Search Results (1,724)

The proliferation of Internet of Things (IoT) devices has significantly expanded the threat landscape for malicious software (malware), rendering traditional signature-based detection methods increasingly ineffective in coping with the volume and evolving nature of modern threats. In response, researchers are utilising artificial intelligence (AI) for a more dynamic and robust malware detection solution. An innovative approach utilising AI is focusing on image classification techniques to detect malware on resource-constrained Single-Board Computers (SBCs) such as the Raspberry Pi. In this method the conversion of malware binaries into 2D images is examined, which can be analysed by deep learning models such as convolutional neural networks (CNNs) to classify them as benign or malicious. The results show that the image-based approach demonstrates high efficacy, with many studies reporting detection accuracy rates exceeding 98%. That said, there is a significant challenge in deploying these demanding models on devices with limited processing power and memory, in particular those involving of both calculation and time complexity. Overcoming this issue requires critical model optimisation strategies. Successful approaches include the use of a lightweight CNN architecture and federated learning, which may be used to preserve privacy while training models with decentralised data are processed. This hybrid workflow in which models are trained on powerful servers before the learnt algorithms are deployed on SBCs is an emerging field attacting significant interest in the field of cybersecurity. This paper synthesises the current state of the art, performance compromises, and optimisation techniques contributing to the understanding of how AI and image representation can enable effective low-cost malware detection on resource-constrained systems. Full article

Acute Appendicitis (AA) is the commonest abdominal digestive surgical emergency, but its etiology is not clarified. Based on histologic observations, an allergic cause has been proposed. In a type I hypersensitivity allergic reaction, there is a Th2 immune response characterized by Th2 cells, eosinophils, basophils, IgE, IL-4, IL-5, and IL-13 serum elevation. Recent studies showed a local appendicular endoluminal and parietal Th2 immune response in acute phlegmonous appendicitis. We performed a prospective single-center study where we evaluated the Th2 blood immune response in 38 patients with acute phlegmonous appendicitis, 27 patients with acute gangrenous appendicitis, and 18 patients with the clinical picture of AA, who underwent appendectomy but had negative histology for AA (negative appendectomy group). Higher levels of basophils were found in phlegmonous appendicitis (p = 0.03), and higher levels of eosinophils were found in the control group (p = 0.003). Effector memory CD4 T cells re-expressing CD45RA were higher in gangrenous (p = 0.020) and central memory CD4 T cells in phlegmonous appendicitis (p = 0.004). The number of Th2 circulating cells was higher in gangrenous appendicitis (p = 0.037), while Th1 circulating cells were higher in phlegmonous appendicitis (p = 0.028). IL-4 blood concentrations were elevated in acute gangrenous appendicitis (p = 0.029). No significant differences were found in the levels of IgE, IL-5, or IL-13 in any of the groups. Thus, a Th2 response was not detected in patients’ serum with phlegmonous appendicitis. Serum levels of IgE, IL-5, and IL-13 were not different among patients with acute phlegmonous appendicitis, acute gangrenous appendicitis, and the negative appendectomy group. These findings are in contrast to our previous work in which we evaluated the Th2 response at the local level, at the appendicular luminal aspect and appendicular wall, in phlegmonous appendicitis and control groups, and we unequivocally showed a Th2 response in phlegmonous appendicitis. Thus, in patients with phlegmonous appendicitis, the local Th2 response is not reflected in the serum levels of immune cells and cytokines. Full article

Secure and auditable data sharing in large-scale Internet of Things (IoT) environments remains a significant challenge due to weak trust coordination, limited scalability, and susceptibility to emerging quantum attacks. This study introduces a hybrid blockchain-based framework that integrates post-quantum cryptography with intelligent anomaly detection to ensure end-to-end data integrity and resilience. The proposed system utilizes Hyperledger Fabric for permissioned device lifecycle management and Ethereum for public auditability of encrypted telemetry, thereby providing both private control and transparent verification. Device identities are established using quantum-entropy-seeded credentials and safeguarded with lattice-based encryption to withstand quantum adversaries. A convolutional long short-term memory (CNN–LSTM) model continuously monitors device behavior, facilitating real-time trust scoring and autonomous revocation via smart contract triggers. Experimental results demonstrate 97.4% anomaly detection accuracy and a 0.968 F1-score, supporting up to 1000 transactions per second with cross-chain latency below 6 s. These findings indicate that the proposed architecture delivers scalable, quantum-resilient, and computationally efficient data sharing suitable for mission-critical IoT deployments. Full article

Mega-event-led urban waterfront regeneration has played a key role in shaping public open spaces, particularly in newly developed areas within the Chinese context. However, public perceptions and their influence on the use of newly built open spaces created through mega-event-led regeneration have not been examined in existing research. To address this gap, this study establishes an integrated assessment framework to evaluate the quality of urban waterfront open spaces. A mixed methods approach was adopted, including direct observations and 770 online questionnaires collected between July and October 2024 at the South Bank of the Qiantang River (SBQR) in Hangzhou, China. Spatial analysis and Importance–Performance Analysis (IPA) were employed to determine priority improvement areas that should inform future waterfront regeneration strategies. The results indicate that inclusiveness emerged as the most important factor for enhancing waterfront open space quality, while spatial aesthetics ranked the lowest. Among the sub-sub factors, elements related to improving water accessibility, enhancing natural surveillance, providing artificial shelters and diverse seating options, introducing distinctive water features, and shaping collective memory through digital technologies are the key priorities for improvement in the future urban waterfront regeneration policies. Finally, the study highlights that the intangible legacies of the Asian Games and the adaptive reuse of informal built heritage have the potential to reshape a distinctive new city image and collective memory, even in the absence of tangible and formally recognised heritage buildings. Full article

The exponential growth of data-driven applications has intensified performance demands on database systems, where cache management represents a critical bottleneck. Traditional cache replacement policies such as Least Recently Used (LRU) and Least Frequently Used (LFU) rely on simple heuristics that fail to capture complex temporal and frequency patterns in modern workloads. This research presents a modular machine learning-enhanced cache management framework that leverages pattern recognition to optimize database performance through intelligent replacement decisions. Our approach integrates multiple machine learning models—Random Forest classifiers, Long Short-Term Memory (LSTM) networks, Support Vector Machines (SVM), and Gradient Boosting methods—within a modular architecture enabling seamless integration with existing database systems. The framework incorporates sophisticated feature engineering pipelines extracting temporal, frequency, and contextual characteristics from query access patterns. Comprehensive experimental evaluation across synthetic workloads, real-world production datasets, and standard benchmarks (TPC-C, TPC-H, YCSB, and LinkBench) demonstrates consistent performance improvements. Machine learning-enhanced approaches achieve $8.4\%$ to $19.2\%$ improvement in cache hit rates, $15.3\%$ to $28.7\%$ reduction in query latency, and $18.9\%$ to $31.4\%$ increase in system throughput compared to traditional policies and advanced adaptive methods including ARC, LIRS, Clock-Pro, TinyLFU, and LECAR. Random Forest emerges as the most practical solution, providing $18.7\%$ performance improvement with only $3.1\%$ computational overhead. Case study analysis across e-commerce, financial services, and content management applications demonstrates measurable business impact, including $8.3\%$ conversion rate improvements and USD 127,000 annual revenue increases. Statistical validation ($p<0.001$, Cohen’s $d>0.8$) confirms both statistical and practical significance. Full article

Symmetry and asymmetry are foundational design principles in artificial intelligence, defining the balance between invariance and adaptability in multimodal learning systems. In audio-visual speaker verification, where speech and lip-motion features are jointly modeled to determine whether two utterances belong to the same individual, these principles govern both fairness and discriminative power. In this work, we analyze how symmetry and asymmetry emerge within a gated-fusion architecture that integrates Time-Delay Neural Networks and Bidirectional Long Short-Term Memory encoders for speech, ResNet-based visual lip encoders, and a shared Conformer-based temporal backbone. Structural symmetry is preserved through weight-sharing across paired utterances and symmetric cosine-based scoring, ensuring verification consistency regardless of input order. In contrast, asymmetry is intentionally introduced through modality-dependent temporal encoding, multi-head attention pooling, and a learnable gating mechanism that dynamically re-weights the contribution of audio and visual streams at each timestep. This controlled asymmetry allows the model to rely on visual cues when speech is noisy, and conversely on speech when lip visibility is degraded, yielding adaptive robustness under cross-modal degradation. Experimental results demonstrate that combining symmetric embedding space design with adaptive asymmetric fusion significantly improves generalization, reducing Equal Error Rate (EER) to 3.419% on VoxCeleb-2 test dataset without sacrificing interpretability. The findings show that symmetry ensures stable and fair decision-making, while learnable asymmetry enables modality awareness together forming a principled foundation for next-generation audio-visual speaker verification systems. Full article

Global climate change demands modeling approaches that are both computationally efficient and physically faithful to the system’s long-term dynamics. Classical Energy Balance Models (EBMs), while valuable, are fundamentally limited by their memoryless exponential response, which fails to represent the prolonged thermal inertia of the climate system—particularly that associated with deep-ocean heat uptake. In this study, we introduce a fractional Energy Balance Model (fEBM) by replacing the classical integer-order time derivative with a Caputo fractional derivative of order $\alpha (0<\alpha \le 1)$, thereby embedding long-range memory directly into the model structure. We establish a rigorous mathematical foundation for the fEBM, including proofs of existence, uniqueness, and asymptotic stability, ensuring theoretical well-posedness and numerical reliability. The model is calibrated and validated against historical global mean surface temperature data from NASA GISTEMP and radiative forcing estimates from IPCC AR6. Relative to the classical EBM, the fEBM achieves a substantially improved representation of observed temperatures, reducing the root mean square error by approximately 29% during calibration (1880–2010) and by 47% in out-of-sample forecasting (2011–2023). The optimized fractional order $\alpha =0.75\pm 0.03$ emerges as a physically interpretable measure of aggregate climate memory, consistent with multi-decadal ocean heat uptake and observed persistence in temperature anomalies. Residual diagnostics and robustness analyses further demonstrate that the fractional formulation captures dominant temporal dependencies without overfitting. By integrating mathematical rigor, uncertainty quantification, and physical interpretability, this work positions fractional calculus as a powerful and responsible framework for reduced-order climate modeling and long-term projection analysis. Full article

Repeated exposure to low-level blast overpressure (BOP) during controlled detonations is an emerging occupational health concern for military breachers and Special Operations Forces personnel, given accumulating evidence that chronic exposure may produce subtle, subclinical neurotrauma. This study derived a latent neuroinjury construct integrating three complementary domains of brain health—post-concussive symptoms, working-memory performance, and circulating biomarkers—to determine whether breachers exhibit coherent patterns of neurobiological alteration. Symptom severity was assessed using the Rivermead Post-Concussion Questionnaire (RPQ), and working memory was assessed with the N-Back task and a panel of thirteen neuroproteomic biomarkers was measured reflecting astroglial activation, neuronal and axonal injury, oxidative stress, inflammatory signaling, and neurotrophic regulation. Experienced Canadian Armed Forces breachers with extensive occupational BOP exposure were compared with unexposed controls. Bayesian latent-variable modeling provided probabilistic evidence for a chronic, subclinical neurobiological signal, with the strongest contributions arising from self-reported symptoms and smaller but consistent contributions from the biomarker domain. Working-memory performance did not load substantively on the latent factor. Several RPQ items and circulating biomarkers showed robust loadings, and the latent neuroinjury factor was elevated in breachers relative to controls (97% posterior probability). The pattern is broadly consistent with subclinical neurobiological stress in the absence of measurable cognitive impairment, suggesting early or compensated physiological alterations rather than overt dysfunction. This multidomain, biomarker-informed framework provides a mechanistically grounded and scalable approach for identifying subtle neurobiological strain in military personnel routinely exposed to repetitive low-level blast. It may offer value for risk stratification, operational health surveillance, and the longitudinal monitoring of neurobiological change in high-risk occupations. Full article

Financial risk early warning systems are essential for proactive risk management in volatile markets, particularly for emerging economies such as China. This study develops a hybrid deep learning model integrating Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRUs) to enhance the accuracy and robustness of financial risk prediction. Using firm-level quarterly financial data from Chinese listed companies, the proposed model is benchmarked against standalone CNN, LSTM, and GRU architectures. Experimental results show that the hybrid CNN–LSTM–GRU model achieves superior performance across all evaluation metrics, with prediction accuracy reaching 93.5%, precision reaching 92.2%, recall reaching 91.8%, and F1-score reaching 92.0%, significantly outperforming individual models. Moreover, the hybrid approach demonstrates faster convergence than LSTM and improved class balance compared to CNN and GRU, reducing false negatives for high-risk firms—a critical aspect for early intervention. These findings highlight the hybrid model’s robustness and real-world applicability, offering regulators, investors, and policymakers a reliable tool for timely financial risk detection and informed decision-making. By combining high predictive power with computational efficiency, the proposed system provides a practical framework for strengthening financial stability in emerging and dynamic markets. Full article

Drug–drug interactions (DDIs) can cause adverse reactions or reduce the efficiency of a drug. Using computers to predict DDIs is now critical in pharmacology, as this reduces risks, improves drug outcomes and lowers healthcare costs. Clinical trials are slow, expensive, and require a lot of effort. The use of artificial intelligence (AI), primarily in the form of machine learning (ML) and its subfield deep learning (DL), has made DDI prediction more accurate and efficient when handling large datasets from biological, chemical, and clinical domains. Many ML and DL approaches are bio-inspired, taking inspiration from natural systems, and are considered part of the broader class of biomimetic methods. This review provides a comprehensive overview of AI-based methods currently used for DDI prediction. These include classical ML algorithms, such as logistic regression (LR) and support vector machines (SVMs); advanced DL models, such as deep neural networks (DNNs) and long short-term memory networks (LSTMs); graph-based models, such as graph convolutional networks (GCNs) and graph attention networks (GATs); and ensemble techniques. The use of knowledge graphs and transformers to capture relations and meaningful data about drugs is also investigated. Additionally, emerging biomimetic approaches offer promising directions for the future in designing AI models that can emulate the complexity of pharmacological interactions. These upgrades include using genetic algorithms with LR and SVM, neuroevaluation (brain-inspired model optimization) to improve DNN and LSTM architectures, ant-colony-inspired path exploration with GCN and GAT, and immune-inspired attention mechanisms in transformer models. This manuscript reviews the typical types of data employed in DDI (pDDI) prediction studies and the evaluation methods employed, discussing the pros and cons of each. There are useful approaches outlined that reveal important points that require further research and suggest ways to improve the accuracy, usability, and understanding of DDI prediction models. Full article

Background: Aging is traditionally characterized by progressive structural and cognitive decline; however, increasing evidence shows that the aging brain retains a remarkable capacity for reorganization. This adaptive neuroplasticity supports cognitive resilience—defined as the ability to maintain efficient cognitive performance despite age-related neural vulnerability. Objective: To synthesize current molecular, cellular, neuroimaging, and electrophysiological neuromarkers that characterize adaptive neuroplasticity and to examine how these mechanisms contribute to cognitive resilience across aging. Methods: This narrative review integrates findings from molecular neuroscience, multimodal neuroimaging (fMRI, DTI, PET), electrophysiology (EEG, MEG, TMS), and behavioral research to outline multiscale biomarkers associated with compensatory and efficient neural reorganization in older adults. Results: Adaptive neuroplasticity emerges from the coordinated interaction of neurotrophic signaling (BDNF, CREB, IGF-1), glial modulation (astrocytic lactate metabolism, regulated microglial activity), synaptic remodeling, and neurovascular support (VEGF, nitric oxide). Multimodal neuromarkers—including preserved frontoparietal connectivity, DMN–FPCN coupling, synaptic density (SV2A-PET), theta–gamma coherence, and LTP-like excitability—consistently correlate with resilience in executive functions, memory, and processing speed. Behavioral enrichment, physical activity, and cognitive training further enhance these biomarkers, creating a bidirectional loop between experience and neural adaptability. Conclusions: Adaptive neuroplasticity represents a fundamental mechanism through which older adults maintain cognitive function despite biological aging. Integrating molecular, imaging, electrophysiological, and behavioral neuromarkers provides a comprehensive framework to identify resilience trajectories and to guide personalized interventions aimed at preserving cognition. Understanding these multilevel adaptive mechanisms reframes aging not as passive decline but as a dynamic continuum of biological compensation and cognitive preservation. Full article

Classical integer-order chaotic maps usually exhibit chaotic degradation under prolonged iterations or finite-precision computation, which may compromise the reliability of chaos-based algorithms. Fractional difference chaotic systems with memory effects offer a promising alternative; however, existing studies rarely provide a systematic and quantitative understanding of how the nonlinear gain parameter, memory strength, and initial condition collectively influence the emergence and robustness of complex dynamics under finite-time iterations. It should be noted that memory effects do not inherently guarantee robust chaotic behavior under finite-precision computation, and appropriate parameter and initial-condition selection remains essential. In this paper, we conduct a systematic numerical dynamical analysis of a logistic-type fractional difference system with power-law memory by leveraging bifurcation diagrams and Lyapunov exponent mappings. Rather than aiming to select optimal parameter points, we propose a quantitative composite chaos evaluation (CCE) framework to identify admissible parameter intervals within which robust finite-time chaotic dynamics can be consistently sustained. Numerical results demonstrate the effectiveness and reliability of the proposed framework, which may facilitate future applications in chaos-enhanced optimization, nonlinear control, and secure communication. Full article

Tin selenide (SnSe) is a sustainable, lead-free IV–VI semiconductor whose layered orthorhombic crystal structure induces pronounced electronic and phononic anisotropy, enabling diverse energy-related functionalities. This review systematically summarizes recent progress in understanding the structure–property–processing relationships that govern SnSe performance in thermoelectric and optoelectronic applications. Key crystallographic characteristics are first discussed, including the temperature-driven Pnma–Cmcm phase transition, anisotropic band and valley structures, and phonon transport mechanisms that lead to intrinsically low lattice thermal conductivity below 0.5 W m⁻¹ K⁻¹ and tunable carrier transport. Subsequently, major synthesis strategies are critically compared, spanning Bridgman and vertical-gradient single-crystal growth, spark plasma sintering and hot pressing of polycrystals, as well as vapor- and solution-based thin-film fabrication, with emphasis on process windows, stoichiometry control, defect chemistry, and microstructure engineering. For thermoelectric applications, directional and temperature-dependent transport behaviors are analyzed, highlighting record thermoelectric performance in single-crystal SnSe at hi. We analyze directional and temperature-dependent transport, highlighting record thermoelectric figure of merit values exceeding 2.6 along the b-axis in single-crystal SnSe at ~900 K, as well as recent progress in polycrystalline and thin-film systems through alkali/coinage-metal doping (Ag, Na, Cu), isovalent and heterovalent substitution (Zn, S), and hierarchical microstructural design. For optoelectronic applications, optical properties, carrier dynamics, and photoresponse characteristics are summarized, underscoring high absorption coefficients exceeding 10⁴ cm⁻¹ and bandgap tunability across the visible to near-infrared range, together with interface engineering strategies for thin-film photovoltaics and broadband photodetectors. Emerging applications beyond energy conversion, including phase-change memory and electrochemical energy storage, are also reviewed. Finally, key challenges related to selenium volatility, performance reproducibility, long-term stability, and scalable manufacturing are identified. Overall, this review provides a process-oriented and application-driven framework to guide the rational design, synthesis optimization, and device integration of SnSe-based materials. Full article

Mild traumatic brain injury (mTBI) disrupts hippocampal network function through coordinated alterations in glial reactivity, synaptic integrity, and adult neurogenesis. Effective therapeutic approaches targeting these interconnected processes remain limited. Lipid-derived molecules capable of modulating these mTBI-induced disturbances are emerging as promising neuroprotective candidates. Here, we investigated the effects of N-stearidonylethanolamine (SDEA), an ω-3 ethanolamide, in a mouse model of mTBI. SDEA treatment attenuated astrocytic reactivity, restored Arc expression, and improved dendritic spine density and morphology in the CA1 hippocampal area. In the dentate gyrus, mTBI reduced Ki-67-indexed proliferation while leaving DCX-positive immature neurons unchanged, and SDEA partially rescued proliferative activity. These effects were accompanied by improvements in anxiety-like behavior and working-memory performance. Together, these findings demonstrate that SDEA modulates several key components of the glia-synapse-neurogenesis axis and supports functional recovery of hippocampal circuits following mTBI. These results suggest that ω-3 ethanolamides may represent promising candidates for multi-target therapeutic strategies in mTBI. Full article

With the rapid development of mobile edge computing (MEC) and unmanned aerial vehicle (UAV) communication networks, UAV-assisted edge computing has emerged as a promising paradigm for low-latency and energy-efficient computation. However, the time-varying nature of air-to-ground channels and the coupling between UAV trajectories and computation offloading decisions significantly increase system complexity. To address these challenges, this paper proposes an Adaptive UAV Edge Intelligence Framework (AUEIF) for joint UAV computation offloading and trajectory optimization under dynamic channels. Specifically, a dynamic graph-based system model is constructed to characterize the spatio-temporal correlation between UAV motion and channel variations. A hierarchical reinforcement learning-based optimization framework is developed, in which a high-level actor–critic module is responsible for generating coarse-grained UAV flight trajectories, while a low-level deep Q-network performs fine-grained optimization of task offloading ratios and computational resource allocation in real time. In addition, an adaptive channel prediction module leveraging long short-term memory (LSTM) networks is integrated to model temporal channel state transitions and to assist policy learning and updates. Extensive simulation results demonstrate that the proposed AUEIF achieves significant improvements in end-to-end latency, energy efficiency, and overall system stability compared with conventional deep reinforcement learning approaches and heuristic-based schemes while exhibiting strong robustness against dynamic and fluctuating wireless channel conditions. Full article