The generalizability of Machine Learning (ML) classifiers in threat detection and classification (TD-TC) is crucial, particularly in Software-Defined Network (SDN)-based Internet of Things (IoT) architectures, where dynamic network management and centralized control introduce unique security challenges. However, there is an ongoing debate regarding the role of Network Identifier Attributes (NIAs) in training and evaluating ML models using publicly available datasets comprising network behaviors. Therefore, this study investigates the influence of NIAs on model generalization by examining learning curves with cross-validation. The results show that Timestamp and Flow ID lead to memorization in classifier models when using the IoTID20 dataset. However, removing all NIAs still resulted in a relatively high classifier accuracy, with Decision Tree (DT) achieving 96.63
The expansion of mobile cloud computing (MCC) has led to a rapid evolution in how data and services are delivered across geographically dispersed infrastructures. While MCC enables flexibility, scalability, and ubiquitous access, it presents critical security and performance challenges. This paper proposes a comprehensive architecture that leverages Software-Defined Networking (SDN), Network Function Virtualization (NFV), and Cloud Access Security Broker (CASB) to enhance the resilience, scalability, and manageability of MCC systems. We introduce a layered architecture where SDN controls the network plane, NFV deploys dynamic security functions, and CASB enforces data governance in cloud services. Through extensive experimentation using Mininet and traffic simulation tools, we evaluate the system’s latency, throughput, scalability, and security performance under various traffic conditions. Results demonstrate that the proposed solution achieves latency overhead under 5ms, successfully mitigates more than 90
Complex interface structures hinder the manual classification of UI components. This paper introduces a prototypical approach for UI element classification that combines a custom browser extension for collecting structured metadata (e.g., DOM structure, CSS properties) with machine learning methods to assign UI components to subject-specific tasks. The approach enables reliable clustering without extensive manual annotation and supports the evolution of adaptive user interfaces. Evaluations in the domains of e-commerce and smart city applications demonstrate that a combination of structural and textual features provides sufficient discriminatory power. The method integrates with intelligent interaction systems and context-aware interfaces, reducing redundancy and addressing growing interface complexity. Similar to semantic segmentation techniques, it enables dynamic adaptation of interaction flows, contributing to improved user experience and functional interface design.
This systematic literature review investigates the influence of technological knowledge on user acceptance, intention, and utilization of Artificial Intelligence (AI), synthesizing findings from 25 empirical studies published between 2021 and early 2025. Employing an integrated Antecedents-Decisions-Outcomes (ADO) and Theories-Contexts-Methods (TCM) framework, the analysis reveals that research in this period predominantly utilizes quantitative SEM methods based on TAM and UTAUT frameworks, largely within higher education contexts. Key findings indicate that technological knowledge facets, particularly Objective Knowledge (e.g., AI literacy) and Self-Efficacy, are critical antecedents to AI adoption, though their influence is often indirect. These factors frequently shape adoption outcomes by impacting core perceptual mediators, notably perceived ease of use, perceived usefulness, and trust. The synthesis highlights significant gaps, including limited methodological diversity, a narrow contextual focus, under-explored theoretical perspectives beyond TAM/UTAUT, and insufficient investigation into actual AI usage behavior. Addressing these gaps through diversified research approaches is crucial for advancing a comprehensive understanding of AI adoption dynamics.
The integration of the Internet of Things (IoT) and Cyber-Physical Systems (CPS) has revolutionized industries by enhancing connectivity, automation, and efficiency. However, these advancements bring significant privacy challenges due to extensive data collection and interconnected ecosystems. This paper provides a comprehensive analysis of privacy challenges in IoT and CPS, categorized into technical, data management, user-centric, and ecosystem-wide dimensions. Key privacy threats, such as data breaches, inference attacks, device tracking, and insider risks, are explored through case studies and empirical evidence. Existing mitigation strategies, including encryption, privacy-enhancing technologies, access control mechanisms, and regulatory frameworks, are evaluated. Despite these efforts, challenges remain, such as scalability of privacy solutions, privacy concerns in emerging technologies like 5G and AI-driven IoT, and the balance between privacy, security, and usability. The paper emphasizes the need for interdisciplinary collaboration to address these challenges, fostering trust and ensuring the sustainable development of IoT and CPS technologies. Future research directions are proposed to develop scalable, user-centric, and adaptive privacy-preserving frameworks that align with evolving technological and regulatory landscapes.
Nowadays, the prevalence of type 2 diabetes in humans is steadily increasing due to dietary habits and sedentary lifestyles. As a result, high-carbohydrate diets have caused imbalanced glucose levels, afflicting individuals with diabetes since medieval times. To address this problem, our paper introduces the innovative concept of MEGMNT-ITL (Metaverse-Edge Glucose Monitoring with Nutrition and Activities in Interoperable Transfer Learning). We present a novel and pioneering approach that combines metaverse technology with real-time glucose monitoring via smartwatches. Our objective is to reduce the chances of developing type 2 diabetes based on a healthy diet and the effects of the subject in practice. Due to resource constraints in smartwatches, we offload workloads to remote edge cloud servers to alleviate the computational issues of local systems by integrating transfer learning methodologies. The adaptive offloading scheme is suggested to allow for more efficient and precise glucose monitoring, further enhanced by considering variables such as nutrition and physical activity. Transfer learning ensures monitoring algorithms adapt to individual user behaviours and environmental conditions, enhancing detection accuracy. Overall, this methodology signifies a significant breakthrough in healthcare technology, with potential benefits for individuals managing diabetes and related conditions.
This paper explores the application of blockchain technology, smart contracts, encrypted Non-Fungible Tokens (NFTs), and the InterPlanetary File System (IPFS) in managing health records. The digitalization of health records, while beneficial, introduces significant challenges in terms of data security, privacy, and management. Our proposed framework leverages the decentralized, immutable, and transparent nature of blockchain technology to address these issues. By integrating RSA, RC4, DES, ChaCha20, Blowfish, and AES encryption algorithms with NFTs, the framework ensures that health records are secure, unique, and accessible only by authorized parties. Smart contracts automate various operations, enhancing system efficiency and reliability. We evaluated the framework’s performance on Binance Smart Chain, Polygon, Fantom, and Celo, focusing on system adaptability, operational speed, and cost efficiency. Our findings suggest that this integrated approach could significantly enhance the management and security of health records, presenting a viable solution for the healthcare sector’s digital management needs.
This study examines sustainable mobility intention in a university population, based on an online survey of 2,192 respondents (1,499 students and 693 staff). Drawing on the Rubicon Model of Action Phases and the Action-Phase Model of Developmental Regulation, we applied Partial Least Squares Structural Equation Modeling (PLS-SEM) to examine how sustainable mobility intention is associated with the cognitive salience of perceived barriers and transport modality during the pre-actional phase. Results show that stronger sustainable commuting intentions are associated with greater cognitive salience of specific barriers, particularly those related to commute time, cost, distance, and infrastructure limitations—indicating intensified appraisal during the pre-actional phase. The association between intention and transport modality was limited to students. These findings suggest that interventions targeting the pre-actional phase should focus on awareness, informational support, and reducing perceived constraints, while acknowledging that different barriers may trigger primary or secondary control strategies in goal pursuit.
The increasing use of electronic devices has raised expectations in how we communicate and share data, particularly in critical scenarios such as natural disasters or human negligence that severely affect existing infrastructure. Unmanned aerial vehicles (UAVs) used as Drone Base Stations (DBS), offer an innovative solution by providing wireless connectivity over cellular networks and distributing critical items to users in areas with limited coverage. This paper proposes three novel models based on the classic p-median problem to solve the optimal candidate site drone localization problem in time. More precisely, in our models, we introduce the time dimension of how drones move during time slots, which determines their performance. Additionally, we incorporate a mechanism that allows drones to move in a more free or restricted mode following their predefined label numbers. Finally, in the objective functions, we consider the cost of moving drones plus the user connectivity costs versus the cost of connecting users only. In our experimental setup, we compare all our models and determine differences in terms of CPU times, obtained objective function values, drone strategy, and associated costs.
In recent years, the rapid development of 5G networks has reinforced the importance of mobile edge computing in ensuring low latency and high bandwidth for content delivery in traffic-intensive applications. In particular, a large part of traffic load on the Internet is due to the repeated access to popular content, placing significant loads on cloud servers. Mobile edge caching mitigates this issue by storing frequently accessed data at edge devices, such as base stations and specialized edge nodes. Thus, this paper proposes a cooperative distributed caching architecture with tailored small base stations (SBSs) as edge nodes optimized through a game-theoretic model that achieves Nash equilibrium. The experiments exhibit that our proposal can reduce the latency of redundant data transmission, mitigate data traffic on the backbone, and alleviate the workload on cloud servers.
This paper presents a blockchain-based system designed to improve the management of intellectual property (IP) rights through the integration of smart contracts, RSA-encrypted Non-Fungible Tokens (NFTs), and the InterPlanetary File System (IPFS). Our approach leverages the unique capabilities of blockchain technology to enhance transparency and security throughout the IP lifecycle, from creation to commercialization. We evaluate various encryption methods to secure NFT metadata, ensuring robust protection for IP data. The system’s performance, including operational efficiency and security, was assessed across four EVM-compatible blockchain platforms: Binance Smart Chain, Polygon, Fantom, and Celo. The study offers a comprehensive analysis of each platform’s strengths and weaknesses in supporting our IP management system, focusing on transaction costs, speed, and resource efficiency. The findings indicate that blockchain technology can significantly aid in the secure, efficient, and transparent management of property rights.
The adoption of Software-Defined Networking (SDN) faces challenges due to existing infrastructure investments and migration costs, leading to the emergence of incremental hybrid SDN deployment. A critical challenge in hybrid SDN implementation is determining the effective node migration sequence while simultaneously considering immediate performance gains and effective load balancing. This paper investigates the impact of different reward systems in a Deep Reinforcement Learning (DRL)-driven SDN migration framework under dynamic network conditions. We propose three distinct reward functions focusing on local goal optimization, global goal optimization, and a combined approach integrating both goals. Experiments conducted on Abilene and GEANT network topologies demonstrate that the combined approach achieves superior results, requiring only 4 and 12 nodes respectively to attain minimum median Maximum Link Utilization (MLU), compared to local goal-oriented approach requiring 5 and 14 nodes. The combined approach maintains efficient convergence times (143 episodes for Abilene and 233 episodes for GEANT), comparable to single-goal reward functions. Our research contributes by: firstly, proposing three distinct reward functions for SDN migration, followed by demonstrating the combined approach’s superiority in achieving both objectives, and finally revealing that local and global goals can effectively complement each other within a well-designed reward system.
During the last decade, a paradigm shift can be observed in digital marketing as the influencer marketing landscape evolves. The Concept of virtual influencers (VIs) sheds light as a notion within this shift. These Vis challenge foundational assumptions of Source Credibility Theory by lacking lived experience and authentic human qualities; thus, it is evident that Vis must be effectively engaged with customers (followers). This systematic literature review synthesizes findings from influencer marketing, parasocial theory, consumer behaviour, and creator marketing to demarcate its role in facilitating consumer decision journeys. The results of the synthesized literature indicate that parasocial interactions better explain the audience’s connection with VIs than the Source Credibility Theory. It is postulated that human influencers are perceived as trustworthy due to their authenticity, whereas VIs sustain influence through content curation. The study further explores the potential for VIs to transition into creator-marketers, leveraging content-driven engagement and direct monetization over the influencer marketer role and outperforming at the top of the customer decision journey. Alongside audience engagement and endorsement, dynamics emerge as key to the longevity of the influence. Academic investigations into VIs as independent content entities are limited. This review highlights this deficiency and suggests a research framework for examining the potential of VIs to transition from their influencer role to content creators, facilitating the customer decision journey.
The waste collection process in Kota Kinabalu, Sabah, Malaysia faces significant communication challenges due to reliance on traditional communication methods and lack of an accessible waste platform, hindering effective coordination among residents, garbage truck drivers, and authorities. Without transparency in collection schedules, missed garbage bin pickups are frequent, leading to overflowing bins and frustration among residents who are unable to track or anticipate the accurate timing of truck arrivals. To address this issue, this project aims to develop a waste collection tracking system called KK Garbage Tracker. Through stakeholder survey and analysis of existing processes, system requirements were identified, leading to the development of a web application for authorities as admins and mobile applications for residents and drivers. The system was implemented using defined modules and underwent testing, including unit testing, integration testing, and system testing, to ensure its functionality and accuracy. Employing a combination of waterfall and iterative prototype methodologies, the usability of the system was assessed through the System Usability Scale. The expected outcome is a user-friendly tracking system that enhances communication and coordination in waste collection operations in Kota Kinabalu, aligning with Malaysia’s digitalization vision.
In modern smart cities, public transportation is essential for urban mobility, utilizing uniform ticketing systems across various routes. However, current transport infrastructure faces critical challenges, including cyberattack vulnerabilities, long processing times, high energy consumption, and operational costs. This paper proposes a blockchain-enabled digital twin framework for secure authentication and enhanced usability in intelligent transport systems within Oslo Smart City. The Digital Twin Blockchain Secure Access Control (DTBC-SAC) framework integrates advanced techniques for data collection, secure authentication, blockchain-enabled security, and digital twin simulations. It aims to enhance public transport infrastructure by reducing processing time, energy consumption, and operational costs while ensuring robust cyber protection for passengers. Additionally, the framework optimizes edge cloud servers that support transport operations, surpassing existing infrastructure in efficiency and security. Simulation results show that DTBC-SAC effectively minimizes processing time, energy usage, and operational costs while mitigating cyber threats. These findings highlight its superior performance over current methods, positioning DTBC-SAC as a transformative solution for intelligent transport systems in smart cities.
This paper addresses the management of software-defined networks (SDN) by minimizing latency for switches and controllers. We consider scenarios with a fixed number of controllers as well as scenarios where the model determines the optimal number of controllers. Our goal is to maximize link capacity utilization while imposing an upper bound to prevent overload and maintain balanced infrastructure. More precisely, we propose a Bilevel Programming approach for the problem where the goal is to optimize latency for the leader and maximize capacity for the lower-level user programming problem. To tackle the proposed model, we derive two alternative models while using linearization techniques to make the problem more computationally tractable using the Gurobi solver. To our knowledge, this is the first time that a bilevel approach is considered for latency management and capacity in SDN, offering an innovative and efficient solution. Based on numerical experiments, we observe that the model with variable controller allocation outperforms the fixed-controller model in terms of CPU time and solution quality for most of the studied Benchmark instances. Our proposed models achieve optimal solutions for most of the tested instances, highlighting the importance of proper mathematical formulations and demonstrating how efficient optimization can enhance SDN performance. Finally, enabling applications in larger and more complex wireless network scenarios.
The integration of Artificial Intelligence (AI) in the banking sector has the potential to significantly enhance decision-making processes, particularly in loan prediction and credit risk assessment. Despite these advancements, traditional loan assessment methods often lead to inaccuracies and biases, increasing the need for more sophisticated AI and machine learning (ML) techniques in the field. This paper reviews the current state of AI applications in the banking industry, focusing on their effectiveness in improving loan predictions while addressing challenges such as data quality, interpretability, and ethical implications. A comprehensive analysis of existing literature reveals the need for transparent and explainable AI models to facilitate stakeholder and customer trust while mitigating biases in financial decision-making. Additionally, the paper identifies critical challenges in AI integration, including the risk of reinforcing societal biases and the potential for fraudulent activities. To address these issues, we propose an AI framework called “EquiLoan” that emphasises the development of an integrated AI solution capable of real-time monitoring, explainability, and bias mitigation. This research aims to contribute valuable insights into the responsible use of AI in the banking industry, promoting fair lending practices and enhancing customer trust, while ensuring ethical considerations are prioritised in financial decision-making.
Leveraging the emerging promise of low-code development platforms (LCDPs), this study investigates their role in enhancing IT-business alignment within Norwegian enterprises, with a specific focus on the social dimension. Employing an exploratory, inductive case study methodology, in-depth interviews reveal that LCDPs primarily foster a collaborative arena for knowledge exchange between IT and business units rather than directly bridge the technical gap. Findings indicate that the use of LCDPs significantly improves social alignment by enhancing communication, shared understanding, and governance practices. In response, the study introduces a novel conceptual framework to assess social alignment maturity, emphasizing the critical importance of targeted training and robust governance mechanisms. Furthermore, the research identifies key organizational enablers and inhibitors ranging from internal readiness to structural barriers that shape the effective deployment of low-code solutions. This study contributes to the limited academic literature on low-code technologies and social alignment by synthesizing insights from empirical data with established information systems theory. It offers both theoretical and practical implications, providing a foundation for further research and guiding practitioners in leveraging LCDPs to facilitate IT-business collaboration, drive strategic agility, and support digital transformation initiatives.
Distal radius fractures (DRFs) are among the most frequently encountered fractures in clinical practice, requiring accurate and timely diagnosis for effective treatment. Thus far, deep learning-based object detection models have shown promise in automating fracture detection, however the potential of newer architectures, such as YOLOv8 and YOLOv11, remains largely unexplored in medical imaging. This study systematically evaluates YOLOv8m, YOLOv8l, YOLOv11m, and YOLOv11l, comparing their performance against the commonly utilized Faster R-CNN model in the literature for DRF detection. A diverse dataset of wrist radiographs, including publicly available and real-world clinical images, is used for training and evaluation. The models are assessed based on precision, recall, F1-score, mean average precision (mAP), and inference speed. The results indicate that YOLOv11l outperforms all other models, achieving the highest precision of 96.5
Small to Medium-sized Enterprises (SMEs) are vital to Europe’s economy but continue lagging behind larger firms in digital transformation (DT). This study identifies key SME-specific DT challenges: financial constraints, limited technical expertise, lack of strategic direction, and organizational resistance, through an AI-first literature review supported by human validation. In response, it introduces a conceptualization of the open-source SMEDT framework, which is do-it-yourself oriented and capability-driven. Designed with attention to contextual constraints such as knowledge dissemination gaps and high DT failure rates, the framework is a work-in-progress artifact for enabling self-guided transformation. The research contributes to the knowledge base on digital transformation in constrained contexts and advances the use of generative AI in early-stage design science. Future work includes advancing the SMEDT framework and its qualitative validation with SME stakeholders.
