nyansatte. Forfatteren mener studentenes utenfrablikk gir verdifulle innsikter som ofte er vanskelige å se internt. Samarbeidet gir bedriftene nyttig dokumentasjon og studentene praktisk erfaring.

• Hauge, Bjørn Gitle (2026). UFO-jakten. [Journal]. VG. doi: https:/www.vg.no/nyheter/i/q6Ozj1/ufo-jakten. Fulltekst i vitenarkiv Vis sammendrag

  Hva er det egentlig som skjer på himmelen over Hessdalen?

• Herrebrøden, Marte Karoline; Skafle, Ingjerd; Marentakis, Georgios & Nordahl-Hansen, Anders (2026). Hearing Protection Devices for Autistic Students Sensitive to Auditory Stimuli: Enhancing School Inclusivity? doi: 10.13140/RG.2.2.15086.57923. Fulltekst i vitenarkiv Vis sammendrag

  Autistic students with auditory sensitivity often experience sensory overload in school due to insufficient adaptation of the auditory environment (Price & Romualdez, 2025). These challenges can negatively affect learning, academic performance, and social-emotional well-being (Keith et al., 2019; Poulsen et al., 2025). Hearing protection devices (HPDs) show potential to support autistic students in managing sensory environments (Pfeiffer et al., 2019; Kwong et al., 2025). Educational professionals require sufficient competence to balance benefits and drawbacks of these technologies to support inclusive practices. Despite the widespread use of hearing protection devices among autistic students, there is limited empirical research and guidance to inform their appropriate use (Herrebrøden et al., 2025). Objective: To explore educational professionals’ experiences and perspectives on the use of HPDs for autistic students in school, including their competence and systemic approaches. The overall aim is to increase awareness of how educational environments and HPDs can support inclusive practices for students with auditory sensitivities.

• Opsahl, André; Sundkvist, Charlotte Haugland; Thøgersen, Joachim Minge & Velvin, Jan (2026). Hva bør økonomistudenter lære? Et triangulert perspektiv. Fulltekst i vitenarkiv
• Palanichamy, Manikandan (2026). International Advisory Board member and Keynote speech on balancing Digital Transformation and Climate Commitments: The Role of Sustainable Data Centers. doi: https:/sites.google.com/view/3icre-2026/committees. Fulltekst i vitenarkiv Vis sammendrag

  As India pursues its ambitious Net Zero 2070 vision alongside rapid digital transformation through initiatives like Digital India, an often-overlooked challenge emerges: the environmental impact of data centers. Globally, data centers consume 1-2% of total electricity, with this figure rising exponentially due to cloud computing, artificial intelligence, and IoT expansion. This keynote examines the critical balance between India's digital ambitions and climate commitments, exploring sustainable solutions for data center operations. Key strategies include energy-efficient cooling systems, renewable energy integration leveraging India's solar and wind potential, AI-driven workload optimization, waste heat recovery through circular economy principles, and green hydrogen as backup power. The presentation emphasizes that achieving sustainable data centers requires collaborative efforts among researchers, industry practitioners, policymakers, and technology innovators. Furthermore, it highlights the crucial role of engineering education in preparing future professionals through case-based learning approaches. Aligning data center sustainability with India's National Green Hydrogen Mission represents both a technical challenge and an opportunity for innovation in the journey toward carbon neutrality.

• Palanichamy, Manikandan (2026). International Advisory Board and Keynote speech on how Electrical Engineering Case Studies Bridge the Gap Between Classroom and Real-World Power Systems. doi: https:/www.hiof.no/english/about/collaboration/educational-projects/green-energy-innovation-through-international-collaboration/arrangementer/h2resa-2026-global-conference.html. Fulltekst i vitenarkiv Vis sammendrag

  Engineering education faces a critical challenge in bridging the gap between classroom theory and real-world practice. Research on industrial requirements shows that problem-solving in ambiguous situations as the most critical competency. This gap becomes particularly evident as India pursues its Net Zero 2070 vision. The National Green Hydrogen Mission, with its allocation of 19,744 crores demands engineers who can translate hydrogen production theories into practical infrastructure. Hard-to-abate sectors like steel, cement, and transport require professionals who understand not just science but also the economic, regulatory, and implementation challenges of deploying green hydrogen at scale. Consider how other high-stakes professions prepare their practitioners. Pilots train on simulators, surgeons practice through clinical cases, and business leaders study real decisions. Engineers should be no different. Yet traditional education often prioritizes abstract problem-solving over contextual application, preparing graduates for implementing real-world renewable energy solutions. The solution lies in industrial case-based learning. Case studies transform concepts into solutions while developing higher-order thinking skills essential for engineering judgment. Case-based education delivers five outcomes, critical for India's energy transition. It builds applied knowledge in contexts like battery lifecycle management or hydrogen storage. It develops critical thinking about technology deployment. It cultivates problem-solving intuition across renewable projects. It fosters contextual thinking about regulatory and economic constraints. Finally, it enables learning from real-world implementations. Bridging this gap requires collective commitment. Students must seek real cases from India's renewable sector. As we gather at H₂RESA 2026, India stands poised with 1164 GW wind and 748 GW solar potential. Achieving Net Zero by 2070 depends on engineers translating theories into practical solutions. Case-based learning provides the essential bridge from theory to reality.

• Birkeland, Marius (2026). Gipsplater. I Bolstad, Erik (Red.), Store Norske Leksikon. Store norske leksikon. doi: https:/snl.no/gipsplater. Fulltekst i vitenarkiv Vis sammendrag

  Gipsplater er bygningsplater av gips. De vanligste gipsplatene har kartongkledd overflate, men det finnes også platetyper med fibermattearmert overflate eller fiberarmert kjerne. Gipsplater brukes særlig som kledning på vegger og i himlinger, men også i konstruksjoner med krav til lydisolasjon og brannmotstand.

• Birkeland, Marius (2026). Armeringskorrosjon. I Bolstad, Erik (Red.), Store Norske Leksikon. Store norske leksikon. doi: https:/snl.no/armeringskorrosjon. Fulltekst i vitenarkiv Vis sammendrag

  Armeringskorrosjon er rust på armeringsstål i armert betong. Når armeringen korroderer (ruster), dannes korrosjonsprodukter som har større volum enn stålet. Dette skaper trykk i betongen (sprengvirkning), slik at betongen får riss- eller sprekkdannelser og kan skalle av, samtidig som armeringen mister styrke fordi tverrsnittet reduseres.

• Birkeland, Marius (2026). Byggfukt. I Bolstad, Erik (Red.), Store Norske Leksikon. Store norske leksikon. doi: https:/snl.no/byggfukt. Fulltekst i vitenarkiv Vis sammendrag

  Byggfukt er overskudd av fukt i en bygningsdel når den står ferdig, sammenlignet med det fuktnivået den får etter en tid med normal bruk. Byggfukt er dermed fukt som finnes i bygningsmaterialer og konstruksjoner i byggeperioden, før bygningen tas i bruk.

• Birkeland, Marius (2026). Spennbetong. I Bolstad, Erik (Red.), Store Norske Leksikon. Store norske leksikon. doi: https:/snl.no/spennbetong. Fulltekst i vitenarkiv Vis sammendrag

  Spennbetong er en form for armert betong der spennarmeringen, vanligvis i form av enkelttråder eller flertrådige strenger av høyfast stål, forspennes (strammes opp på forhånd) for å forbedre konstruksjonens bæreevne og bruksytelse. Forspenningen gir en trykkspenning i betongen som motvirker strekkpåkjenninger fra egenlast og ytre laster.

• Birkeland, Marius (2026). Forspenning av betong. I Bolstad, Erik (Red.), Store Norske Leksikon. Store norske leksikon. doi: https:/snl.no/forspenning_av_betong. Fulltekst i vitenarkiv Vis sammendrag

  Forspenning av betong er en teknikk innen betongkonstruksjoner der armeringen påføres strekk slik at betongen settes i trykk før eller under belastning. Formålet er å redusere eller forhindre rissdannelse, begrense deformasjoner og muliggjøre større spennvidder enn i tradisjonelle slakkarmerte betongkonstruksjoner.

• Palanichamy, Manikandan; P, Rachchelle; M, Kesava & P, Karthikeyan (2026). Digital Shadow Framework for Degradation Assessment and RUL-Oriented Analysis of a PEM Fuel Cell System. Fulltekst i vitenarkiv Vis sammendrag

  This study presents a Digital Shadow–based framework for monitoring Proton Exchange Membrane (PEM) fuel cell degradation under dynamic Worldwide Harmonized Light-Duty Test Procedure (WLTP) operating conditions. A 1 kW PEM fuel cell stack was subjected to scaled WLTP load profiles, and synchronized electrical and thermal time-series data were continuously acquired. A physics-informed Electro-Thermal Stress Indicator was developed to capture operational stress during transient load variations. A Degradation Index (DI) was constructed to represent the cumulative health state of the system. Results show gradual DI increase indicating stable early-life degradation behavior, with Remaining Useful Life exceeding 95%. Voltage degradation remained minimal, suggesting limited irreversible electrochemical damage. The proposed data-driven framework enables real-time degradation monitoring without feedback control, establishing a foundation for predictive maintenance and lifetime assessment of automotive fuel cell systems.

• P, Rachchelle; Palanichamy, Manikandan; P, Karthikeyan & P, Govindaraj (2026). Experimental Characterization and Performance Evaluation of a PEM Fuel Cell under WLTP Driving Cycle Conditions. Fulltekst i vitenarkiv Vis sammendrag

  This study experimentally investigates the dynamic performance of a 1 kW Proton Exchange Membrane (PEM) fuel cell stack under Worldwide Harmonized Light-Duty Test Procedure (WLTP) driving cycle conditions, representing realistic automotive operation. The fuel cell was subjected to scaled WLTP dynamic power profiles, and synchronized electrical parameters including voltage, current, and power were measured. Results demonstrate stable operation with effective power delivery during rapid load transitions. A MATLAB-based Proportional-Integral (PI) controller was implemented to enhance dynamic power tracking during transient variations. Comparative analysis between uncontrolled and PI-controlled responses shows improved tracking accuracy of 3.72% and reduced deviations during load changes. Polarization characteristics confirm expected electrochemical behavior across the operating range. The integrated electro-thermal performance analysis validates the capability of PEM fuel cells to handle realistic automotive power demands, supporting their feasibility for zero-emission transportation and sustainable mobility systems.

• Palanichamy, Manikandan; Lin, Gwendolyn Claire; P, Rachchelle; M, Kesava.; P, Karthikeyan & Jain, I. P. [Vis alle 7 forfattere av denne artikkelen] (2026). Separating Reversible and Irreversible Operating Stress in PEM Fuel Cells under Dynamic Driving Conditions. Fulltekst i vitenarkiv Vis sammendrag

  PEM fuel cells used in transportation operate under continuously varying load conditions, where rapid current changes induce voltage fluctuations and temperature responses. Many studies evaluate this behavior using average performance indicators that describe overall trends but do not clearly distinguish temporary operating stress from stress that persists under dynamic conditions. As a result, identifying driving phases that impose critical operational stress remains challenging. The novelty of this work lies in the development of a recovery-based stress assessment approach that evaluates coupled electrical and thermal responses under realistic driving conditions. The analysis examines voltage and temperature deviations during dynamic load variations and assesses their recovery characteristics to distinguish between temporary and non recovering operational stress. Experimental voltage, current, and temperature data are obtained from a PEM fuel cell system operating under a Worldwide Harmonized Light Vehicles Test Procedure driving cycle. MATLAB based signal processing techniques analyze the dataset. Dynamic stress events are identified from simultaneous voltage and temperature deviations during load transients. Events that show rapid voltage and temperature recovery represent temporary operating stress, whereas events with delayed or incomplete recovery indicate non recovering operational stress. An operating stress index quantifies critical stress periods within the driving cycle. Results indicate that most dynamic stress events recover effectively. However, rapid acceleration combined with delayed thermal response produces sustained voltage deviation and incomplete temperature stabilization. Recovery behavior therefore serves as a reliable indicator for distinguishing reversible disturbances from persistent operational stress and supports durability assessment of PEM fuel cells in hydrogen powered transportation applications.

• Palanichamy, Manikandan; Lin, Gwendolyn Claire; P, Rachchelle; P, karthikeyan; Jain, I. P. & Lal, Chhagan (2026). AI-Enabled Green Hydrogen Strategies for Energy Transition and Climate Goals. Fulltekst i vitenarkiv Vis sammendrag

  Global decarbonization efforts need transformative clean energy solutions. Green hydrogen is recognized as a critical pathway for achieving climate neutrality, especially in hard-to-electrify sectors. Artificial intelligence (AI) offers remarkable potential to overcome the cost and efficiency barriers that currently constrain hydrogen's role in contributing to climate change. However, a significant gap exists between what AI can theoretically do and what it actually does in green hydrogen systems today. This gap limits policy-driven transitions toward carbon neutrality. Through a comprehensive review of AI techniques applied to green hydrogen production, this study proposes an integrated three-layer AI-driven digital twin framework that systematically addresses fragmentation in current approaches. The framework comprises of three layers such as, Layer 1 of multi-source data integration, Layer 2 of hybrid AI processing), and Layer 3 of explainable AI-driven outputs for human decision support and automated control. Component-level validation demonstrates that machine learning algorithms improve energy efficiency by up to 20%, while AI-driven predictive maintenance reduces maintenance costs by 50% and extends equipment lifespan by 25%. The proposed framework addresses critical challenges including data quality, model interpretability, and regulatory compliance through systematic implementation strategies. By integrating multiple AI techniques with physics-based constraints and explainability mechanisms, the framework provides a practical roadmap for researchers, engineers, and policymakers to accelerate cost-competitive green hydrogen deployment. The study calls for further validation in diverse industrial control applications to establish broader deployment guidelines.

• Torgersen, Geir (2026). Om oversvømmelser etter uværet. [Radio]. NRK. Fulltekst i vitenarkiv
• Johansen, Frode Ramstad & Andersson, Gunnar (2026). Holmen, Trosvik | Holmen kan bli et forbilde – men da må vi gripe muligheten nå. Fulltekst i vitenarkiv Vis sammendrag

  Fredrikstad kommune og Høgskolen i Fredrikstad bør vurdere et samarbeid om å utvikle Holmen som et levende laboratorium.

• Johansen, Frode Ramstad (2026). HiØs nye Prinsipper for KI og vurdering: Hva, hvorfor og hvordan? – Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  En time med felles kritisk dialog om HiØs nye prinsipper for bruk av generativ kunstig intelligens (GenKI) og vurdering (se lenke under). Webinaret starter med en kort presentasjon av prinsippene og prosessene både i forkant av, og underveis i utarbeidelsen. Et lite panel, med representanter fra alle fakultetene, åpner dialogen. I webinaret løfter vi problemstillinger, lufter frustrasjoner, og deler kunnskap - på kryss av fag og roller.

• Johansen, Frode Ramstad (2026). Prinsipper for bruk av generativ kunstig intelligens for studenter. Fulltekst i vitenarkiv
• Johansen, Frode Ramstad (2026). Ingeniører i team ...eller hvordan bli kjent med andre og samtidig trene teamutvikling. Fulltekst i vitenarkiv
• Lindahl, Sondre (2026). Noen grenser for KI. Fulltekst i vitenarkiv Vis sammendrag

  Foredrag om grenser for KI for politikere i Østfold Fylkeskommune

• Davidsen, Jan Grimsrud; Hellum, Vidar; Robbersmyr, Kjell Gunnar; Jakobsen, Mette Mo; Sjursen, Aslak Darre & Frøiland, Eirik [Vis alle 7 forfattere av denne artikkelen] (2026). Status for utvikling av automatisk sortering av fisk ved hjelp av kunstig intelligens. doi: https:/www.fhf.no/nyheter-og-arrangementer/arrangementer/pukkellaks-2026/. Fulltekst i vitenarkiv
• Lindahl, Sondre & Borgebund, Harald (2026). Ingen EU-debatt – en fundamental politisk svikt. doi: https:/www.dn.no/innlegg/politikk/eu/eos/ingen-eu-debatt-en-fundamental-politisk-svikt/2-1-1928747. Fulltekst i vitenarkiv Vis sammendrag

  Norske politikeres fravær i denne debatten undergraver Norges langsiktige interesser. De prioriterer kortsiktige gevinster.

• Johansen, Frode Ramstad (2026). Fredrikstad sentrum, Byutvikling | Fredrikstad sentrum: Hvordan samarbeid kan vekke sentrum. Fulltekst i vitenarkiv Vis sammendrag

  Leserinnlegget beskriver at Fredrikstad sentrum samlet sett viser positiv utvikling, samtidig som deler av gågata har lav aktivitet. Det vises til tidligere erfaringer og forskning som fremhever samarbeid mellom gårdeiere, næringsliv og kommune som avgjørende for sentrumsutvikling, og det pekes på behov for felles, konkrete tiltak framover. Tidligere studentarbeider trekkes inn.

• Palanichamy, Manikandan; Nordby, Reidar Johannes; Halvorsen, Thor-Erik & Aarø, Aksel Gluppe (2026). Development of a Differential Drive Ultrasonic Navigation System with Integrated Wireless Beacon for Autonomous Mobile Robotics. doi: 10.1109/ICCECE69715.2026.11493066. Fulltekst i vitenarkiv Vis sammendrag

  This paper investigates ultrasonic sensor technologies for autonomous navigation systems. It explores differential amplifier configurations, wireless beacon networks, and real-time signal processing techniques to enable accurate obstacle detection and person-following capabilities. A novel differential drive approach was implemented to enhance ultrasonic transmission power, while MEMS microphone technology provided cost-effective reception compared to traditional piezoelectric solutions. Wireless beacon positioning was integrated using ESP32 microcontrollers and XBee communication protocols to enable distributed sensing networks. Experimental validation demonstrated effective range performance with direct transmission reaching 9 meters and obstacle detection at 2 meters using Time of Flight calculations. Environment mapping algorithms achieved 0.5-meter spatial resolution for navigation applications. The research contributes fundamental design principles for ultrasonic navigation systems, demonstrating feasibility while identifying critical integration constraints. Results provide valuable insights for developing practical autonomous navigation solutions in robotics, smart mobility, and assistive technology applications.

• Palanichamy, Manikandan; Tsekouras, George & Gopalakrishnan, Sundar (2026). Machine Learning Applications in Intelligent Hydrogen Storage Systems for Enhanced Storage Efficiency and Safety. doi: 10.1109/IDCIoT67589.2026.11455894. Fulltekst i vitenarkiv Vis sammendrag

  This paper presents comprehensive artificial intelligence (AI) and machine learning (ML) applications to address critical challenges in hydrogen storage systems for clean energy transition. Traditional hydrogen storage faces limitations in efficiency, safety, and operational optimization. This research demonstrates how advanced AI algorithms transform storage system design and operation through intelligent composite tank optimization, material discovery via ML models, and real-time monitoring using artificial neural networks. Machine learning-based predictive maintenance enables proactive fault detection, while intelligent safety protocols utilize pattern recognition for autonomous threat assessment. Case studies across different economies showcase deep learning applications in learning for dynamic optimization. Results demonstrate ML-enhanced storage systems achieve superior performance in efficiency metrics, safety parameters, and adaptive response capabilities. These AI applications represent a paradigm shift toward intelligent hydrogen storage infrastructure, establishing ML-driven solutions as essential for large-scale hydrogen economy implementation and sustainable energy transition.

• Palanichamy, Manikandan (2026). AI-Driven Optimization of Sustainable Hydrogen Transportation Systems in the Global Energy Transition. doi: 10.1109/IDCIoT67589.2026.11455742. Fulltekst i vitenarkiv Vis sammendrag

  This study examines the transformative role of artificial intelligence (AI) in optimizing sustainable hydrogen transportation systems during the global energy transition. AI technologies demonstrate significant potential for enhancing operational efficiency and safety across various hydrogen delivery modes through advanced applications including machine learning (ML)-based route optimization, predictive maintenance algorithms, and real-time monitoring systems. Regional analysis across Europe, Asia Pacific, and the United States reveals diverse approaches to infrastructure development and regulatory frameworks adapted to local energy policies and market conditions. While AI implementation has successfully improved operational efficiency and risk management capabilities, significant challenges persist in infrastructure standardization and cross-network technology integration. This paper highlights critical technical barriers and policy implications influencing the scalability of AI-enhanced hydrogen transportation systems. These insights contribute to the growing knowledge base essential for developing robust, and sustainable hydrogen transport infrastructure while supporting the global clean energy transition. The research underscores the necessity of coordinated technological and policy approaches.

• Prøitz, Lin (2025). KI gjør hærverk i studentenes skriverom. doi: https:/www.f-b.no/ki-ved-hogskolen-i-ostfold-behov-for-langsom-tid-na/o/5-59-3780571. Fulltekst i vitenarkiv Vis sammendrag

  All mulighet til langsom tenkning og refleksjon knuses. Det er min første tanke etter å ha gått gjennom studentenes refleksjoner om bruk av KI ved Høgskolen i Østfold. De kjenner selv at de mister muligheten til å vokse og dannes langsomt. Et interessant funn når de sammenligner KI-genererte tekster og med studentskrevne tekster, er at de KI-genererte tekstene fremstår som om det ikke er noen hjemme i teksten. Det uttrykker studentene selv når de reflekterer over høstsemesterets oppgave i kommunikasjonsfag.

• Birkeland, Marius (2025). Fagkonsulentuttalelse – Konstruksjonsanalyse (Erduran, 2026). Fulltekst i vitenarkiv Vis sammendrag

  Faglig vurdering av manus for lærebok i konstruksjonsanalyse brukt i høyere utdanning.

• Birkeland, Marius (2025). Asbestsanering. I Bolstad, Erik (Red.), Store Norske Leksikon. Store norske leksikon. doi: https:/snl.no/asbestsanering. Fulltekst i vitenarkiv Vis sammendrag

  Asbestsanering er arbeid som gjennomføres for å hindre spredning av asbestfibre fra materialer som inneholder asbest. Sanering utføres ved fjerning, innkapsling eller innbygging, og må skje under strenge sikkerhetstiltak.

• Borgebund, Harald (2025). Militær opprusting, Klima og miljø | Klima må vike for våpen: Europas motvilje mot reformer og opprustning kan bli katastrofalt. Fulltekst i vitenarkiv Vis sammendrag

  Norge og Europa må ruste opp for å avskrekke angrep, og for å ha et effektivt forsvar. Klimapolitikk og reguleringer står i veien og må ofres for å trygge samfunnet.

• Bergsjø, Leonora Onarheim; Leonardsen, Ann-Chatrin Linqvist; Heggernes, Pinar; Said, Fahad Faisal; Engeness, Irina & Christoffersen, Trine Eker (2025). Kunstig eller intelligent? Høyere utdanning i identitetskrise. doi: https:/www.arendalsuka.no/programsoek/details/28499. Fulltekst i vitenarkiv
• Kivedal, Bjørnar Karlsen (2025). Climate risks and housing markets. Fulltekst i vitenarkiv
• Kivedal, Bjørnar Karlsen (2025). Nonperformance Factors in Wind and Brass Band Competitions. Fulltekst i vitenarkiv
• Dechsling, Anders; Cipriano, Christina; Børtveit, Line; Fjermestad, Krister Westlye; Karlsen, Ann-Charlott Beatrice Amundsen & Akerbæk, Tore Marius [Vis alle 8 forfattere av denne artikkelen] (2025). Examining the effectiveness of Second Step Early Learning versus free play on socio-emotional learning in kindergarten: a cluster randomized controlled trial protocol. doi: 10.1186/s13063-025-09047-7. Fulltekst i vitenarkiv
• Tomczak, Danuta Aniela (2025). Russland før og nå - utfordringer og forventninger. Fulltekst i vitenarkiv Vis sammendrag

  Foredrag på FoU gruppens seminar

• Tomczak, Danuta Aniela & Orban, Franck (2025). Polen etter valget i USA. Fulltekst i vitenarkiv
• Tomczak, Danuta Aniela (2025). Visjoner av samfunnet i fremtiden - før og nå. Fulltekst i vitenarkiv Vis sammendrag

  Dette er et innlegg på seminaret i Forskningsgruppen Business, Development & Governance

• Tellnes, Lars G. F.; Pamies, Ramon Francisco Porras & Kjøniksen, Anna-Lena (2025). Environmental Performance Indicators for Circular business models based on Eco-efficiency: Cases on New business models and Product-service systems. Fulltekst i vitenarkiv
• Biedacha, Maciej (2025). Comparative LCA of Mycelium and melamine-urea-formaldehyde binders (MUF) for medium density fiberboards. Fulltekst i vitenarkiv Vis sammendrag

  Background The growing interest in sustainable construction practices has prompted research into alternatives to conventional synthetic binders used in engineered wood products. This paper investigates the environmental impact of replacing melamine-urea-formaldehyde (MUF)—a widely used synthetic resin in medium-density construction boards—with mycelium, a bio-based binding agent derived from fungal biomass. This topic is important due to the high environmental footprint associated with MUF production and use, including formaldehyde emissions, reliance on non-renewable resources, and challenges in end-of-life disposal. In contrast, mycelium offers a renewable, biodegradable alternative that may reduce the carbon footprint and support circular material strategies in the building sector (Alaux et al. 2024; Candido et al. 2024; Maierhofer et al. 2024). While existing studies have examined the use of mycelium in low-density applications such as insulation panels, little is known about its suitability and environmental performance when used in structural or medium-density boards. There is currently a knowledge gap concerning the life cycle environmental performance of mycelium versus MUF in this context. This paper addresses that gap by conducting a comparative life cycle assessment (LCA) of MUF and mycelium binders intended for application in medium-density wood-based construction boards. The goal is to inform material choices in sustainable building design and contribute to the development of fully bio-based construction materials. Keywords: LCA, Mycelium, MUF, Medium density fibreboards. Experimental This study employs a comparative Life Cycle Assessment to evaluate the environmental performance and carbon footprint of alternative binders used in particle board manufacturing. The assessment follows the methodology specified in EN 15804+A2, ensuring compliance with current European standards for construction products. Two types of binders were analyzed: • Bio-based mycelium, representing an innovative, renewable alternative. • Melamine-urea-formaldehyde, a conventional fossil-based resin used widely in the wood industry. Due to the early-stage development of the MyCobuild project, inventory data for mycelium were obtained from peer-reviewed literature. In contrast, the life cycle inventory for MUF was based on data compiled during the 2024 development of an Environmental Product Declaration (EPD) for particle boards. A follow-up study incorporating primary data from laboratory-scale and pilot-scale production of mycelium is in an initial phase and primary data are expected to be collected within next six months. The functional unit is defined as 1 kg of dry binder material. The system boundary covers a cradle-to-gate scope, including raw material acquisition, binder synthesis, and associated emissions—particularly CO₂ emissions during the biological growth phase of mycelium. LCA modeling adheres to the principles of EN 15804+A2, with additional references to standard emission quantification protocols (e.g., EN ISO 14067 for carbon footprint of products, where applicable). These ensure methodological consistency and allow for robust environmental comparisons between the two binder systems. This comparative framework supports a scientifically grounded evaluation of emerging bio-based binders and contributes to ongoing efforts to decarbonize engineered wood products in line with circular economy goals. Results and Discussion The comparative LCA conducted in this study reveals a substantially lower carbon footprint for mycelium-based binders compared to conventional MUF formulations. Within the cradle-to-gate system boundary, the results for mycelium show net negative carbon emissions. This outcome is attributed to the biological process by which mycelium absorbs carbon from organic substrates during its growth phase. Although fungi lack the ability to photosynthesize, they play a significant role in carbon sequestration by metabolizing and storing carbon derived from decomposed organic matter. CO₂ emissions resulting from mycelial metabolic activity were accounted for in the analysis and represent approximately 20% of the total emissions associated with mycelium production (Livne et al. 2022; Nussbaumer et al. 2024). Despite these emissions, the overall incubation process is characterized by low energy demand, further contributing to the favorable environmental profile of mycelium. In comparison, MUF production is associated with significantly higher fossil-based carbon emissions. Specifically, the climate change impact category related to fossil sources is more than six times greater for MUF than for mycelium (see Figure 1). These findings underscore the potential of mycelium as a low-impact, bio-based alternative binder in particle board manufacturing. The study does not include later stages of life cycle however scenarios considered in the literature indicate that mycelium’s performance at and of life is better than considered alternative (Alaux et al. 2024). Figure 1. Figure presenting results. Conclusions This study demonstrates the promising potential of mycelium-based binders in reducing the carbon footprint of particle board production. The comparative LCA, conducted in accordance with EN 15804+A2, revealed that mycelium exhibits significantly lower fossil-based emissions than conventional MUF binders. Notably, the cradle-to-gate results for mycelium showed net negative carbon emissions, primarily due to its ability to sequester carbon from organic substrates during growth. Although metabolic processes in fungi contribute to some CO₂ emissions, these accounted for only a minor share of the total impact. The low energy demand of the incubation process further enhances the environmental performance of mycelium. Conversely, MUF production involves substantial fossil resource use and higher embodied energy. These findings highlight the environmental viability of mycelium as a binder in engineered wood products. The study also underscores the importance of expanding LCA research to include emerging bio-based materials in structural applications. Future work will incorporate laboratory-scale data to refine the inventory and validate the results under real production conditions. Overall, this research supports the transition toward more sustainable material choices in the wood industry.   References Alaux N, Vašatko H, Maierhofer D, Saade MRM, Stavric M, Passer A (2024) Environmental potential of fungal insulation: a prospective life cycle assessment of mycelium-based composites. Int J Life Cycle Assess 29(2):255–272. Bjånesøy S, Kinnunen A, Einarsdóttir H, Heinonen J (2023) Carbon storage in the built environment: a review. Environ Res: Infrastruct Sustain 3(4):042003. Candido A, Amiri A, Junnila S, Pittau F (2024) Mycelium-wood composites as a circular material for building insulation. Front Sustain Cities 6:1412247. Hawkins HJ, Cargill RI, Van Nuland ME, Hagen SC, Field KJ, Sheldrake M, Soudzilovskaia NA, Kiers ET (2023) Mycorrhizal mycelium as a global carbon pool. Curr Biol 33(11):R560–R573. Livne A, Wösten HA, Pearlmutter D, Gal E (2022) Fungal mycelium bio-composite acts as a CO₂-sink building material with low embodied energy. ACS Sustain Chem Eng 10(37):12099–12106. Maierhofer D, Alaux N, Vašatko H, Saade M, Stavric M, Passer A (2024) The influence of biogenic carbon assessment assumptions on biogenic global warming results – case study of an innovative mycelium-based composite block. In: IOP Conf Ser: Earth Environ Sci 1363(1):012060. IOP Publishing. Nussbaumer M, Karl T, Benz JP (2024) Quantification of fungal biomass in mycelium composites made from diverse biogenic side streams. Fungal Biol Biotechnol 11(1):20. Volk R, Schröter M, Saeidi N, Steffl S, Javadian A, Hebel DE, Schultmann F (2024) Life cycle assessment of mycelium-based composite materials. Resour Conserv Recycl 205:107579.

• Karlsen, Joakim (2025). Å nå ut tidlig: Digital rekonstruksjon og formidling av Gjellestadfunnet. Fulltekst i vitenarkiv Vis sammendrag

  Riksantikvaren, i samarbeid med Østfold fylkeskommune, inviterer til seminar for ansatte i Sametinget og regionalforvaltningen som arbeider med skjøtsel, formidling og tilrettelegging av arkeologiske kulturminner, og med prosjekter som har fått tilskudd fra Statsbudsjettet kap. 1429 post 70 (tidligere) BARK.

• Lindahl, Sondre (2025). Grenser for KI. doi: https:/shows.acast.com/grenser-for-ki. Fulltekst i vitenarkiv Vis sammendrag

  I denne podkasten skal vi gjennom 9 episoder utforske ulike grenser for KI. Du får høre fra ledende forskere på KI, eksperter på valgmanipulasjon og demokrati og ledende stemmer i teknologidebatten i Norge som i ulike episoder diskuterer hvilke grenser som finnes for KI, hva som kan gjøres og hva som bør gjøres.

• Lindahl, Sondre (2025). SOVER VI I TIMEN? OSS OG KUNSTIG INTELLIGENS. Fulltekst i vitenarkiv
• Larsen, Elin Strand (2025). Falske opplysninger, notoriske løgnere og faktasjekk. Valganalyse 2025. doi: https:/www.kristiania.no/forskning/forskningsgrupper/polcom/Valganalyse/valganalyse-2025/falske-opplysninger-notoriske-lognere-og-faktasjekk/. Fulltekst i vitenarkiv Vis sammendrag

  Faktasjekk i valgkamper handler ikke lenger bare om å etterprøve påstander fra politikere. I årets valgkamp prøvde faktasjekkerne å avsløre digitale påvirkningsforsøk og hvordan vi skal orientere oss i en stadig mer konfliktfylt verden.

• Badshah, Imtiaz; Wu, Hong & Khattak, Imran Ullah (2025). Annual confidential report of the civil service of Pakistan. Fulltekst i vitenarkiv Vis sammendrag

  Abstract: Performance appraisal is an important management tool used for assessing employee job performance, to make decisions regarding employee incentives and rewards, and providing feedback or training to develop employee skills and competencies. It is generally agreed that performance appraisal can benefit employees and organizations, but for that, the performance appraisal techniques and instruments need to be fair and objective. Employees may positively react to the performance appraisal systems if the design of such systems involve employee input and the evaluation corresponds to the respective job descriptions. The current study aims at examining the effectiveness of performance appraisal system in the public sector of Pakistan. The study is qualitative in nature and draws on the literature in ‘new public management’. The study led to explore that the process of performance appraisal of civil servants in Pakistan is highly subjective. It is used as a tool for making decisions regarding employee promotions, but the evaluations are mainly based on criteria irrelevant to merit because of the use of subjective and generic evaluative instruments. Since evaluations are entirely secret, that is why one of the core objectives of performance appraisal, which is performance improvement is not achieved through performance appraisal system. More specifically, the study indicates that factors such as an administrative culture rife with cronyism, high centralization, and conflict of interests between policy-makers may contribute to the design of a performance appraisal system that lacks fairness and objectivity. It transpired that the positive effects of the performance appraisal system in the Pakistani public sector are sparse because it is not designed in collaboration with employees, it lacks transparency and its utility for performance improvement is vague. We suggest practical implications and directions for future research.

• Badshah, Imtiaz & Imtiaz, Sajida (2025). Embracing Digital Transformation: The Future of Accounting profession in Norway. Fulltekst i vitenarkiv Vis sammendrag

  Abstract:- This study seeks to elucidate the impact of digital transformation on the accounting profession, specifically within the accounting environment in Norway. To achieve this, the study divides the accounting environment into four distinct perspectives as follows: 1. The Accounting profession, 2. Traditional academic institutions 3. Professional accounting institutions 4. Accounting regulations The advent of digital technologies and their impact on the future of the accounting profession is a multifaceted topic. To provide a deeper understanding of this complexity, this study will construct a theoretical framework by drawing insights from institutional theories, including those of Scott (1995), DiMaggio and Powell (1993), and Meyer and Rowan (1977), alongside structural theories from scholars like Mintzberg (1979) and Galbraith (1973). This exploratory study is designed to examine how digital technologies are influencing the future of the accounting profession. Data collection will involve interviews with respondents. Moreover, published documents can provide valuable data insights. To gain insights into the accounting curriculum at traditional academic and professional institutions, content analysis will be employed. The qualitative approach of this study facilitates a comprehensive exploration of the future trends of the accounting profession in Norway. The increased use of digital technologies such as blockchain, big data, could-based analytical tools and artificial intelligence are easing the job of accountants. Accountants are intensively using digital-based solutions for recording and tracking financial transactions, which have markedly reduced the efforts (and time) of accountants in recording and tracking financial transactions. These solutions significantly enhanced the accountant’s ability to prepare and communicate financial information to management for decision making. Additionally digital accounting solutions provide other benefits such as increased functionality, faster processing, improved external reporting. However, all above benefits are not without challenges such as control and information security.

• Haugen, Øystein (2025). Digital Twins: Models Everywhere. doi: https:/modelsward.scitevents.org/KeynoteSpeakers.aspx?y=2025#0. Fulltekst i vitenarkiv Vis sammendrag

  Finally, everybody embraces models - in the form of "Digital Twins"! What a relief for someone who has worked with models for many years in several different domains. But we have not conquered all challenges concerning digital twins, and this talk will look at digital twins in the light of experiences from working with models and model-driven engineering for a professional lifetime. What is a model in the era of digital twins and what should it be? Do we need specific languages for describing digital twins? How should digital twins evolve as their importance increases?

• Esnoul, Coralie; Jee, Eunkyoung; Simensen, John Eidar; Ryan, Ita; Sánchez-Gordón, Mary & Lau, Nathan (2025). Foreword to the 6th International Workshop on Engineering and Cybersecurity of Critical Systems (EnCyCriS). I Esnoul, Coralie; Jee, Eunkyoung; Simensen, John Eidar; Ryan, Ita; Sánchez-Gordón, Mary & Lau, Nathan (Red.), 2025 IEEE/ACM 6th International Workshop on Engineering and Cybersecurity of Critical Systems (EnCyCriS). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331538101. doi: 10.1109/encycris66464.2025.00005. Fulltekst i vitenarkiv Vis sammendrag

  This marks the 6th edition of the International Workshop on Engineering and Cybersecurity of Critical Systems (EnCyCriS) at the ICSE conference. For industrial critical systems, although the previous premises of increasing system interconnectivity, decentralization, and introduction of new, more intelligent technologies still hold true, there is an increased societal awareness with regards to cybersecurity. This has led to clearer regulation, sharper requirements, and higher expectations for industry. At the same time, the availability of readily deployable competence, methods, tools, and solutions is lacking, which should be considered a critical societal risk. In the current international political climate, cyber security, and safety of critical infrastructures across industry, are more important than ever before. The EnCyCriS workshop facilitates discourse and discussion amongst researchers, practitioners, and students who are working on challenges and solutions related to industrial critical infrastructure. It has a particular focus on sharing industry experience and project results pertaining to cyber threats on critical systems, secure systems engineering, and attack detection and response mechanisms.

• Pedersen, Asbjørn Olav (2025). Årsavslutning av kommuneregnskap. Fulltekst i vitenarkiv
• Pedersen, Asbjørn Olav (2025). Forum for økonomisjefer. Fulltekst i vitenarkiv
• Pedersen, Asbjørn Olav (2025). Forum for økonomisjefer. Fulltekst i vitenarkiv
• Pedersen, Asbjørn Olav (2025). Forum for økonomisjefer. Fulltekst i vitenarkiv
• Demi, Selina & Bjørnar, Guttormsen, (2026). A Gamified Approach to Security Awareness: Exploring Gamification and Data Visualization in Phishing Awareness Training. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Phishing remains one of the most prevalent forms of cyberattacks, and traditional awareness training approaches often struggle to sustain engagement or address recurring gaps such as poor retention and generic content. This thesis explores how gamification and data visualization can be used to enhance phishing awareness training, investigating both the current limitations and gaps in existing training approaches and how a prototype incoporating these elements could be designed and built to address them. An exploratory approach was used, consisting of a survey, the prototype and semi-structured interviews with cybersecurity professionals. The survey (n=23) gathered input on current training practices and contained questions that was later shaped into requirements for the prototype. This prototype incorporated story-driven scenarios, level-based progression, targeted feedback alongside other elements. The prototype was evaluated through interviews with three cybersecurity professionals. The findings suggest that current phishing awareness training faces challenges related to engagement and motivation, retention, feedback, progress tracking and realism. Combining gamification and data visualization elements in the prototype was positively received by the cybersecurity professionals as a promising approach to addressing these gaps.

• Demi, Selina & Sofie, Sand, (2026). Generative AI-enabled Cybersecurity Awareness and Training Framework for Organizations Operating in Critical Infrastructure. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Organizations operating in critical infrastructure face increasing cybersecurity risks, while traditional awareness and training programs often remain generic, static, and insufficiently adapted to employees’ roles and operational environments. This thesis investigates the limitations of current cybersecurity awareness training in critical infrastructure sectors and explores how generative artificial intelligence can support more effective, adaptive, and engaging training approaches. The study applied a qualitative multi-method research design combining a systematic literature review with ten semi-structured interviews involving professionals from sectors including defense, energy, healthcare, finance, the public sector, and research. The findings from both data sources were analyzed and synthesized to identify recurring challenges, generative AI-enabled opportunities, and requirements for responsible implementation. The results show that current awareness programs are frequently characterized by one-size-fits-all content, limited practical relevance, low engagement, and a persistent gap between cybersecurity knowledge and secure behavior. These limitations are particularly significant in critical infrastructure environments, where employees may operate under time pressure and where human error may have severe operational and societal consequences. The findings further suggest that generative AI may address several of these challenges through personalized and role-specific content, adaptive learning paths, realistic simulations, interactive scenarios, real-time feedback, and continuous content updates. Based on the combined findings, the thesis proposes a conceptual framework for generative AI-enabled cybersecurity awareness and training structured around the dimensions of knowledge, attitude, and behavior. The framework connects identified training challenges with corresponding generative AI capabilities and expected organizational outcomes. It also emphasizes governance, privacy, security, human oversight, and quality assurance as necessary conditions for responsible implementation. The study concludes that generative AI has considerable potential to strengthen cybersecurity awareness and training in critical infrastructure organizations. However, it should be treated as a support mechanism within a broader human-led and context-aware awareness strategy rather than as a standalone solution.

• Demi, Selina & Hovland, Pedersen, Håkon Halland (2026). Quantum Migration and Business Readiness in Norway. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  The advent of cryptographically relevant quantum computers poses a significant long-term threat to widely deployed public-key cryptographic algorithms. Post-Quantum-Cryptography (PQC) has emerged as the primary mitigation strategy, supported by recent international standardisation efforts. While the technical foundations of PQC are well established, there is limited empirical research on how organisations perceive the quantum threat, assess their readiness and approach the migration to quantum-resistant cryptographic solutions. This thesis investigates organisational awareness, risk perception and readiness related to post-quantum cryptography within a Norwegian organisational context. The study adopts a qualitative research approach and applies the Technology–Organisation–Environment (TOE) framework as an analytical lens to examine technological, organisational and enviro-nmental factors influencing preparedness for quantum migration. Data was collected through semi-structured interviews with selected organisational stakeholders across relevant industries. The analysis identifies awareness as a layered phenomenon: surface-level recognition of the quantum threat is near-universal among respondents, while operational understanding of harvest-now-decrypt-later (HNDL) exposure, of the organisation’s own cryptographic estate and of what migration would practically require - remains thin and unevenly distributed. Technological barriers centre on the absence of cryptographic inventories and the abstract, invisible character of cryptography as an organisational asset; organisational barriers include competing priorities, resource constraints and a distinct deficit in procurement competence; environmental influences are dominated by the advisory, non-binding character of current regulatory signals. Across these dimensions, the study identifies a set of deferral mechanisms – authority-signal, vendor-trust and compliance-waiting deferral – through which organisations justify postponing migration, together with a temporal framing pattern termed the Doomsday Paradox. The goal of this thesis is that the results can contribute to both academic research and practical cybersecurity governance by providing empirical insight into organisational decision-making related to post-quantum cryptography. It aims to inform practitioners, policymakers and researchers on the implications of these findings for PQC migration strategy, regulatory communication and organisational readiness assessment in Norway.

• Ogul, Hasan & Sonia, El Ouahabi Amaiua, (2026). LLM-Enhanced Reinforcement Learning for Dynamic JSSP. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Dynamic job shop scheduling requires real-time adjustments whenever the shop floor changes due to machine breakdowns, new job arrivals, processing-time variations, or other operational disruptions. Although reinforcement learning offers a promising approach for learning adaptive scheduling policies, it still faces challenges related to reward design, learning efficiency, and decision interpretability in dynamic environments. This thesis proposes an LLM-enhanced reinforcement learning architecture for dynamic job shop scheduling. The language model is used in three ways during training: it rates the actions the agent chooses, it changes the reward weights when a disruption happens, and it picks a high-level scheduling strategy. The system is tested in a simulated job shop with machine breakdowns, new jobs, and changing processing times, and is compared with twelve common dispatching rules and with a PPO reinforcement learning agent. The results show that the way the language model is used matters. Using it to choose a high-level strategy worked best. It reduced the average makespan by about 2.6\% compared with the best dispatching rule, and also finished more jobs on time. Rating individual actions and changing the reward weights helped with meeting deadlines, but did not shorten the schedule. These findings indicate that LLM guidance was most effective when used to select a high-level scheduling strategy, rather than to judge individual dispatching decisions.

• Sica, Arianna (2026). The ethics of emotionally intelligent social robots. Høgskolen i Østfold. ISSN 9788278254677. Fulltekst i vitenarkiv Vis sammendrag

  Advances in Artificial Emotional Intelligence (AEI) are enabling the creation of social robots that appear to recognize and respond to human emotions. These developments open new possibilities for companionship, care, and intimacy, but they also raise urgent ethical and social questions. If robots can convincingly simulate empathy and affection, what does this mean for trust, authenticity, and the nature of human relationships? Could such machines provide meaningful support for those who are isolated, or might they foster dependency, manipulation, and the displacement of human connections? This dissertation investigates these challenges by focusing on social robots designed for companionship and intimacy, including sex and love robots. It examines how AEI blurs boundaries between humans and machines, reshaping how emotions and relationships are experienced and understood. The project is therefore an ethical analysis of AEI in social robots, aimed at clarifying key concepts, analysing core ethical questions, and preparing the ground for responsible evaluation and governance of the technology. I argue that AEI social robots present four main ethical challenges. First, they raise questions of authenticity and deception. Robots cannot truly feel emotions, yet their design encourages users to believe that they do – what does it mean to form relationships built on such simulation? Second, they problematize trust. Robots may be reliable, but can it ever make sense to extend feelings of trust to a machine, and if so, under which conditions? Third, they blur the boundaries of moral status. If people treat AEI robots as relational partners rather than objects, should our ethical frameworks change in response, and what social and ethical consequences follow? Finally, in the case of sex and love robots, AEI intensifies debates around intimacy and sexuality. Could such robots provide genuine comfort and care, or do they risk deepening dependency and isolation? Taken together, these challenges suggest that AEI social robots may not simply expand human emotional possibilities, but also fundamentally alter them. Whether this alteration supports or undermines human flourishing remains an open question – one that this dissertation seeks to explore in detail.

• Marentakis, Georgios & Vakhajevitsj, Mahauri, Hassan (2026). Mobile Sound Classification with Multimodal Feedback: Evaluating Visual Representations for Environmental Sound Awareness. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Headphone users with active noise cancellation are acoustically isolated from their environment, unable to hear sounds that may be important for safety or social awareness. This thesis addresses this problem by designing, implementing, and evaluating a mobile sound awareness system that classifies environmental sounds in real time on a mobile device and notifies users through multimodal feedback. The system introduces two novel design elements that consist of animated action-based visual representations that depict the physical actions producing each sound, and a binary routine–urgent distinction that communicates the urgency of a detected sound. Two within-subjects’ experiments were conducted with a combined 31 participants. Experiment 1 compared three visual representation formats (text-only, static icon and animation). All formats achieved near-ceiling accuracy when auditory context was available. Under visual-only conditions with brief exposure, animation and static icon outperformed text-only. Animation produced the fastest reaction times, highest subjective ratings, and was preferred by 14 of 18 participants overall. Experiment 2 compared three notification feedback conditions (visual-only, visual-plus-haptic and visual-plus-audio) during a realistic dual-task scenario where participants read articles while monitoring sound notifications. Visual-only notifications produced a 31% miss-rate, with five participants missing half or more of target notifications despite high reading comprehension. Adding haptic feedback lowered the miss-rate, but the improvement over visual-only was not statistically significant. Audio feedback through the headphones achieved 100% detection across all participants, significantly outperforming visual-only. Audio was rated highest on trust and effectiveness without being perceived as significantly more disruptive. Ten of 12 participants preferred audio for daily use. The findings demonstrate that visual-only notifications are insufficient for acoustically isolated users during cognitively demanding tasks, and that audio feedback delivered through the very headphones that create the isolation provides the most effective solution channel. The thesis contributes empirical evidence for the design of mobile sound awareness systems, including the first evaluation of animated action-based representations for environmental sound visualization and evidence that audio feedback through headphones restores environmental awareness without significant disruption.

• Chanda, Sukalpa & Holstad, Johansson, Magnus (2026). Visual Similarity Is Not Enough. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  In industrial defect search, the most visually similar image is not always the most relevant one. Images from the same object category can look almost identical, while showing different defect states. This thesis studies how a normal reference image and a short text defect description can be combined into one retrieval query, so the search is guided by both object context and defect intent. The proposed method uses frozen DINOv3 patch features, CLIP text features, and a trainable text-guided combiner. The final model combines concat fusion and sparsemax cross-attention to preserve the normal image context while conditioning the query on the requested defect. It is evaluated on MVTec AD and VisA using in-domain retrieval, model ablations, cross-dataset transfer, and combined-dataset training. In-domain, the method retrieves the intended defect with around 0.65 R@1 on MVTec AD and 0.41 on VisA. This shows that a normal reference image and a defect prompt can form a useful retrieval query for known defect search. Performance is strongest when defect labels correspond to visually consistent groups. Among the attention-normalization variants, sparsemax gave the most consistent results in the tested setup, with a small effect on MVTec AD and a clearer gain on VisA. Label-aware loss and combined-dataset training helped one dataset but hurt the other, while cross-dataset transfer was weak in both directions. The thesis concludes that composed retrieval is a promising tool for searching known industrial defect examples, but that reliable generalization across defect datasets remains an open problem.

• Magnusson, Lars Vidar & Morgan, Waage, (2026). Using Generative AI Base Models and Techniques for Generating Graphical Assets in 2D Games. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  This thesis investigates how free and open-source generative AI models and tech- niques can be used to generate graphical assets for 2D games. It also examines what challenges limit their use in a complete asset pipeline. The study uses con- trolled experiments with locally hosted di!usion-based workflows in ComfyUI. It generates characters, backgrounds, objects, and short animations under fixed conditions. The outputs are evaluated through manual assessment. Two criteria are used: quality and consistency. Quality is defined as structural integrity and usability. Consistency is defined as alignment with a predefined cartoon style. The results show that generative AI can produce usable 2D assets. Backgrounds and simple objects achieve higher success rates than characters. Workflows with guidance for structure and style produce better results. The results also show that consistency across asset categories remains di”cult to maintain. Outputs vary between runs. Animation generation produces unreliable results. These findings show that current tools can support parts of the 2D asset pipeline. They require controlled workflows, repeated generation, and human selection to achieve a consistent and production-ready result.

• Lampropoulos, Georgios; Schucan, Christian & Løve, Hansen, Ole Marcus (2026). Incident Response Training Serious Game in Virtual Reality: Exploring Game Mechanics for Cybersecurity Preparedness to Support Business Sustainability. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  As businesses increasingly depend on digital infrastructure, cybersecurity has become a critical enabler for long-term economic and social sustainability. Related work has established strong connections between cybersecurity and the United Nations (UN) Sustainable Development Goals (SDGs), and cybersecurity labour shortages make effective education and training methods more important than ever. Virtual Reality (VR) and serious games have been supported by previous research as particularly effective approaches for achieving this, demonstrating equal or superior outcomes compared to traditional learning methods in terms of engagement, motivation, and knowledge acquisition. However, there is a lack of research connecting immersive technologies, incident response training, and specific SDG targets. This thesis addresses this gap by exploring how an incident response training serious game in VR can enhance cybersecurity preparedness and support business sustainability. The study combines the design, development, and evaluation of a VR serious game prototype for incident response training with a literature analysis connecting the findings to the SDGs and business sustainability. The prototype was informed by best practices in cybersecurity education, gamification, and VR design, using the COFELET, Stenmap, and SGDDEdu frameworks, and implements task-based and stressor-based game mechanics derived from defined learning objectives. The evaluation was conducted with nine Norwegian higher education students using a custom survey tool, including pre- and post-test knowledge assessments, structured observations, and surveys on both the game mechanics and sustainability. The results provide insights into which game mechanics are most effective for incident response training and how cybersecurity connects to the SDG agenda. Text Button Interactions and Readable Instructions were found to be the most educational game mechanics, with Finding Object and Sorting Objects being the most engaging and motivating. Knowledge assessment scores improved across all assessed areas after a single play-through, and the participants rated the game positively in all surveyed aspects. The sustainability findings identified SDG target 8.8 (promoting safe and secure working environments) as the most strongly connected to incident response training, receiving a unanimous maximum rating, with SDGs 8, 9, 9.1, 16, and 16.4 also rating highly. The findings of this study contribute practical guidelines for VR serious game design for incident response training, strengthen existing research connecting cybersecurity to the SDG agenda, and highlight how immersive incident response training can be designed to directly support the social and economic sustainability of businesses.

• Katta, Vikash & Radin, Morik, (2026). Real-Time Deepfake Video Detection: An Evaluation of Open-Source Methods on Consumer-Grade Hardware. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  With the rapid advancement and accessibility of real-time deepfake technology, robust deepfake detection is crucial for mitigating fraud and social engineering attacks. This thesis evaluates whether open-source, real-time detection on consumer-grade hardware is a viable approach for addressing the growing threat. Through a comprehensive narrative review of the deepfake detection field, two open-source detection models were selected for evaluation: a partially pre-configured solution based on InceptionResNet V1 used as a baseline and an image-based detection model named Tiny-LaDeDa, which was adapted for video detection through a frame-extraction method. Both models were evaluated on two real-world deepfake datasets (Deepfake-Eval-2024 & Celeb-DF V2) to assess classification performance and real-time capabilities on Apple M1 hardware. The findings suggest that real-world deepfake content subjected to realistic media degradation significantly reduces the classification performance of open-source, real-time detection models. The results confirm that academic benchmark scores are not reliable predictors of real-world performance. However, the baseline model demonstrated strong performance on controlled deepfake content resembling video conferencing scenarios, suggesting viability in specific contexts if false positive rates are reduced. Furthermore, the evaluations revealed that consumer-grade hardware is capable of running these models in real-time, with both model implementations achieving latencies under 80 ms per frame on the tested Apple M1 hardware. This indicates that the current limitations for real-time detection lie in model robustness, not hardware capabilities. Future work should further evaluate reliability under real-world conditions and improve the models to reduce false positives without increasing latencies.

• Lampropoulos, Georgios & Nayef, Chaaban, Omar (2026). Designing and Evaluating a 3D Interactive Educational Game to Teach Programming Concepts. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Programming can be challenging for beginners because the concepts are non-tangible, abstract, and difficult to visualize. This thesis focuses on the design, development, and evaluation of The Farm, a 3D interactive educational game developed in Unity to help beginners understand basic JavaScript concepts by combining game-based learning with visual metaphors. The game used a farm as its environment, so learners could relate to familiar objects and interactions that represented concepts such as variables, arrays, objects, conditionals, loops, and functions. A mixed-methods evaluation was conducted with 10 participants who interacted with the game and then completed a questionnaire to assess usability, engagement, and the overall educational experience, and to gather feedback through open-ended questions. Participants liked the interactive approach to learning and found the game usable, engaging, and helpful, and rated most design choices highly. The thesis concludes that metaphor-based 3D game environments can be useful as supplementary learning tools for introductory programming education.

• Khakpour, Alireza; Schucan, Christian & Kedi, Krasniqi, (2026). A Scenario-Based Analysis for Precious Metal Recovery in Norwegian E-Waste Systems. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  This thesis investigated the associations between precious metal recovery, sustainability impacts, and electronic waste (e-waste) recovery strategies in Norway’s e-waste system. E-waste has increased largely due to the growing number of electronic devices, and it consists of high amounts of important precious metals like gold, platinum, and silver. To manage the issue of large amounts of metals getting lost, I established an analytical decision-support artifact to assess the metal recovery rate in Norway’s e-waste system. This artifact consisted of four scenarios in total that were based around recovery and collection rate adjustments. With the first scenario representing the current scenario, the second scenario represented the scenario of a 15% recovery rate increase, while the two last scenarios represented collection rate increases (of 20% and 50%). In order to calculate and assess how much precious metals were recovered and lost, I integrated three public datasets which information that consisted of e-waste generation, metal composition, and recycling rates. This allowed me to calculate the lost and recovered amounts of gold, platinum, and silver, specifically within the timeframe from 2008 to 2018. A structured approach was used through Python and SQL/MySQL to clean and combine the data, model the scenarios, and access the sustainability impacts. Whereas Power BI, and DAX measures, were utilized to visualize all its key aspects displaying information such as the precious metal recovery rate, sustainability impacts, and the availability to toggle and explore the different scenarios and their outputs. By enhancing the collection and recovery rates, and what their impacts are on recovery rate outputs, environmental factors (mining avoided and CO2 avoided), and the economic value, the model uses all this to assess the four scenarios provided. The main results of this thesis exposed critical trade-offs, from both a recovery rate and sustainability perspective, through the recovery strategies used. The scenario in which the recovery rate was increased, the results showed that the recovery rate was high, meaning more metals were recovered compared to lost. Likewise, looking at the sustainability impact, specifically the economic perspective, the economic losses were minimized largely. But it is important to also be aware that, even though the recovery ratio is positive, the total amounts of metals and environmental holdings that were recovered were low. Moving on, the scenarios in which the collection rate was increased, the results showed that the metals recovered were way higher. Additionally, the environmental impacts, in terms of mining- and CO2 saved, had way higher amounts of reductions compared to the recovery-increase scenario. However, the collection-increase scenarios also led to a worse recovery rate efficiency and economic losses, with more amounts of metals being lost rather than recovered. The main contributions from this thesis is the creation of the decision-support artifact itself, which consisted of a combination of recovered and lost precious metals, scenario modeling, sustainability impact metrics, and finally, a dashboard visualizing all these critical information into one combined artifact allowing decision-makers to examine e-waste outcomes. This solution is neither predictive nor operational, but instead it justifies data-driven assessment regarding how collection- and recovery parameter increasements affects the recovery and sustainability impacts in the Norwegian e-waste system.

• Katta, Vikash; Ottestad, Ole Jakob & Asam, Bhatti, Toqir (2026). FRA STATISK TIL DYNAMISK RISIKOSTYRING: TILNÆRMINGER FOR KONTINUERLIG INTEGRASJON AV TRUSSEL- OG SÅRBARHETSINFORMASJON. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Denne masteroppgaven undersøker hvordan man kan etablere en mer dynamisk kobling mellom risikovurdering, etterlevelse av standarder, rammeverk, lover og forskrifter, og kontinuerlig innkommende trussels– og sårbarhetsinformasjon. Arbeidet med etterlevelse er basert på lovverk, standarder og rammeverk som ISO- 27001 og 27002, NIST, NSMs grunnprinsipper for IKT-sikkerhet, CIS-Controls, og sikkerhetsloven. Disse utgjør et statisk grunnlag for risikoregistre og etterlevelsesoppfølging som blir gjennomført av de fleste organisasjonene i dag. Ny trussel- og sårbarhetsinformasjon er derimot dynamisk og endrer seg kontinuerlig. Hovedutfordringen som adresseres i oppgaven er hvordan man kan automatisere samspillet mellom disse. Hvordan etterlevelse informasjonen kan aktivt støtte vurdering av nye trusler og sårbarheter, og hvordan ny sårbarhet- og trusselinformasjon kan tilbakeføres for å oppdatere risikoregistrene. Gjennom en kombinasjon av litteraturstudie, analysering av eksisterende rammeverk, løsninger og verktøy og utvikling av en Proof Of Concept modell i samarbeid med Norsk Nukleær Dekommisjonering, vil oppgaven identifisere både muligheter og utfordringer ved å bygge en dynamisk prosess innen risikostyring med strategisk og operasjonelt sikkerhetsarbeid til betraktning. Resultatene gir praktiske anbefalinger for hvordan organisasjoner kan innpasse dynamisk risikoinformasjon for bedre beslutningsgrunnlag og risikokommunikasjon til relevante aktører som kan bidra til å ta beslutninger tidlig.

• Nichele, Stefano; Guichard, Etienne & Daniel, Eggereide, (2026). Exploring a Neural Cellular Automaton Approach to Solving Sudoku-Bench. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  This thesis investigates to what extent a Neural Cellular Automaton (NCA) can learn to solve puzzles from Sudoku-Bench. Sudoku-Bench is a benchmark designed to test reasoning over tightly coupled constraints, using Sudoku-like puzzles with different grid sizes and additional rule types. The aim of the project was to explore whether an NCA, which updates a grid through learned local rules, can learn useful Sudoku-solving behavior and generalize this behavior to Sudoku-Bench puzzles. The thesis follows an experimental development process. A preliminary NCA model was first developed for standard Sudoku, before the architecture was extended to support different grid sizes, rule-specific input channels, and a final training pipeline based on generated puzzle stacks. Before the final training run, the generated training data was compared with the Sudoku-Bench challenge set, and puzzles with identical input representations were removed to avoid direct overlap between the training data and the benchmark inputs. The final model was then trained using masked cross-entropy loss and evaluated on generated test data and the Sudoku-Bench challenge. The results show that the model learned meaningful Sudoku-solving behavior, but that its final performance on Sudoku-Bench was limited. The best single checkpoint and NCA-step combination solved 5 of the 100 challenge puzzles, all of which were $4 \times 4$ puzzles. Across all saved checkpoints and NCA steps, 6 puzzles were solved at least once. The model did not solve any $6 \times 6$ or $9 \times 9$ benchmark puzzles. Additional trace analysis showed that the model often tended to stabilize its predictions around one training rollout, even though the pool training was intended to encourage continued improvement across multiple rollouts. Overall, the project shows that solving Sudoku-Bench with an NCA may be feasible, but that the current implementation is only a first step toward that goal. The code developed during the project can be found at \url{https://github.com/Hubiele/Sudoku-Bench}.

• Katta, Vikash; Rabi, Maben & Jacob, Holth, (2026). A Modular Platform for Evaluating Digital Twins for Cybersecurity in Vehicular IoT Platooning. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Vehicular Internet-of-Things (IoT) platooning relies on continuous exchange of state information between vehicles. This creates a cyber-physical security problem: manipulated or inconsistent telemetry can influence physical leader-follower behaviour, while conventional channel protection does not by itself assess whether reported values are plausible. This thesis investigates how Digital Twins can support cybersecurity evaluation in controlled leader-follower platooning, and what practical limitations arise when such monitoring is implemented. The main contribution is TwinGuard, a modular experimental workbench for platooning cybersecurity experiments. TwinGuard converts live MQTT telemetry and replay datasets into canonical frames, runs Digital Twin components for contextual interpretation, and applies a ScenarioTwin layer for targeted consistency checks. Attack injection is implemented outside the Digital Twin pipeline through an MQTT tamper relay, so monitoring is based on observed telemetry effects rather than direct attack commands. The platform also supports visualization, recording, replay, evidence inspection, and risk-oriented clean trials produced no false positives. Velocity spoofing was detected most consistently, with 59 of 61 trials detected, while false-brake trials were detected in 36 of 60 trials. These results show that TwinGuard can support repeatable, evidence-based cybersecurity experimentation. The validation should therefore be interpreted as evidence that the workbench operates end-to-end under controlled conditions, not as a complete intrusion-detection benchmark.

• Nichel, Stefano & Vemund, Vestgøte, (2026). Applying Diffusion Models to Solve Scheduling Problems. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  This thesis explores the use of diffusion, a class of generative machine learning models commonly associated with image generation, to solve scheduling problems. Diffusion is compared with traditional optimization methods. Conventional approaches, such as exact algorithms (e.g., linear programming and constraint programming) and metaheuristics (e.g., genetic algorithms and simulated annealing), can produce high-quality solutions for scheduling tasks. However, in practice, factors beyond the quality of the solution may be equally important. Runtime efficiency, flexibility in implementing different objectives, such as ability to handle dynamic changes like machine wear are all important practical considerations. Traditional methods often struggle with these aspects, motivating research into novel methods that can address them more effectively. Diffusion may offer fast generation times and greater flexibility in solution generation, making it a promising alternative for scheduling applications. This thesis investigates how a diffusion model can be developed for scheduling and evaluates its ability to generate high-quality schedules compared to traditional methods. In addition, diffusion is assessed with respect to practical considerations, where runtime efficiency and flexibility may be as important as schedule quality. Several challenges associated with applying diffusion to scheduling problems are highlighted, including the generation of infeasible schedules, and different approaches for addressing these challenges are explored. The results indicate that diffusion is a promising approach for scheduling tasks, capable of generating quality schedules while maintaining low runtime and a high degree of flexibility

• Rabi, Vikash katta, Maben & Mo, Kirkeluten, Torstein (2026). A Modular Platform for Evaluating Digital Twins for Cybersecurity in Vehicular IoT Platooning. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Vehicular Internet-of-Things (IoT) platooning relies on continuous exchange of state information between vehicles. This creates a cyber-physical security problem: manipulated or inconsistent telemetry can influence physical leader-follower behaviour, while conventional channel protection does not by itself assess whether reported values are plausible. This thesis investigates how Digital Twins can support cybersecurity evaluation in controlled leader-follower platooning, and what practical limitations arise when such monitoring is implemented. The main contribution is TwinGuard, a modular experimental workbench for platooning cybersecurity experiments. TwinGuard converts live MQTT telemetry and replay datasets into canonical frames, runs Digital Twin components for contextual interpretation, and applies a ScenarioTwin layer for targeted consistency checks. Attack injection is implemented outside the Digital Twin pipeline through an MQTT tamper relay, so monitoring is based on observed telemetry effects rather than direct attack commands. The platform also supports visualization, recording, replay, evidence inspection, and risk-oriented clean trials produced no false positives. Velocity spoofing was detected most consistently, with 59 of 61 trials detected, while false-brake trials were detected in 36 of 60 trials. These results show that TwinGuard can support repeatable, evidence-based cybersecurity experimentation. The validation should therefore be interpreted as evidence that the workbench operates end-to-end under controlled conditions, not as a complete intrusion-detection benchmark.

• Blauhut, Daniela & Stavrum, Myrvold, Marianne (2026). SmartMaking: En digital og brukervennlig plattform som gir veiledning, opplæring og støtte til bærekraftig bruk ved digital fabrikasjon. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Prosjektet undersøker hvordan en digital læringsplattform kan bidra til tilgjengelighet, forståelse og bærekraftig bruk av MakerSpace og FabLab ved Høgskolen i Østfold avdeling Halden (HiØ). Prosjektet tar utgangspunkt i to forskningsspørsmål som omhandler utforming av en brukervennlig og pedagogisk plattform som kan engasjere og støtte studenters læring om bærekraftig prototyping og digital fabrikasjon. Prosjektet er gjennomført som en kvalitativ studie basert på litteraturgjennomgang, kartlegging av praksis ved flere studiesteder, intervjuer, observasjoner, brukertesting og iterativ prototyping. Funnene viser at studenter opplever utfordringer med å tilegne seg og huske informasjon fra opplæringen, samtidig som det finnes begrenset opplæringsmateriale tilgjengelig utenfor kursene. Bærekraft er i liten grad integrert i dagens praksis, og det mangler gode systemer for resirkulering og gjenbruk. Funnene indikerer også at en oversiktlig og strukturert plattform med innhold som tekst, bilder, videoer og quiz, kan bidra til økt engasjement, forståelse og bedre navigasjon. Universell utforming, gestaltprinsipper og pedagogiske prinsipper er sentrale for å sikre en brukervennlig plattform som støtter kunnskapsdeling, samt praktisk og selvstendig arbeid. Som en del av prosjektet ble det utviklet en digital læringsplattform, en resirkuleringsstasjon og et konseptuelt rammeverk for bærekraftig arbeid med digital fabrikasjon. Rammeverket integrerer bærekraft gjennom fabrikasjonsprosessen, inkludert resirkulering, gjenbruk og dokumentasjon. Prosjektet konkluderer med at en digital læringsplattform som formidler rammeverkets budskap, kombinert med en resirkuleringsstasjon, kan bidra til økt tilgjengelighet, forståelse og bærekraftig bruk av MakerSpace og FabLab ved HiØ Halden. Løsningen kan engasjere og støtte studenters læringsprosess, samt fremme en mer bærekraftig praksis innen digital fabrikasjon. Samtidig er det behov for videre forskning på bærekraft i digital fabrikasjon i en IT-sammenheng.

• Blauhut, Daniela & Lassen, Grimsrud, Jennifer (2026). SmartMaking: En digital og brukervennlig plattform som gir veiledning, opplæring og støtte til bærekraftig bruk ved digital fabrikasjon. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Prosjektet undersøker hvordan en digital læringsplattform kan bidra til tilgjengelighet, forståelse og bærekraftig bruk av MakerSpace og FabLab ved Høgskolen i Østfold avdeling Halden (HiØ). Prosjektet tar utgangspunkt i to forskningsspørsmål som omhandler utforming av en brukervennlig og pedagogisk plattform som kan engasjere og støtte studenters læring om bærekraftig prototyping og digital fabrikasjon. Prosjektet er gjennomført som en kvalitativ studie basert på litteraturgjennomgang, kartlegging av praksis ved flere studiesteder, intervjuer, observasjoner, brukertesting og iterativ prototyping. Funnene viser at studenter opplever utfordringer med å tilegne seg og huske informasjon fra opplæringen, samtidig som det finnes begrenset opplæringsmateriale tilgjengelig utenfor kursene. Bærekraft er i liten grad integrert i dagens praksis, og det mangler gode systemer for resirkulering og gjenbruk. Funnene indikerer også at en oversiktlig og strukturert plattform med innhold som tekst, bilder, videoer og quiz, kan bidra til økt engasjement, forståelse og bedre navigasjon. Universell utforming, gestaltprinsipper og pedagogiske prinsipper er sentrale for å sikre en brukervennlig plattform som støtter kunnskapsdeling, samt praktisk og selvstendig arbeid. Som en del av prosjektet ble det utviklet en digital læringsplattform, en resirkuleringsstasjon og et konseptuelt rammeverk for bærekraftig arbeid med digital fabrikasjon. Rammeverket integrerer bærekraft gjennom fabrikasjonsprosessen, inkludert resirkulering, gjenbruk og dokumentasjon. Prosjektet konkluderer med at en digital læringsplattform som formidler rammeverkets budskap, kombinert med en resirkuleringsstasjon, kan bidra til økt tilgjengelighet, forståelse og bærekraftig bruk av MakerSpace og FabLab ved HiØ Halden. Løsningen kan engasjere og støtte studenters læringsprosess, samt fremme en mer bærekraftig praksis innen digital fabrikasjon. Samtidig er det behov for videre forskning på bærekraft i digital fabrikasjon i en IT-sammenheng.

• Blauhut, Daniela & Dombestein, Heggelund, Jonas (2026). Utforsking og implementering av en interaktiv oversikt for Makerspace: Design og utvikling av et brukersentrert dashbord for utstyrsstatus. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  Denne oppgaven presenterer design og utvikling av et webbasert sanntidsdashbord for Makerspacet ved Høgskolen i Østfold. Bakgrunnen for prosjektet er at mange Makerspaces mangler digitale verktøy for oversikt over maskinstatus og brukeraktivitet, noe som skaper utfordringer for både studenter og ansatte. Prosjektet tar utgangspunkt i to forskningsspørsmål: hvordan et brukervennlig sanntidssystem kan designes og utvikles for å gi brukere oversikt over maskinstatus, romaktivitet og miljøinformasjon, og hvilke typer informasjon studenter og ansatte opplever som mest relevante i et Makerspace. Den metodiske tilnærmingen er forankret i forskning gjennom design, kombinert med brukersentrert design. Empiriske data ble samlet inn gjennom observasjoner ved HiØ og OsloMet, semistrukturerte intervjuer og brukertesting underveis i utviklingen, samt en avsluttende brukervennlighetstest med ti informanter. Dashbordet ble utviklet i rammeverket Nuxt og integrerer sanntidsdata fra 3D-printere via PrusaLink, loddestasjon via termisk kamera, laserkutter via akselerometer, storformatskriver via HP DesignJet API, samt miljødata fra Makerspacet. Funnene fra brukervennlighetstesten viser at samtlige informanter navigerte dashbordet uten veiledning, og at brukerne i størst grad verdsetter informasjon om maskinstatus og tilgjengelighet. Informasjonsbehovet varierer mellom brukergrupper, der ansatte og studentassistenter i større grad etterspør tekniske detaljer og statistikk. Prosjektets største bidrag er integrasjonen av sanntidsdata fra flere ulike maskiner i ett felles grensesnitt, noe som skiller seg fra eksisterende løsninger som typisk fokuserer på én maskintype om gangen.

• Nichele, Stefano & Shokoufeh, Ghadersouri, (2026). Exploring Improvements to Neural Cellular Automata for ARC-AGI-1 and ARC-AGI-2. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  This thesis explores the use of Neural Cellular Automata (NCA) for solving ARC-AGI tasks in a task-wise few-shot setting. ARC-AGI is designed as an IQ-test-like benchmark for AI systems, where the goal is to test reasoning, abstraction, and generalization from only a few input-output examples. The work begins by implementing a baseline NCA model and evaluating it on a sampled subset of ARC-AGI-1 tasks. The baseline results are then analyzed to identify where the model performs well and where it fails, with particular attention to local transformations, overfitting, unseen colors, rotations, and more complex reasoning patterns. Based on this analysis, several model variants are developed and compared. The first improvement direction focuses on architectural modifications to the NCA perception mechanism, including extended fixed-filter and learnable-filter variants. The second direction investigates training and architectural strategies aimed at improving generalization, including color permutation, orientation-aware perception, and combined variants. Selected models are then evaluated on the full ARC-AGI-1 and ARC-AGI-2 datasets using exact-match performance, task-level comparisons, and qualitative inspection of model predictions. The results show that NCA models can solve some ARC-AGI tasks, especially those based on local or iterative spatial transformations. Learnable and rotating perception filters improved performance compared with the baseline, while color permutation helped on some tasks involving unseen colors. However, the improvements were limited and task-dependent, and ARC-AGI-2 remained especially challenging. The thesis concludes that NCAs are a promising but still limited approach for ARC-style reasoning, and that stronger mechanisms for global coordination, compositional reasoning, and robust generalization are needed.

• Demi, Selina & Forberg, Reinaas, Kevin (2026). Building Cyber Threat Intelligence Teams and Programs in Critical Infrastructure Organizations: Lessons Learned from Industry Practices. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  The thesis investigates and examines how critical infrastructure organizations can build and mature Cyber Threat Intelligence (CTI) capabilities to enhance their cybersecurity. It also examines how CTI capabilities are integrated into organization processes within a large critical infrastructure organization in the energy sector. Using a qualitative single-case study, five semi-structured interviews with cybersecurity and risk management employees were conducted. In addition to the interviews, document analysis was performed to help strengthen the interpretation of the findings. The results were analyzed using the Cyber Threat Intelligence Capability Maturity Model (CTI-CMM) as a theoretical lens to interpret CTI capability development. The results show that CTI capability is an adaptive process driven by organizational needs, stakeholder needs and continuous learning. The findings also show that CTI supports security operations, cyber risk management, and strategic decision-making at the same time, highlighting the importance of cross-functional integration of CTI. Some important success factors for effective CTI operations were found to be communication, stakeholder trust and translating technical intelligence into business relevance. The study also shows how maturity models can support capability development while highlighting limitations connected to maturity assessment in operational environments. The thesis contributes to research by providing insight into CTI capability development within a critical infrastructure by identifying lessons learned in building mature CTI programs.

• Gurav, Aniket Anand; Chanda, Sukalpa; Pedersen, Marius & Krishnan, Narayanan Chatapuram (2026). Handwritten Text Recognition and Generation for Historical Document. Norges teknisk-naturvitenskapelige universitet. ISSN 9788235301468. Fulltekst i vitenarkiv Vis sammendrag

  Historical Handwritten Text Recognition (HTR) and handwriting generation are central to preserving cultural heritage, but low-resource languages and scripts face persistent challenges: limited annotated data, diacritics, multi-writer variability, and document degradation. Existing approaches, often tuned for high-resource languages, generalize poorly to unseen words, provide limited style control, and overfit when data is scarce. This thesis improves Zero-Shot Learning (ZSL) recognition and enables diverse, realistic handwriting generation . Six interconnected contributions form a unified framework. ResPho(SC)Net employs a ResNet-18 backbone with Pyramidal Histograms of Characters (PHOC) and Pyramidal Histograms of Shapes (PHOS) embeddings for robust unseen word recognition. HTLDIFF introduces diffusion-based line-level handwriting generation, while Word-Diffusion extends this to Out-of-Vocabulary (OOV) words using PHOS conditioning. Beyond Memorization offers training-free style diversification via writer embedding injection. ResNet-TPP enhances sub-word feature disentanglement through parallel temporal pyramid pooling (TPP), and CA-DIFF leverages multi-stage cross-attention for style consistency and improved writer classification. Overall, the recognition contributions improve robustness to unseen words and diverse handwriting styles, while the generation contributions increase stylistic diversity and realism for augmentation, leading to more reliable downstream HTR in low-resource settings. Together, these methods strengthen generalization across historical collections and provide practical benefits for archives, libraries, and cultural institutions.

• Jamal, Abdullah & Kishor, Nand (2026). Energy Flexibility as a Service (EFaaS) scenarios from Verdal public buildings, where the role of final consumers/prosumers is enabled actively in the energy market. Zenodo. doi: 10.5281/ZENODO.19132993. Fulltekst i vitenarkiv Vis sammendrag

  This scientific report presents the progress of Work Package 3 (WP3), “Design of energy flexibility solutions”, in the PERSIST project. It describes the development and first implementation of a common technical framework for Energy Flexibility as a Service (EFaaS) and applies this framework to Verdal municipality. In the project proposal, Deliverable D3.1 is defined as “EFaaS scenarios on Verdal municipality”. In this report, that objective is met through a detailed case study of Verdal’s municipal public-building portfolio, which acts as the primary test site. The work is based on hourly data from 1 January to 31 December 2024.

• Ahmad, Shahzad; Chanda, Sukalpa; Pedersen, Marius & Rawat, Yogesh S (2026). Remote Intelligent Health Monitoring. Norges teknisk-naturvitenskapelige universitet. ISSN 9788235300881. Fulltekst i vitenarkiv Vis sammendrag

  Remote health monitoring using non-contact visual sensing has emerged as a promising direction for scalable, accessible, and privacy-preserving healthcare technologies. However, reliable estimation of physiological signals from facial videos remains challenging due to motion artifacts, illumination variation, environmental noise, and limited supervision. In parallel, meaningful interpretation of physiological change in realistic settings also requires awareness of behavioral and contextual information. This thesis develops a coherent research framework for Remote Intelligent Health Monitoring (RIHM), organized around two complementary branches: non-contact physiological sensing and contextual video understanding. The first part of the thesis focuses on remote photoplethysmography (rPPG) and camera-based SpO2 estimation. It shows that temporal stability is a foundational requirement for reliable physiological representation learning from weak and indirectly observed video signals. To address this challenge, the thesis introduces stability-aware learning strategies and continuous-time neural modeling through Liquid Time-Constant (LTC) networks, enabling adaptive temporal representations that improve robustness to motion, noise, and signal inconsistency. For camera-based SpO2 estimation, the thesis further develops physiology-aware modeling of AC/DC signal components together with continuity-regularized and physics-aware learning objectives, improving temporal coherence and physiological plausibility. The second part of the thesis focuses on contextual video understanding through zero-shot and compositional action recognition using vision-language models. To improve temporal reasoning in video, lightweight temporal adaptation mechanisms are introduced that enhance motion sensitivity without requiring full retraining of large pretrained backbones. To improve compositional generalization and reduce shortcut bias, the thesis further develops explanation-conditioned and composition-aware learning strategies that strengthen robustness to unseen action-object combinations. Taken together, the thesis shows that reliable remote health intelligence from video benefits from structured modeling of latent dynamics. In the physiological branch, this structure appears through temporal stability, continuous-time dynamics, physiological priors, and continuity-aware constraints. In the contextual branch, it appears through temporally efficient adaptation and shortcut-aware compositional reasoning. Rather than presenting a fully integrated end-to-end monitoring system, the thesis establishes the methodological foundations and conceptual bridge needed for future RIHM systems that combine physiological sensing with contextual understanding in a more reliable, interpretable, and scalable manner.

• Katta, Vikash & Tommy, Stefanac, (2026). Implementing Autonomous drones in the Fire Department. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  According to data gathered by the Commission of the European Union (EU), within twenty years the European drone sector alone is expected to employ directly more than 100,000 people and to have an economic impact exceeding €10 billion per year. The integration of autonomous drones into emergency services has the potential to significantly enhance situational awareness, influence response strategies, and support decision-making under uncertainty. This master’s thesis investigates how autonomous drones may support fire departments during the early phase of emergency incidents, with particular emphasis on data needs, perceived effects on response time, and decision-making processes. The study is based on qualitative data collected through tabletop use-case scenarios, using a conceptual model of a fully autonomous drone, conducted with two Norwegian fire departments. Three representative incident types were examined: hazardous materials (CBRNE), wildfires, and drowning incidents. Each scenario was carried out both with and without autonomous drone support, followed by semi-structured group feedback sessions with participating firefighters. The data were analysed using thematic analysis to identify recurring patterns and perspectives across departments and use cases. The findings indicate that firefighters perceive autonomous drones as a valuable tool for enhancing situational awareness through early access to real-time visual data, thermal imaging, and environmental information. Participants consistently reported that such information could support earlier and more informed decision-making, particularly with regard to resource allocation and safety considerations. However, the study does not empirically measure actual performance outcomes such as verified response-time reductions. The results therefore primarily reflect perceived operational value rather than objectively measured effects. The findings should be interpreted in light of the study’s context, including the use of tabletop exercises and varying levels of prior drone experience among participants. While these factors may influence the consistently positive assessments, the study provides valuable insight into how end users perceive the role of autonomous drones in fire service operations. The thesis concludes by outlining directions for future research, including live field trials to evaluate measurable performance effects and further explore organizational and technological requirements for implementation.

• (hasan.ogul@hiof.no), Hasan Ogul & Muhammad, Waqas, (2025). Benchmarking Code Embeddings for Automatic Grading Across Multiple Programming Assignment Datasets. Høgskolen i Østfold. Fulltekst i vitenarkiv Vis sammendrag

  As class sizes continue to grow and students submit increasing amounts of code, automated assessment of programming assignments has become more important than ever. Manual grading is not only time-consuming but can also be inconsistent and difficult to scale. Recent progress in pretrained code models offers a promising alternative by allowing student programs to be represented as dense embeddings that capture both their structure and meaning. This thesis explores how well different embedding models can be used to predict student grades across two educational programming datasets. The study compares five pretrained models CodeBERT-base, CodeBERTa-small, CodeBERT-MLM, UniXcoder, and CodeT5-base using a unified benchmarking pipeline. Two datasets are first preprocessed: OSF, which contains docstring–code pairs, and CSEDM, which consists of code-only submissions. Embeddings are extracted using either mean pooling or layer aggregation and then fed into lightweight regression models. Grade prediction is performed using Ridge Regression as well as two XGBoost configurations, one regularized and one high-capacity. Model performance is evaluated using RMSE, MAE, R², Quadratic Weighted Kappa (QWK), and the Concordance Index (C-index). The results show that normalized mean-pooled embeddings usually work better than deeper layer aggregation methods. CodeBERTa-small and CodeT5-base both perform well on the two datasets. Ridge Regression often does better than XGBoost for code-only submissions. Each dataset’s features affect how hard it is to make predictions. For example, grade imbalance in OSF and more score variation in CSEDM both influence model performance. The error analysis also shows that it's hardest to guess mid-range grades and very short code submissions correctly. In conclusion, the results highlight both the strengths and limitations of methodologies based on word or code patterns for forecasting student performance. This study establishes a systematic approach for evaluating code models previously trained for educational applications. Additionally, it offers recommendations for future research regarding the provision of automated feedback, the development of data collections, and the assessment of progress measurement in programming education.

• Kivedal, Bjørnar Karlsen (2025). Climate Risks and Housing Markets. OsloMet - storbyuniversitetet. Fulltekst i vitenarkiv
• Petrichenko, Lubova; Jamal, Abdullah; Mutule, Anna; Kishor, Nand; Lazdins, Roberts & Kozadajevs, Jevgenijs (2025). Evaluation of existing schemes supporting directly or indirectly investments at consumer level into assets that contribute to long-term changes in electricity production or consumption. Zenodo. doi: 10.5281/ZENODO.15771818. Fulltekst i vitenarkiv Vis sammendrag

  The transition to renewable energy and the increasing decentralization of energy systems have underscored the need to evaluate regulatory frameworks that influence consumer investments in long-term electricity production and consumption assets. Understanding whether these frameworks promote or hinder such investments is critical to achieving national and global sustainability goals.

• Øystein, Haugen, & Preeti, Grewal, (2025). AI-Powered Energy Hunter. Høgskolen i Østfold. Fulltekst i vitenarkiv

Se alle arbeider i NVA