Search Results (72)

Accurate malware family classification from dynamic sandbox reports continues to be a fundamental cybersecurity challenge. Most prior works depend on random splits that tend to overestimate accuracy, whereas deployment requires robustness under temporal drift as well as changing behaviors. We present a leakage-aware pipeline that transforms CAPEv2 sandbox JSON reports into structured visual heatmaps and evaluate models under stratified and chronological splits. The pipeline rigorously flattens behavioral keys, builds normalized representations, and benchmarks Random Forest, MLP, CNN64, HybridNet, and a modern ResNeXt-50 backbone. On the Avast–CTU CAPEv2 dataset containing ten malware families, Random Forest achieves nearly state-of-the-art accuracy (97.2% accuracy, 0.993 AUC) with high efficiency on CPUs, making it attractive for triage. ResNeXt-50 achieves the best overall performance (98.4% accuracy, 0.998 AUC) and provides visual interpretability via Grad-CAM, enabling analysts to verify predictions. We further quantify efficiency trade-offs (inference throughput and GPU memory) and report ablation studies on vocabulary size and keyset choices. These results affirm that though ensemble methods are still robust, heatmap-based CNNs provide better accuracy, interpretability, and robustness against drift. Full article

Artificial intelligence (AI) is playing an increasingly important role in driving sustainable change across coastal and marine environments. Artificial intelligence offers strong support for environmental decision-making by helping to process complex data, anticipate outcomes, and fine-tune day-to-day operations. In busy coastal zones such as the Mediterranean where tourism and boating place significant strain on marine ecosystems, AI can be an effective means for marinas to reduce their ecological impact without sacrificing economic viability. This research examines the contribution of artificial intelligence toward the development of environmental sustainability in marina management. It investigates how AI can potentially reconcile economic imperatives with ecological conservation, especially in high-traffic coastal areas. Through a focus on the impact of social and technological context, this study emphasizes the way in which local conditions constrain the design, deployment, and reach of AI systems. The marinas of Ibiza and Monaco are used as a comparative backdrop to depict these dynamics. In Monaco, efforts like the SEA Index^® and predictive maintenance for superyachts contributed to a 28% drop in CO₂ emissions between 2020 and 2025. In contrast, Ibiza focused on circular economy practices, reaching an 85% landfill diversion rate using solar power, AI-assisted waste systems, and targeted biodiversity conservation initiatives. This research organizes AI tools into three main categories: supervised learning, anomaly detection, and rule-based systems. Their effectiveness is assessed using statistical techniques, including t-test results contextualized with Cohen’s d to convey practical effect sizes. Regression R² values are interpreted in light of real-world policy relevance, such as thresholds for energy audits or emissions certification. In addition to measuring technical outcomes, this study considers the ethical concerns, the role of local communities, and comparisons to global best practices. The findings highlight how artificial intelligence can meaningfully contribute to environmental conservation while also supporting sustainable economic development in maritime contexts. However, the analysis also reveals ongoing difficulties, particularly in areas such as ethical oversight, regulatory coherence, and the practical replication of successful initiatives across diverse regions. In response, this study outlines several practical steps forward: promoting AI-as-a-Service models to lower adoption barriers, piloting regulatory sandboxes within the EU to test innovative solutions safely, improving access to open-source platforms, and working toward common standards for the stewardship of marine environmental data. Full article

Wheeled rovers are widely used in lunar and planetary exploration missions owing to their mechanical simplicity and energy efficiency. However, they face serious mobility challenges on sloped soft terrain, especially in terms of sideslip and loss of attitude angle when traversing across slopes. Previous research proposed using wheelbase extension/contraction and intentionally sinking wheels into the ground, thereby increasing shear resistance and reducing sideslip. Building upon this concept, this study proposes a novel recovery method that integrates beam extension/contraction and Archimedean screw-shaped wheels to enable lateral movement without rotating the rover body. The beam mechanism allows for independent wheel movement, maintaining stability by anchoring stationary wheels during recovery. Meanwhile, the helical structure of the screw wheels helps reduce lateral earth pressure by scraping soil away from the sides, improving lateral drivability. Driving experiments on a sloped sandbox test bed confirmed that the proposed 2DPPL (two-dimensional push-pull locomotion) method significantly reduces sideslip and prevents a drop in attitude angle during slope traversal. Full article

Augmented Reality (AR) is an innovative tool in education, enhancing learning experiences across multiple domains. This literature review explores the application of AR in education, with a particular focus on geographical learning. The study begins by tracing the historical development of AR, distinguishing it from Virtual Reality (VR) and highlighting its advantages in an educational context. The integration of AR into learning environments has been shown to improve engagement, comprehension of abstract concepts, and collaboration among students. The use of AR in geographical education through interactive applications, such as GeoAR and AR Sandbox, improves the exploration of spatial relationships, topographic maps, and environmental changes. Studies demonstrate that AR enhances students’ ability to recall information and understand geographical processes more effectively than with traditional methods. Furthermore, AR Sandbox implementations, including Illuminating Clay, SandScape, and AR Sandbox, are analyzed and compared. The paper also discusses future developments in AR for geography education for AR Sandbox, such as the integration of a mobile application for extended learning and improving computing solutions through Raspberry Pi. These advancements aim to make AR systems more accessible and to increase the benefits to both students and professors. Full article

The convergence of blockchain and metaverse technologies is poised to redefine how Global Value Chains (GVCs) create, capture, and distribute value, yet scholarly insight into their joint impact remains scattered. Addressing this gap, the present study aims to clarify where, how, and under what conditions blockchain-enabled transparency and metaverse-enabled immersion enhance GVC performance. A systematic literature review (SLR), conducted according to PRISMA 2020 guidelines, screened 300 articles from ABI Global, Business Source Premier, and Web of Science records, yielding 65 peer-reviewed articles for in-depth analysis. The corpus was coded thematically and mapped against three theoretical lenses: transaction cost theory, resource-based view, and network/ecosystem perspectives. Key findings reveal the following: 1. digital twins anchored in immersive platforms reduce planning cycles by up to 30% and enable real-time, cross-border supply chain reconfiguration; 2. tokenized assets, micro-transactions, and decentralized finance (DeFi) are spawning new revenue models but simultaneously shift tax triggers and compliance burdens; 3. cross-chain protocols are critical for scalable trust, yet regulatory fragmentation—exemplified by divergent EU, U.S., and APAC rules—creates non-trivial coordination costs; and 4. traditional IB theories require extension to account for digital-capability orchestration, emerging cost centers (licensing, reserve backing, data audits), and metaverse-driven network effects. Based on these insights, this study recommends that managers adopt phased licensing and geo-aware tax engines, embed region-specific compliance flags in smart-contract metadata, and pilot digital-twin initiatives in sandbox-friendly jurisdictions. Policymakers are urged to accelerate work on interoperability and reporting standards to prevent systemic bottlenecks. Finally, researchers should pursue multi-case and longitudinal studies measuring the financial and ESG outcomes of integrated blockchain–metaverse deployments. By synthesizing disparate streams and articulating a forward agenda, this review provides a conceptual bridge for international business scholarship and a practical roadmap for firms navigating the next wave of digital GVC transformation. Full article

Malware continues to evolve rapidly, posing significant challenges to network security. Traditional signature-based detection methods often struggle to cope with advanced evasion techniques such as polymorphism, metamorphism, encryption, and stealth, which are commonly employed by cybercriminals. As a result, these conventional approaches frequently fail to detect newly emerging malware variants in a timely manner. To address this limitation, Zero-Shot Learning (ZSL) has emerged as a promising alternative, offering improved classification capabilities for previously unseen malware samples. ZSL models leverage auxiliary semantic information and binary feature representations to enhance the recognition of novel threats. This study proposes a Transductive Zero-Shot Learning (TZSL) model based on the Vector Quantized Variational Autoencoder (VQ-VAE) architecture, integrated with a malware knowledge graph constructed from sandbox behavioral analysis of ransomware families. The model is further optimized through hyperparameter tuning to maximize classification performance. Evaluation metrics include per-family classification accuracy, precision, recall, F1-score, and Receiver Operating Characteristic (ROC) curves to ensure robust and reliable detection outcomes. In particular, the harmonic mean (H-mean) metric from the Generalized Zero-Shot Learning (GZSL) framework is introduced to jointly evaluate the model’s performance on both seen and unseen classes, offering a more holistic view of its generalization ability. The experimental results demonstrate that the proposed VQ-VAE model achieves an F1-score of 93.5% in ransomware classification, significantly outperforming other baseline models such as LeNet-5 (65.6%), ResNet-50 (71.8%), VGG-16 (74.3%), and AlexNet (65.3%). These findings highlight the superior capability of the VQ-VAE-based TZSL approach in detecting novel malware variants, improving detection accuracy while reducing false positives. Full article

Using the PRISMA-ScR methodology, this scoping review systematically analyzes how EU laws and regulations influence the development, adoption, and deployment of AI-driven digital solutions in energy generation, transmission, distribution, consumption, and markets. It identifies key regulatory barriers such as stringent risk assessments, cybersecurity obligations, and data access restrictions, along with enablers like regulatory sandboxes and harmonized compliance frameworks. Legal uncertainties, including AI liability and market manipulation risks, are also examined. To provide a comparative perspective, the EU regulatory approach is contrasted with AI governance models in the United States and China, highlighting global best practices and alignment challenges. The findings indicate that while the EU’s risk-based approach to AI governance provides a robust legal foundation, cross-regulatory complexity and sector-specific ambiguities necessitate further refinement. This paper proposes key recommendations, including the integration of AI-specific energy sector guidelines, acceleration of standardization efforts, promotion of privacy-preserving AI methods, and enhancement of international cooperation on AI safety and cybersecurity. These measures will help strike a balance between fostering trustworthy AI innovation and ensuring regulatory clarity, enabling AI to accelerate the clean energy transition while maintaining security, transparency, and fairness in digital energy systems. Full article

The rapid proliferation of ransomware variants necessitates more effective detection mechanisms, as traditional signature-based methods are increasingly inadequate. These conventional methods rely on manual feature extraction and matching, which are time-consuming and limited to known threats. This study addresses the escalating challenge of ransomware threats in cybersecurity by proposing a novel deep learning model, LSTM-EDadver, which leverages Generative Adversarial Networks (GANs) and Carlini and Wagner (CW) attacks to enhance malware detection capabilities. LSTM-EDadver innovatively generates adversarial examples (AEs) using sequential features derived from ransomware behaviors, thus training deep learning models to improve their robustness and accuracy. The methodology combines Cuckoo sandbox analysis with conceptual lattice ontology to capture a wide range of ransomware families and their variants. This approach not only addresses the shortcomings of existing models but also simulates real-world adversarial conditions during the validation phase by subjecting the models to CW attacks. The experimental results demonstrate that LSTM-EDadver achieves a classification accuracy of 96.59%. This performance was achieved using a dataset of 1328 ransomware samples (across 32 ransomware families) and 519 normal instances, outperforming traditional RNN, LSTM, and GCU models, which recorded accuracies of 90.01%, 93.95%, and 94.53%, respectively. The proposed model also shows significant improvements in F1-score, ranging from 2.49% to 6.64% compared to existing models without adversarial training. This advancement underscores the effectiveness of integrating GAN-generated attack command sequences into model training. Full article

This paper presents three 3D reconstructions of different analogue models used to reproduce, interpret, and describe the geological setting of a seismogenic area in Southern Italy—the Messina Strait. Three-dimensional analysis is a technique that allows for less sparse and more congruent and coherent information about a study zone whose complete understanding reduces uncertainties and risks. A thorough structural and geodynamic description of the effects of low-angle normal faulting in the same region through analogue models has been widely investigated in the scientific literature. Sandbox models for fault behaviour during deformation and the effects of a Low Angle Normal Fault (LANF) on the seismotectonic setting are also studied. The deformational patterns associated with seismogenic faults, rotational behaviour of faults, and other related problems have not yet been thoroughly analysed. Most problems, like the evolution of normal faults, fault geometry, and others, have been cited and briefly outlined in earlier published works, but a three-dimensional approach is still significant. Here, we carried out a three-dimensional digital model for a complete and continuous structural model of a debated, studied area. The aim of this study is to highlight the importance of fully representing faults in complex and/or non-cylindrical structures, mainly when the shape and dimensions of the fault(s) are key parameters, like in seismogenic contexts. Full article

The current cybersecurity ecosystem is proving insufficient in today’s increasingly sophisticated cyber attacks. Malware authors and intruders have pursued innovative avenues to circumvent emulated monitoring systems (EMSs) such as honeypots, virtual machines, sandboxes and debuggers to continue with their malicious activities while remaining inconspicuous. Cybercriminals are improving their ability to detect EMS, by finding indicators of deception (IoDs) to expose their presence and avoid detection. It is proving a challenge for security analysts to deploy and manage EMS to evaluate their deceptive capability. In this paper, we introduce the Hydrakon framework, which is composed of an EMS controller and several Linux and Windows 10 clients. The EMS controller automates the deployment and management of the clients and EMS for the purpose of measuring EMS deceptive capabilities. Experiments were conducted by applying custom detection vectors to client real machines, virtual machines and sandboxes, where various artifacts were extracted and stored as csv files on the EMS controller. The experiment leverages the cosine similarity metric to compare and identify similar artifacts between a real system and a virtual machine or sandbox. Our results show that Hydrakon offers a valid approach to assess the deceptive capabilities of EMS without the need to target specific IoD within the target system, thereby fostering more robust and effective emulated monitoring systems. Full article

In the context of global warming, rising sea levels are intensifying seawater intrusion in coastal areas. Due to the complex hydrodynamic conditions and increasing groundwater over-extraction in these regions, understanding the patterns of seawater intrusion is crucial for effective prevention and control. This study employed a sandbox model to investigate both vertical and horizontal seawater intrusion into a coastal unconfined aquifer with an impermeable dam under varying conditions of sea level rise, coastal slope, and groundwater pumping rate. Additionally, a two-dimensional SEAWAT model was developed to simulate seawater intrusion under these experimental conditions. The results indicate that sea level rise significantly increases the extent and intensity of seawater intrusion. When sea level rises by 3.5 cm, 4.5 cm, and 5.5 cm, the areas of the saline wedge reached 362 cm², 852 cm², and 1240 cm², respectively, with both horizontal and vertical intrusion ranges expanding considerably. When groundwater extraction is superimposed, vertical seawater intrusion is notably intensified. At an extraction rate of 225 cm³/min, the vertical intrusion areas corresponding to sea level rises of 3.5 cm, 4.5 cm, and 5.5 cm were 495 cm², 1035 cm², and 1748 cm², respectively, showing significant expansion, and this expansion becomes more pronounced as sea levels rise. In contrast, slope variations had a significant impact only on vertical seawater intrusion. As the slope decreased from tanα = 1/5 to tanα = 1/9, the upper saline wedge area expanded from 525 cm² to 846 cm², considerably increasing the vertical intrusion range. Finally, the combined effects of groundwater extraction and sea level rise exacerbate seawater intrusion more severely than either factor alone, presenting greater challenges for coastal water resource management. Full article

Light non-aqueous phase liquids (LNAPLs), which include various petroleum products, are a significant source of groundwater contamination globally. Once introduced into the subsurface, these contaminants tend to accumulate in the vadose zone, causing chronic soil and water pollution. The vadose zone often contains lens-shaped bodies with diverse properties that can significantly influence the migration and distribution of LNAPLs. Understanding the interaction between LNAPLs and these lens-shaped bodies is crucial for developing effective environmental management and remediation strategies. Prior research has primarily focused on LNAPL behavior in homogeneous media, with less emphasis on the impact of heterogeneous conditions introduced by lens-shaped bodies. To investigate the impact of lens-shaped structures on the migration of LNAPLs and to assess the specific effects of different types of lens-shaped structures on the distribution characteristics of LNAPL migration, this study simulates the LNAPL leakage process using an indoor two-dimensional sandbox. Three distinct test groups were conducted: one with no lens-shaped aquifer, one with a low-permeability lens, and one with a high-permeability lens. This study employs a combination of oil front curve mapping and high-density resistivity imaging techniques to systematically evaluate how the presence of lens-shaped structures affects the migration behavior, distribution patterns, and corresponding resistivity anomalies of LNAPLs. The results indicate that the migration rate and distribution characteristics of LNAPLs are influenced by the presence of a lens in the gas band of the envelope. The maximum vertical migration distances of the LNAPL are as follows: high-permeability lens (45 cm), no lens-shaped aquifer (40 cm), and low-permeability lens (35 cm). Horizontally, the maximum migration distances of the LNAPL to the upper part of the lens body decreases in the order of low-permeability lens, high-permeability lens, and no lens-shaped aquifer. The low-permeability lens impedes the vertical migration of the LNAPL, significantly affecting its migration path. It creates a flow around effect, hindering the downward migration of the LNAPL. In contrast, the high-permeability lens has a weaker retention effect and creates preferential flow paths, promoting the downward migration of the LNAPL. Under conditions with no lens-shaped aquifer and a high-permeability lens, the region of positive resistivity change rate is symmetrical around the axis where the injection point is located. Future research should explore the impact of various LNAPL types, lens geometries, and water table fluctuations on migration patterns. Incorporating numerical simulations could provide deeper insights into the mechanisms controlling LNAPL migration in heterogeneous subsurface environments. Full article

Automated sandbox-based analysis systems are dominantly focused on sequences of API calls, which are widely acknowledged as discriminative and easily extracted features. In this paper, we argue that an extension of the feature set beyond API calls may improve the malware detection performance. For this purpose, we apply the Cuckoo open-source sandbox system, carefully configured for the production of a novel dataset for dynamic malware analysis containing 22,200 annotated samples (11,735 benign and 10,465 malware). Each sample represents a full-featured report generated by the Cuckoo sandbox when a corresponding binary file is submitted for analysis. To support our position that the discriminative power of the full-featured sandbox reports is greater than the discriminative power of just API call sequences, we consider samples obtained from binary files whose execution induced API calls. In addition, we derive an additional dataset from samples in the full-featured dataset, whose samples contain only information on API calls. In a three-way factorial design experiment (considering the feature set, the feature representation technique, and the random forest model hyperparameter settings), we trained and tested a set of random forest models in a two-class classification task. The obtained results demonstrate that resorting to full-featured sandbox reports improves malware detection performance. The accuracy of 95.56 percent obtained for API call sequences was increased to 99.74 percent when full-featured sandbox reports were considered. Full article

Nowadays, some of the most expensive real estate is not “real” at all. Several investors are purchasing land in the virtual world of the Metaverse. To be more accurate in the wording, they are buying “meta-estates”. This work is dedicated to opening a debate about this emerging research field within the real estate discipline. It begins by discussing market segmentation, ownership, and pricing by comparing the traditional real estate market with the virtual estate market. Furthermore, this study involved interviews with six seasoned Metaverse land investors who participated in two Analytic Hierarchy Processes (AHPs). The first AHP ranked 14 investment typologies, while the second focused on ranking and discussing the most important characteristics of meta-estates that influence the formation of prices. As a result, the most appealing investments identified were day-trading, virtual land trading (buying to resell), and virtual land development (transforming and reselling). The primary characteristics of meta-estates considered by investors include the platform (e.g., Earth 2, Sandbox), the location within the platform (proximity to famous neighbours), and the architectural design of the buildings (designed by renowned architects). It is evident that the Metaverse represents a new frontier for land investors, and the primary aim of this study was to encourage other researchers to explore and investigate this evolving field. Full article

The aim of this paper is to discuss how young Latinx men living in Australia negotiate, embody, and complicate existing dominant and racialized masculinities. Queer and feminist theories are used to explore how Latinx men negotiate and embody masculinities, sexualities, and being ‘other’ in a White dominant cultural context. These tensions were explored through semi-structured in-depth interviews and a creative visual method known as sandboxing with twenty-one Latinx men. Sandboxing aims to elicit conversation and allows for the reflection and sharing of a visual and symbolic representation of participants’ lives. The findings suggest masculinities are lived and embodied alongside negotiating racialization and sexualities. The fluidity of masculinities surfaces in participants’ reflexive engagement with masculinities and the nuances in negotiating and simultaneously reproducing gender binary norms. Participants’ careful negotiation in engaging with feminine culture led to developing the concept ‘feminine threshold’, a theoretical contribution offered in this article, in understanding how Latinx men negotiate masculinities. Full article