Search Results (253)

As disaster imagery and social media reports become vital for crisis response, automated assessment systems must address challenges of multimodal integration, privacy-aware learning, and computational efficiency. To address these challenges, we propose TriDA, a privacy-aware and efficiency-conscious multimodal disaster classification framework that fuses image features with text representations through a late-fusion design. A classifier-head DP-SGD stage is used to report training-record-level differential privacy accounting for paired image–text samples under the stated private optimization protocol. To study efficiency-oriented simplification, structured neuron pruning reduces redundant capacity in the classification head while preserving predictive utility. Experiments on the multimodal damage identification dataset show that TriDA maintains strong classification performance, exhibits a controlled privacy–utility trade-off under increasing DP noise, and achieves quantifiable classifier-head parameter and MAC reductions through pruning. These findings position TriDA as a controlled empirical framework for privacy-aware and resource-conscious multimodal disaster assessment. Full article

The increasing integration of distributed energy resources (DERs), including photovoltaic systems, battery energy storage systems, electric vehicles, and flexible loads, is transforming modern power systems and creating new challenges for coordination, control, and cybersecurity. Conventional Virtual Power Plant (VPP) architectures typically rely on centralized energy management systems, which may face scalability limitations, communication bottlenecks, cybersecurity risks, and reduced resilience to failures. This paper presents a blockchain-enabled decentralized Virtual Power Plant framework for secure and autonomous coordination of distributed energy resources. The proposed architecture combines blockchain technology, smart contracts, IoT-based communication infrastructure, and decentralized energy management functions within a unified multi-layer coordination framework. Smart contracts automate energy scheduling, peer-to-peer transaction validation, and settlement processes, reducing dependence on centralized control entities. Lightweight blockchain consensus mechanisms are employed to improve scalability while limiting computational overhead. The effectiveness of the proposed framework is evaluated through simulation-based case studies involving decentralized DER coordination, peer-to-peer energy trading, and resilience assessment under node-failure conditions. Its performance is compared with that of a conventional centralized VPP architecture in terms of scalability, coordination reliability, communication overhead, transaction transparency, and fault tolerance. The results indicate that the decentralized framework improves operational resilience, coordination transparency, and scalability under increasing DER participation while maintaining satisfactory energy balancing performance. Although blockchain-based coordination introduces additional transaction latency, the proposed approach enhances cybersecurity, reduces dependence on centralized control structures, and provides a flexible foundation for future intelligent smart-grid applications. Full article

This paper proposes a decentralized data trading approach based on the Automated Market Maker (AMM) mechanism, aiming to address the bottlenecks in data trading efficiency and fairness via the collaborative innovation of market-oriented pricing mechanisms and automated trading processes. We focus on constructing an automatic pricing and matching mechanism based on liquidity pools. Subsequently, mathematical modeling and simulations reveal the slippage generation mechanism in data liquidity pools under trading shocks and imbalances. To address these issues, a novel dual slippage optimization mechanism integrating dynamic trade splitting and alternating order sorting is proposed, which decomposes orders into sub-orders and reorganizes their timing, establishing a dynamic equilibrium model. Experiments show that the method reduces the average slippage amplitude from 2.1% to 0.5%, representing a 76.2% reduction, and significantly enhances price stability and market efficiency. This research explores the mechanism of applying AMM to data asset trading and mitigates the limitations of AMM in this scenario, providing a theoretical and empirical foundation for building low-cost, high-fairness data trading systems through mechanism innovation and technical optimization. Full article

The convergence of Software-Defined Networking (SDN) and Blockchain (BC) creates a symbiotic relationship in which SDN’s programmable global visibility complements BC’s decentralized, immutable trust model to address critical cybersecurity vulnerabilities and cyber attacks. Addressing the fragmentation in the current literature, this study rigorously investigates BC and SDN (B-SDN) integration with the primary objectives of: (1) differentiating impacts across varied sectors, including the Internet of Things (IoT), Smart Grids, and Vehicular Ad Hoc Networks (VANETs) and more; (2) analyzing critical performance metrics such as energy efficiency and scalability; (3) classifying mitigation, detection, and prevention schemes for specific threats; (4) examining novel Artificial Intelligence (AI) methods; and (5) identifying open challenges and future research directions. Methodologically, this study conducts a survey of state-of-the-art B-SDN studies to investigate six key areas: Industry-specific applications, security mechanisms, defense strategies, defenses against specific attacks, AI integration, and implementation performance. The findings demonstrate that B-SDN integration shows strong potential in simulated and prototype environments to mitigate specific high-impact threats, such as Distributed Denial of Service (DDoS), Man-in-the-Middle (MiTM), and spoofing, across various domains, including IoT, 5G/6G, VANETS, and Smart Grid. Despite the benefits and advantages promised by B-SDN, several limitations continue to exist, including the latency–security trade-off inherent to consensus protocols and scalability constraints in large-scale deployments. Finally, open research challenges persist in AI-driven automation, particularly in Federated Learning (FL) and in the development of standardized interoperability protocols required to enable the transition from conceptual models to operational systems. Full article

The proliferation of digital financial news offers unprecedented opportunities for automated analysis of market-moving events. This paper presents a framework for event-based sentiment analysis of financial news that leverages Large Language Models (LLMs). The approach brings together three complementary ideas: Retrieval-Augmented Generation (RAG) for contextual enhancement, Graph Neural Networks (GNNs) for modeling relationships between events, and a multi-agent ensemble for orchestrated reasoning. The methodology targets well-known difficulties in financial text processing, including domain-specific terminology, implicit event detection, and temporal reasoning, and it combines transformer-based event extraction with sentiment classification enhanced by external knowledge retrieval. We evaluate six model configurations on an aggregated corpus of 14,851 financial news samples. On the event-detection task, every configuration reaches a weighted F1-score of 100%; we show that this is a ceiling effect produced by a binary event/no-event formulation over a highly imbalanced dataset rather than evidence of a difficult problem being solved, and we discuss what it implies for how such systems should be evaluated. On three-way sentiment classification, the strongest configuration—the multi-agent ensemble—reaches 87.4% accuracy, narrowly ahead of a RoBERTa (Robustly Optimized BERT Pretraining Approach) baseline at 87.2%; however, because the gaps reported between models are small and we did not run significance testing, we report them as indicative rather than definitive. The GNN component is described as part of the proposed design, but it has not yet been validated experimentally, and we state this limitation explicitly. The framework produces interpretable, structured outputs suited to downstream use in algorithmic trading, risk assessment, and investment decision support, and the paper contributes a reusable financial NLP pipeline together with a candid account of where the current evidence is, and is not, convincing. Full article

The transition toward smart grids and Industry 4.0 demands a fundamental shift in maintenance strategies, as manual inspection methods are increasingly being supplanted by automated monitoring systems. Among the advanced technologies for smart inspection, infrared imaging has advantages including non-contact operation, intuitive visualization, and predictive capabilities, which has become a cornerstone for autonomous inspection of critical power infrastructure. This review provides recent advancements in infrared imaging, with a specific focus on automated power system inspection. The discussion starts with an overview of the fundamental principles and system architectures, emphasizing the pivotal role of infrared detectors. A detailed analysis traces the technological evolution from traditional photon detectors to current uncooled microbolometers, and critically assesses emerging low-dimensional materials. The analysis highlights inherent performance trade-offs among sensitivity, operating temperature, and fabrication cost. Subsequently, the review explores advanced signal processing algorithms, such as real-time non-uniformity correction and adaptive noise suppression, which are typically implemented on FPGA platforms. Advanced optical configurations—encompassing computational imaging, lensless designs, and scattering suppression methods—are also discussed, demonstrating how their convergence enhances image fidelity and operational reliability in complex field environments. Representative application paradigms are surveyed, including drone-based transmission line inspections, patrol robots in substations, and fault diagnosis in photovoltaic plants; for each, operational efficacy and economic benefits are assessed. Despite considerable progress, several challenges persist, notably the performance–stability–cost trilemma in novel detector development, the substantial computational demands of end-to-end optimized systems, and a lack of standardization. Finally, the review outlines future research directions, such as high-performance uncooled arrays, AI-driven co-design of optics and algorithms, and the development of standardized, low-cost, intelligent inspection platforms. Full article

The rapid advancement of Generative Artificial Intelligence (Gen-AI) is fundamentally reshaping labor markets, creating an urgent need for higher education institutions to adapt their program capacities and curricula. This paper proposes a data-driven Robust Multi-Objective Planning (RMOP) framework to translate heterogeneous Gen-AI labor shocks into actionable, program-level decisions regarding enrollment scaling and curriculum design. Grounded in O*NET micro-task structures, we model occupational evolution as a dynamic system of substitution, augmentation, and insulation driven by logistic technology diffusion. Our simulations across STEM, trade, and arts occupations reveal sharply divergent trajectories: Information Security Engineers face a 62% total impact dominated by substitution, whereas Electricians retain over 80% insulation, and Musicians experience high exposure but low substitution. To bridge these macro-level forecasts with immediate institutional maneuvers, the framework couples an AI-adjusted Grey Model (GM(1,1)) demand model with a Program Effectiveness Index (PEI) to yield discrete enrollment policy levers (Expand, Contract, and Adjust). For curriculum optimization, we employ Ridge regression to rank employability-related curriculum drivers and NSGA-II to generate Pareto portfolios under competing institutional objectives, including employability, instructional cost, ethics, and environmental impact. Final implementable recommendations are selected through entropy-weighted TOPSIS, where student well-being and education equity are treated as supplementary decision criteria rather than direct prediction targets. In addition, an Automation Risk Score (ARS) and a K-means TC clustering module are used to illustrate potential transfer paths across broader institutional settings. Internal scenario checks show that the AI-adjusted GM(1,1) reduces average hold-out MAPE from 7.0% to 5.8% relative to the baseline GM(1,1), and that NSGA-II achieves slightly stronger Pareto coverage than MOPSO and MODE under the same curriculum-portfolio setting. These checks are interpreted as preliminary decision-support evidence rather than external predictive validation. Overall, RMOP is presented as a scenario-based decision-support framework that links Gen-AI occupational exposure, enrollment adjustment, and curriculum portfolio design. Full article

Urban sustainability management increasingly relies on large volumes of heterogeneous environmental data generated by smart city infrastructures. While these data streams offer significant potential for evidence-informed policymaking, environmental governance, and public engagement, their effective use is often constrained by technical barriers and persistent data-skills gaps among non-specialist stakeholders. Using urban air quality as a policy-relevant and data-rich sustainability domain, this paper presents a proof-of-concept dashboard that investigates how large language model (LLM)-enabled natural language interfaces can lower barriers to querying, analysing, and visualising urban environmental data. The system translates natural language questions into executable database queries and automatically generates visualisations over air-quality datasets. A controlled comparative benchmark of proprietary and open-source LLMs is conducted to assess their suitability for text-to-SQL generation in this application context. In this benchmark, proprietary GPT-based models achieved the highest observed query accuracy and robustness among the evaluated models, highlighting practical trade-offs between performance, transparency, reproducibility, and long-term governance. This paper makes a twofold contribution: First, it demonstrates the technical feasibility of an LLM-enabled natural language access layer for smart-city environmental data. Second, it uses the implemented system as a concrete case through which to analyse the trust, transparency, inclusivity, vendor-dependency, and data-quality challenges that arise when such systems are incorporated into sustainability-oriented decision-support workflows. The study provides a transferable design contribution for urban sustainability data access by showing how natural language interfaces, model benchmarking, automated visualisation, and governance-aware system design can be combined to support more inclusive interaction with complex environmental datasets. Full article

Sign language understanding for automated vehicles sits at the intersection of accessibility, intelligent transportation, and safety-critical human–machine interaction. The existing sign-language recognition systems are largely confined to controlled environments, limiting their utility in mobility scenarios characterized by lighting variation, motion blur, and partial occlusion. This paper proposes STCM-HVNet, a safety-aware hybrid multimodal architecture integrating four components: a spatial visual encoder, a MediaPipe-based pose encoder, a bidirectional LSTM temporal encoder, and a context-aware fusion and safety decision module. The architecture is formulated as a multi-task system that jointly predicts sign category, interaction intent, and urgency level, and incorporates confidence-aware rejection and fail-safe action mapping. Experiments are conducted on two Arabic sign-language resources. On the RGBArS image benchmark (31 classes, 7856 images), the proposed pipeline achieves a Top-1 accuracy of 45.38%, Top-3 accuracy of 75.15%, and Macro-F1 of 0.4479, outperforming LinearECOC, kNN-5, and Bagged Trees baselines. On the Arabic sign-language video benchmark (12 classes, 479 clips), the BiLSTM temporal encoder achieves a Top-1 accuracy of 93.15% and Macro-F1 of 0.9383, outperforming frame-aggregation (87.67%) and CNN-LSTM (89.04%) baselines. Ablation results confirm complementary contributions from the visual and pose branches. A safety-threshold analysis and a Monte Carlo dropout comparison demonstrate that the proposed safety decision/gating layer provides a controllable trade-off between prediction coverage and reliability. Full article

Electrocardiogram (ECG) arrhythmia classification remains critical for automated cardiac diagnosis, yet feature extraction methods are frequently adopted without systematic comparative evaluation. This study presents a controlled comparative analysis of four signal processing techniques—Mel-Frequency Cepstral Coefficients (MFCC), Discrete Wavelet Transform (DWT), Hilbert–Huang Transform (HHT), and Synchrosqueezing Wavelet Transform (SSWT)—for ECG feature extraction. Using the MIT-BIH Arrhythmia Database with ANSI/AAMI EC57:1998 standard mapping, we trained Cascade Forest classifiers on each feature set under identical preprocessing and SMOTE-based class balancing conditions to ensure a fair comparison. DWT features achieved superior performance (accuracy: 98.79%, macro-F1: 92.93%, precision: 94.39%) compared to MFCC (88.30% macro-F1), SSWT (84.54% macro-F1), and HHT (83.59% macro-F1), particularly for clinically challenging minority arrhythmia classes. However, DWT’s performance advantage incurred substantial computational cost (10,050 s), while MFCC provided competitive results with a 62% lower computational burden. These findings provide evidence-based guidance for feature extraction method selection in interpretable ECG classification systems, demonstrating critical performance-efficiency trade-offs relevant to clinical deployment contexts. Full article

The rapid integration of large language models (LLMs) into education has shifted research focus from questions of capability, such as what LLMs can do and how accurately—to questions of deployability, including how they can be operated effectively for many learners at once. In school environments, system reliability, scalability, and real-time responsiveness are critical, as delays or interruptions can directly reduce learner engagement, particularly during synchronous activities. This study evaluates the performance of open-source LLaMA models ranging from 1 billion to 70 billion parameters across one-, dual-, triple-, and quad-GPU configurations suitable for educational settings. Performance is assessed using four key indicators: success rate (percentage of completed requests), generation speed (tokens per second), throughput (completed responses per second), and latency (time until full response generation). These metrics were measured under progressively increasing numbers of simultaneous users to identify system capacity limits and trade-offs between model size, responsiveness, and scalability. The results indicate that smaller models (1B–3B) deliver faster, more stable performance under concurrent use, while larger models (8B–70B) experience substantial slowdowns and reduced reliability, even on high-end GPU systems. These findings suggest that effective educational deployment should prioritize empirical performance and infrastructure compatibility over model size alone. The paper concludes by proposing a practical framework to guide educators, administrators, and developers in selecting and configuring locally hosted GPU systems that balance model capability, response speed, and resource efficiency for real-time applications such as AI tutors, classroom chatbots, and automated feedback tools. Full article

The escalating sophistication of ransomware requires defensive strategies that are both proactive against zero-day attacks and operationally efficient. Existing solutions often force a trade-off—sacrificing low false-positive rates for broad detection, or vice versa. This work introduces an integrated framework designed to transcend this limitation. Our dual-stage approach synergizes pre-encryption behavioral analysis with definitive post-encryption confirmation. The first stage employs a specialized artificial immune system (AIS) that monitors a curated set of 47 features, including API-call n-grams and file entropy dynamics, to identify malicious activity before file encryption begins. This pre-emptive analysis is complemented by an enhanced, cross-platform R-Locker mechanism, which uses Windows named pipes and symbolic links to deploy honeyfiles that trap ransomware during I/O operations, providing a high-fidelity trigger for automated containment. We subjected this framework to a rigorous evaluation against 3500 real-world ransomware samples and 12,000 benign applications. The results demonstrate a 98.2% detection rate with a 0.8% false-positive rate, achieving a mean response time of 1.3 s. A key finding is the framework’s efficiency on both Windows and Linux (the only platforms tested), with the AIS and R-Locker modules consuming a combined 101 MB of memory. While the system excels in real-time detection, we note that its current memory forensics capability for key recovery is incompatible with certain ransomware families due to architectural obfuscations. Our findings suggest that the integrated approach performs well under laboratory conditions; further real-world validation is required to confirm robustness in diverse environments. Full article

Recognizing human actions from static images in complex industrial environments remains challenging due to insufficient feature representation and high computational complexity. This issue is particularly critical in power-grid safety monitoring, where improper worker postures (e.g., bending, climbing, falling) can lead to severe accidents and personal injuries, necessitating automated monitoring systems that operate reliably on resource-constrained edge devices. This study proposes a bio-inspired lightweight recognition framework that integrates an improved YOLO-Pose model with a gated recurrent unit (GRU) network. The scientific motivation is grounded in the observation that the human musculoskeletal system achieves highly efficient motion perception through three key mechanisms: hierarchical muscle coordination providing intrinsic rotation invariance, proprioceptive feedback enabling real-time error correction, and selective neural gating reducing redundant information transmission. These biological principles directly inspire our technical contributions: polar-coordinate encoding provides rotation invariance, three-stage filtering mimics proprioceptive feedback, and GRU gating mirrors selective information propagation. Unlike prior approaches that treat pose-based action recognition as a generic computer vision problem, this work explicitly incorporates anatomical structural constraints into the computational pipeline. The framework addresses three research gaps: (1) existing methods lack biomechanically derived invariance properties; (2) GCN-based approaches use fixed topologies that fail to adapt to occlusion patterns; (3) the trade-off between model complexity and accuracy remains unsatisfactory for edge deployment. Experiments on the self-constructed SKPose dataset demonstrate that the proposed method achieves 95.04% accuracy, outperforming ST-GCN by 3.67 percentage points and 2s-AGCN by 1.94 percentage points, with an inference speed of 48 FPS on 8.7 M parameters in underground power-grid environments and provides practical support for biomimetic perception systems and industrial safety monitoring. Full article

Precise in situ characterization of the mechanical properties of lunar regolith is critical for future lunar base construction and resource exploitation. However, existing detection methods predominantly rely on indirect inversion from rover wheel-soil interactions, which exhibit limitations in accuracy, real-time capability, and detection depth. Furthermore, specialized automated equipment capable of adapting to the complex lunar surface environment remains lacking. To address these challenges, this study presents the design and development of a novel autonomous in situ penetration-shear apparatus. The device automatically executes penetration and shear operations while recording real-time data, with a maximum penetration force of 25 N, shear torque of 2.5 N·m, penetration depth of 300 mm, and rotation angle of 360°. Given the maximum normal load constraint of 16 N imposed by the lunar rover platform, 24 probe configurations—varying in conicity, projected area, and vane number—were systematically evaluated using lunar soil simulants with three particle size distributions and two density levels. Multi-objective optimization was conducted to maximize detection efficiency, specifically penetration depth and shear torque, subject to a lightweight payload constraint (16 N). The multi-objective optimization reveals a fundamental trade-off: smaller conicity angles and projected areas favor deeper penetration, while larger projected areas enhance shear torque response. Under the 16 N constraint, the Pareto analysis identifies that a combination of moderate projected area, small conicity, and fewer vanes achieves the most balanced performance across all soil conditions. Results further demonstrate that increasing particle size and density substantially suppress both penetration capability and shear torque response, with compaction being the dominant factor limiting probe advancement under constrained normal loading. Results indicate that the optimal probe configuration comprises a 15° conicity, 324 mm² projected area, and two vanes, achieving an average penetration depth of 51.61 mm and average shear torque of 0.06 N·m across all test conditions. This study validates a complete automated system for characterizing lunar soil mechanical properties and provides an efficient, reliable hardware solution for future unmanned lunar exploration missions through optimized probe design. These findings establish a solid technical foundation for deep, high-precision in situ investigation of lunar soil structure and mechanical parameters, with significant implications for lunar base site selection and In Situ Resource Utilization (ISRU). Full article

The Index Selection Problem (ISP) refers to the task of automatically identifying the most appropriate set of indexes for a given database workload that can minimize execution costs. However, ISP is a fundamental yet complex challenge in database management systems. In the era of data-intensive applications, efficient index strategies are increasingly necessary to maintain scalability and responsiveness. This paper presents a novel automated index selection algorithm for centralized databases that employs a Double Deep Q-Network (DDQN) as the classical learning backbone and extends it with quantum-enhanced variants. Two hybrid quantum variants were proposed: Quantum Double Deep Q-Network Mixed (QDDQNM), which incorporates a residual classical pathway, and Quantum Double Deep Q-Network Boosted (QDDQNB), a boosted model without residuals. All variants were systematically evaluated using the TPC-H benchmark at two small scale factors, 10 MB and 100 MB. Experimental results show that the evaluated Deep Reinforcement Learning (DRL)-based methods improve on the SMARTIX baseline within this proof-of-concept setting. The quantum-enhanced models achieved higher best-run accumulated rewards in the reported experiments, but they also incurred substantially higher simulation cost. The results therefore suggest interesting hybrid learning behavior under the tested conditions, while also highlighting that practical scalability and cost-performance trade-offs remain important limitations for future work. Full article