Blekinge Institute of Technology
Not a member yet
13576 research outputs found
Sort by
A hybrid CNN-transformer framework optimized by Grey Wolf Algorithm for accurate sign language recognition
No full textThis paper introduces the Gray Wolf Optimized Convolutional Transformer Network, a combined deep learning framework aimed at accurately and efficiently recognizing dynamic hand gestures, especially in American Sign Language (ASL). The model integrates Convolutional Neural Networks (CNNs) for spatial feature extraction, Transformers for temporal sequence modeling, and Grey Wolf Optimization (GWO) for hyperparameter tuning. Extensive experiments were conducted on two benchmark datasets, ASL Alphabet and ASL MNIST to validate the model’s effectiveness in both static and dynamic sign classification. The proposed model achieved superior performance across all key metrics, including a accuracy of 99.40%, F1-score of 99.31%, Matthews Correlation Coefficient (MCC) of 0.988, and Area Under the Curve (AUC) of 0.992, surpassing existing models such as PCA-IGWO, KPCA-IGWO, GWO-CNN, and AEGWO-NET. Real-time gesture detection outputs further demonstrated the model’s robustness in varied environmental conditions and its applicability in assistive communication technologies. Additionally, the integration of GWO not only accelerated convergence but also enhanced generalization by optimally selecting model configurations. The results show that GWO-CTransNet offers a powerful, scalable solution for vision-based sign language recognition systems, combining high accuracy, fast inference, and adaptability in real-world applications.
Public policy implications of entrepreneurship research
No full textAfter examining 4247 articles in 10 journals between 2010 and 2020, we found only 11.7% (n = 497) included explicit policy recommendations, and just 1.4% (n = 60) had explicit policy implications subsections. We analysed these 60 articles using behaviourally anchored rating scales (BARS) to assess four key aspects of the policy implications for quality and practicality. Policy implications were generally based on methodologically transparent research, but often lacked rigour in detailing policy and regulatory requirements, and typically demanded significant resources for implementation. Our results contribute to theory by improving our understanding of the why, when and how of policy implications of entrepreneurship research. Also, our BARS framework allows for a differentiation between quality and practical applicability, thereby reframing assumptions by showing that impactful policy implications require both methodological quality and implementation feasibility. Overall, the research-policy gap in entrepreneurship results not merely from disengagement but also research-policy misalignment in content and form
Behavioural Change After Cybersecurity Awareness Training : An Empirical Study Using Evilginx Phishing Simulations
No full textBackground. Cyberattacks increasingly use social engineering tactics like phishing to exploit human vulnerabilities rather than technical system flaws. These dangers are especially dangerous because they may be avoided by using advanced tools like Evilginx, which acts as a reverse proxy. Therefore, it's very important to do realistic phishing simulations to test and improve how aware employees are of security issues. Goals. This study empirically examines the influence of cybersecurity awareness training on employee behavior by assessing their responses to Evilginx-powered phishing simulations prior to and subsequent to a focused training intervention. Methods. We used a quasi-experimental methodology to give 50 employees two phishing simulations, one before and one after training. We gathered behavioral information such as click- through rates, credential submissions, reporting rates, and response times. Results. The results show that behavior changed in a big way after instruction. The number of people who clicked through dropped by 57%, the number of people who submitted credentials declined by 72%, and the number of people who reported jumped by 340%. People also replied to emails that seemed suspicious 65% faster. Final Thoughts. Realistic simulations in phishing awareness training can make employees far more resistant to advanced attacks. Tools like Evilginx are useful for finding behavioral weaknesses and testing the efficiency of training programs in real life
”Jag planerar mitt liv efter mensen” – Kvinnors upplevelser av vardagliga begränsningar vid dysmenorré : En allmän litteraturstudie
No full textBakgrund: Smärtsamma menstruationer, även kallat primär dysmenorré är det vanligastegynekologiska tillståndet bland kvinnor och drabbar mellan 45-95% av menstruerandekvinnor, där kraftig dysmenorré bedöms drabba ~10-25%. Kvinnor som lider av dysmenorréhar ofta sänkt livskvalitet på grund av vardagliga begränsningar, där sänkningen står iförhållande till symtomens svårighetsgrad. Syfte: Syftet är att belysa kvinnors upplevda begränsningar i vardagen vid primärdysmenorré. Metod: Studien är en litteraturöversikt av både kvalitativ och kvantitativ forskning där sexkvalitativa och sex kvantitativa artiklar från CINAHL och PubMed användes. Studien haranvänt Fribergs analysmodell. Resultat: Resultatet presenteras i fyra huvudkategorier med sammanlagt sex underkategorier.Huvudkategorierna är följande: Smärtans konsekvenser i det dagliga livet, Begränsningar idet sociala livet, Fysiska begränsningar samt Livskvalitet och psykiska besvär. Slutsats: Primär dysmenorré visade sig ha en stor begränsande effekt på drabbade kvinnorsvardag och livskvalitet
Operative System for Strategic Sustainable Development—Coordinating Analysis, Planning, Action, and Use of Supports Such as the Sustainable Development Goals, Planetary Boundaries, Circular Economy, and Science-Based Targets
No full textThe number of sustainability supports keeps growing. However, they are typically not capable of bridging planetary-to-organizational scales and therefore in themselves provide limited guidance for how organizations can contribute strategically to global sustainability. This is severe since the power of change that organizations represent is essential for humanity to succeed with the transition to sustainability. Furthermore, a specific sustainability support typically lacks a clear description of what gap it is designed to fill. This may add confusion to an already complex area and can result in organizations selecting supports that are not the most appropriate for the organization's purpose and transition plan. The aim of this study was therefore to prototype an approach for SWOT analyses of sustainability supports and for guiding effective and coordinated use of them. We tested the approach for some examples of supports that are currently receiving much attention: the sustainable development goals, planetary boundaries, circular economy, and science-based targets. We argue that the tested approach can be generalized and makes it possible to clarify limitations and value of any support from a systemic, systematic, and strategic sustainability perspective and, therefore, also to clarify how the supports relate to each other and can be selected and used in an effective and coordinated way. This is similar to how an operative system of, say, a smartphone, facilitates an effective and coordinated use of apps. We argue that the presented approach addresses one of the most dangerously, still most commonly, missed pieces in public and private discourses on sustainability.
Dual Effects of Hybrid Working on Performance : More Work Hours or More Work Time
No full textWork in software development companies has become increasingly hybrid with employees altering days of working in the office with days of working remotely from home. Yet, little is know about the efficiency of such way of working because the current scale of remote working is unprecedented. In this paper, we present our findings from a company-wide survey at Storebrand - a large-scale Norwegian fintech company, focusing on perceived performance. Our analysis of 192 responses shows that most employees report being able to perform the planned tasks. Further, half of respondents perceive to have increased work hours. Through qualitative analysis of open-ended commentaries of respondents we learned that remote working has dual effects on the perceived work hours - some employees report working longer hours and others report having more work time due to efficient use of the time throughout the day. Finally, we recommend managers to discuss and address the concerning habits of employees caused by increased connectivity and inability to stop working, before these lead to burnout and disturbances in the work/life balance.
A context in mutation : How the gig economy is changing the rules of the game for entrepreneurship?
No full textConsidering rapid digital transformation and recent changes in the macro-level “rules of the game” in entrepreneurship, we aim to problematize and understand the progressive encounter and new relationship between opportunity-driven and necessity-driven entrepreneurs within the context of the gig economy. In such a setting, both traditional entrepreneurial roles and the very division between them are brought into question. Their encounter implies a deviation from the basic assumptions of what entrepreneurship entails: being the bearer of risk, taking on uncertainty and individual responsibility, and pursuing unlimited profit potential. This deviation results in a change in the rules of the game. Consequently, we learn more about how the context of moving from a pre-gig economy toward the gig economy has implications for our societal understanding of entrepreneurship as a phenomenon.
A Systematic Literature Review of Hybrid Adaptive Scheduling Algorithmsfor Dynamic Fog ComputingEnvironments
No full textThis thesis presents a systematic review of the literature on scheduling algorithmsin fog computing, focusing on Hybrid Adaptive Scheduling Algorithms. The challengesof scalability, energy efficiency, flexibility, and latency have been addressedin this study by evaluating the efficacy of HASA against classical approaches likeRule-Based and AI-Driven Algorithms within resource-constrained dynamic environmentssuch as healthcare monitoring, smart cities, and industrial applications.Through a comparative analysis, HASAs demonstrate superior performance by balancingadaptability, energy optimization, and task distribution. Their lightweightAI models, together with a decentralized architecture and predictive mechanisms,can enable ultra-low latency with efficient resource management, hence becominghighly suitable for delay-sensitive applications. This work also investigates a fewother strategies, such as hierarchical scheduling architectures and adaptive learningmechanisms, which would enhance the robustness and applicability of HASAs to awide range of fog computing scenarios. The results provide a foundational frameworkfor improving scheduling methodologies in fog computing and point toward futureenhancements for scalable and efficient fog computing systems
How to start strong : Different approaches for crafting project briefs with corporate sponsors for global innovation and STEM-based team projects
No full textThe project brief is considered a pivotal component in corporate-sponsored student projects, yet many university teaching teams lack guidelines and best practices to define, frame, evaluate, and change these briefs with sponsors. We examine a rich data set of 68 project briefs used by 19 partner universities across a decade (2012–2022), drawn from a long-running project course taught at Stanford University and a related academic spinoff consortium called the SUGAR Network. All projects share a similar pedagogy for global innovation challenges and STEM-based team projects. Our study found that corporate sponsors sought seven different types of project outcomes. Although nearly two-thirds (65%) of sponsors adopted “how might we” phrasing in their briefs, other wording like “we dream…” was also used to provoke more imaginative thinking. Moreover, slightly over a third (35%) of briefs focused more on mid-term innovation horizons than near or far-term horizons. Based on these findings, we present a two-question guide for crafting a project brief with corporate sponsors to help student projects start from a stronger position
Digital Sovereignty for Collaborative AI Engineering
No full textIn Information Systems, implementing digital sovereignty is essential for improving transparency and establishing trust among stakeholders. This need for digital sovereignty is more prevalent in crowdsourcing platforms, where the stakeholders are often unknown to each other. AI marketplaces belong to the category of crowdsourcing information systems, where individuals and organizations collaborate to share various AI artifacts with one another. These marketplaces act as platforms that enable artifact exchange, thus accelerating the AI application development process through a multi-stakeholder approach to collaborative AI engineering. This work, investigates techniques for implementing digital sovereignty to promote collaboration among the stakeholders. Digital sovereignty thrives by empowering true owners with control and the ability to make independent decisions over their digital footprint. Depending on the application context, the type of control and the decision-makers change accordingly. For governments, digital sovereignty means the ability to manage citizens’ personal data and ensure data residency within a political region. For individual technology users, digital sovereignty refers to the ability to manage and control the interoperability of personal data across similar platforms. Nevertheless, digital sovereignty focuses on transferring control to the true owner by eliminating intermediaries or centralized organizations. The scope of this work lies in achieving digital sovereignty for marketplace platforms that operate in the context of exchanging data and other AI software artifacts. The Horizon 2020 projects, BonsApps and dAIEdge, provide a functional crowdsourcing AI marketplace with beta stakeholders, which also serves as a source for gathering requirements and validating concepts. The main contributions of this work are translating digital sovereignty definitions and requirements into the context of collaborative AI, as well as designing and implementing technical solutions to empower stakeholders of the underlying information system with digital sovereignty over their digital assets