Search Results (6,206)

In today’s fast-changing digital environment, cyber-physical systems face escalating security challenges due to increasingly sophisticated cyberattacks. Artificial Intelligence (AI) has emerged as a powerful enabler of modern cyberattack detection, offering scalable, accurate, and adaptive solutions to counter dynamic threats. This paper provides a comprehensive review of recent advancements in AI-based cyberattack detection, focusing on Machine Learning (ML), Deep Learning (DL), Reinforcement Learning (RL), Federated Learning (FL), and emerging techniques such as generative AI, neuro-symbolic AI, swarm intelligence, lightweight AI, and quantum Computing. We evaluate the strengths and limitations of these approaches, highlighting their performance on benchmark datasets. The review discusses traditional signature-based Intrusion Detection Systems (IDS) and their limitations against novel attack patterns, contrasted with AI-driven anomaly-based and hybrid detection methods that improve detection rates for unknown and zero-day attacks. Key challenges, including computational costs, data quality, privacy concerns, and model interpretability, are analysed alongside the role of Explainable AI (XAI) in enhancing trust and transparency. The impact of computational resources, dataset representativeness, and evaluation metrics on AI model performance is also explored. Furthermore, we investigate the potential of lightweight AI for resource-constrained environments like IoT and edge devices, and quantum computing’s role in advancing detection efficiency and cryptographic security. The paper also draws attention to future research directions, particularly the development of up-to-date datasets, integration of hybrid quantum–classical models, and optimisation of asynchronous FL protocols to address evolving cybersecurity challenges. This study aims to inspire innovation in AI-driven cyberattack detection, fostering robust, interpretable, and efficient solutions for securing complex digital environments. Full article

Background: Microplastics and nanoplastics (MNPs) are ubiquitous environmental contaminants from plastic degradation, leading to human exposure through ingestion, inhalation, and dermal contact. While emerging evidence suggests potential health effects, comprehensive human-specific data remain limited. Objective: To systematically review evidence on MNP exposure and health impacts across human organ systems. Methods: Following PRISMA guidelines, we searched Embase, Environment Complete, MEDLINE, and Scopus for peer-reviewed English-language studies published between 2020 and 2025 that reported MNP exposure in adult human populations and addressed at least one organ system. Thirty studies met inclusion criteria, and all clinical studies were assessed for risk of bias using the Newcastle–Ottawa Scale (NOS) Results: Clinical studies consistently detected MNPs in human blood, thrombi, feces, and respiratory and reproductive tissues. Higher MNP burdens correlated with increased disease severity across cardiovascular, gastrointestinal, respiratory, musculoskeletal, and reproductive systems. In vitro studies using human-derived cell lines demonstrated that MNPs penetrate cells and disrupt cellular processes, inducing oxidative stress, cytotoxicity, mitochondrial dysfunction, inflammation, DNA damage, and apoptosis. Toxic effects were size-, polymer-, and concentration-dependent, with smaller particles exhibiting greater cellular uptake and toxicity. Conclusions: Human MNP exposure is widespread and associated with adverse biological effects across multiple organ systems. Further interdisciplinary research is needed to establish causal relationships and inform risk assessment and regulatory frameworks for plastic-associated contaminants. Other: This research received no external funding. The research protocol was registered with PROSPERO (Registration ID number CRD420261284559). Full article

Background/Objectives: Chronic hepatitis B (CHB) and type 2 diabetes mellitus (T2DM) frequently coexist. This study aimed to investigate the impact of T2DM and glycemic control on antiviral efficacy in CHB patients. Methods: This single-center, retrospective cohort study included treatment-naïve CHB patients who initiated nucleos(t)ide analogue (NA) therapy between January 2019 and January 2024. The primary endpoint was a complete virological response (CVR), defined as achieving HBV DNA levels below 20 IU/mL after 48 weeks of treatment. Results: The CHB + T2DM group (n = 81) demonstrated a significantly lower CVR rate than the CHB group (n = 106) (26.0% vs. 41.2%, p = 0.038). Multivariate analysis identified T2DM as an independent negative predictor of a CVR (OR = 0.400, 95% CI: 0.196–0.815, p = 0.012). Within the CHB + T2DM subgroup, adequate glycemic control (HbA1c < 7%) was associated with a higher CVR (38.7% vs. 16.7%, p = 0.034). Patients newly diagnosed with diabetes at enrollment showed a higher rate of HBeAg loss than those with pre-existing diabetes (57.1% vs. 10.0%, p = 0.036). Regarding antiviral regimens, entecavir-treated CHB + T2DM patients had a lower CVR than CHB controls (18.8% vs. 46.2%, p = 0.015). Furthermore, tenofovir-based regimens showed a more favorable antiviral trend than entecavir in CHB patients with T2DM. Conclusions: Comorbid T2DM was an independent risk factor for impaired antiviral efficacy in CHB patients. Optimal glycemic control may improve virological outcomes. These findings suggest that the early diagnosis and management of T2DM could enhance antiviral treatment efficacy in CHB patients. Full article

Operational logs are a central information source for monitoring and diagnosing complex information systems, yet the effect of log-sequence representation on anomaly detection remains underexplored. This paper investigates three families of sequence embeddings, E1 (template-ID lookup), E2 (semantic), and E3 (hybrid), for log-based anomaly detection. Each embedding is paired with CNN, LSTM, and Transformer heads under a unified training protocol. We conduct controlled experiments on diverse public corpora to assess in-domain and cross-dataset generalization. We report PR–AUC (primary), AUROC, F1, and precision at recall $\ge 0.9$, with 95% bootstrap confidence intervals. Beyond accuracy, we analyze the impact of sequence length, parser choice, and out-of-vocabulary (OOV) rates at both token and template levels within and across datasets. The results suggest that representation choice can meaningfully influence detection performance, particularly under distribution shift. Open-vocabulary semantic and hybrid embeddings can improve robustness to OOV effects, but transfer gains are inconsistent, and degradation often persists under strict cross-dataset transfer. Full article

Malaria is a major global health concern caused by Plasmodium parasites, among which Plasmodium falciparum is responsible for the most severe and fatal cases. The emergence of drug resistance to existing antimalarial therapies necessitates the discovery of novel molecular targets and chemically distinct inhibitors. Current study employed an integrated in silico drug discovery pipeline combining high-throughput structure-based virtual screening of 1549 deep-sea marine microbial metabolites with MM-GBSA binding free-energy estimation, QikProp-based ADME/Tox profiling, and 100 ns molecular dynamics (MD) simulations to link rapid screening with dynamic verification of binding stability. Molecular docking against Plasmodium falciparum dihydroorotate dehydrogenase (PfDHODH; PDB ID: 7KZ4) yielded five top-ranked compounds with Glide scores ranging from −12.02 to −10.61 kcal·mol⁻¹, which is higher than the Primaquine (−6.920 kcal·mol⁻¹; a clinically approved antimalarial reference compound). MM-GBSA analysis further refined hit selection, producing binding free energies (ΔG_bind) between −63.28 and −31.37 kcal·mol⁻¹. The selected lead compounds included (±)-puniceusine P, aspergilol F, tersaphilone C, 4-carbglyceryl-3,3′-dihydroxy-5,5′-dimethyldiphenyl ether, and 15-O-methyl ML-236A. The top hits were subjected to 100 ns MD simulations in Desmond, demonstrating stable protein–ligand complexes, particularly for (±)-puniceusine P and 15-O-methyl ML-236A (protein backbone root mean square deviation (RMSD; ~0.8–1.0 Å). ADME profiling indicated acceptable predicted physicochemical and pharmacokinetic properties. Overall, these in silico findings highlight deep-sea marine microbial metabolites as promising PfDHODH inhibitor candidates requiring experimental validation. Full article

Intrusion detection in Internet of Things (IoT) environments is challenged by severe class imbalance, evolving attack patterns, and the limited computational resources of edge devices. To address these challenges, this paper proposes a lightweight transfer-learning framework based on a combined architecture of Convolutional Neural Network and Gated Recurrent Unit (CNN–GRU) for IoT intrusion detection. The model is first pretrained on a large-scale source dataset containing mixed benign and attack traffic, then adapted to a smaller and structurally different target dataset using partial finetuning. To enable efficient edge adaptation, early convolutional layers are frozen while only the GRU and classification head are updated on the target domain. A leakage-free, group-aware data preparation strategy with overlapping temporal windows is employed to ensure reliable evaluation. Experimental results demonstrate that the proposed lightweight transfer approach achieves solid macro-level detection performance while reducing training cost compared to full finetuning. Additional analysis using a CPU-based inference proxy shows low latency and a small model footprint. This supports the feasibility of edge deployment. The results confirm that lightweight transfer learning offers an effective balance between detection performance and adaptation efficiency for resource-constrained IoT intrusion detection systems. Full article

Security measures are essential in the automotive industry to detect intrusions in-vehicle networks. However, developing a one-size-fits-all intrusion detection system (IDS) is challenging because each vehicle has a unique data profile. This is due to the complex and dynamic nature of the data generated by vehicles regarding their model, driving style, test environment, and firmware update. To address this issue, a universal IDS has been developed that can be applied to all types of vehicles without the need for customization. Unlike conventional IDSs, the universal IDS can adapt to data distribution shifts caused by changes in driving style, vehicle platform, or firmware updates. In this study, a new hybrid approach has been developed, combining Pearson correlation with deep learning techniques. This approach has been tested using data obtained from four distinct mechanical and electronic vehicles, including Tesla, Sonata, and two Kia models. The data has been combined into two frequency datasets, and wavelet transformation has been employed to convert them into the frequency domain, enhancing generalizability. Additionally, a statistical method based on independent rule-based systems using Pearson correlation has been utilized to improve system performance. The system has been compared with eight different IDSs, three of which utilize the universal approach, while the remaining five are based on conventional techniques. The accuracy of each system has been evaluated through benchmarking, and the results demonstrate that the hybrid system effectively detects intrusions in various vehicle models. Full article

As additive manufacturing (AM) expands into high-end industries, ensuring both technical performance and dimensional accuracy remains a challenge. This paper addresses the challenge of integrating recycled materials into the field of plastic extrusion additive manufacturing technologies by conducting a study on the evaluation and optimization of thermoplastic extrusion parameters to improve the dimensional accuracy of additively manufactured parts from virgin and recycled polyethylene terephthalate glycol (PETG, rPETG) and acrylonitrile styrene acrylate (ASA), both in virgin and recycled form. To carry out the study, 180 three-point bending specimens were additively manufactured on the QIDI Q1 Pro 3D printer by thermoplastic extrusion of PETG, rPETG, ASA, rASA (45 specimens for each type of material), using the following variable parameters: layer height deposited in one pass L_h = (0.10–0.20) mm and filling percentage—I_d = (50–100)%. After manufacturing the specimens, the dimensional characteristics that will be determined by measurement were defined: L—length, WA—width A, HA—height A, WA’—width A’, and HA’—height A’. Dimensional accuracy was assessed through 900 measurements using a DeMeet 400 coordinate measuring machine and analyzing the arithmetic means, dispersions, and mean square deviations. The results of the study confirm the superior dimensional stability of virgin materials (18.77–20.04%) compared to recycled materials. The analysis demonstrates that by optimizing the process parameters, filaments from recycled materials (rPETG and rASA) can achieve acceptable precision, with average deviations of 0.25–0.78% from the nominal dimensions. The present study validates the use of rPETG and rASA as a viable alternative for applications that do not require critical tolerances. Full article

Ziziphus lotus (L.) Lam., an extremophyte shrub native to the Mediterranean basin, yields underexplored fruits as a source of therapeutic agents. This study combined in vitro and in silico approaches to evaluate the antioxidant potential of Z. lotus fruits and predict their potential to inhibit xanthine oxidase (XO), a key enzyme in reactive oxygen species generation and oxidative stress-related pathologies. The ethyl acetate extract from the hydroalcoholic macerate was enriched in total phenolics (281.33 ± 1.5 μg GAE/mg) and flavonoids (127.26 ± 5.89 μg RE/mg) and displayed remarkable effects against the ABTS^•+ radical cation (IC₅₀ = 18.49 ± 1.47 μg/mL) and phenanthroline reducing power (A_0.5 = 8.38 ± 0.69 μg/mL), together with measurable xanthine oxidase inhibition (IC₅₀ = 170.4 ± 5.90 μg/mL). The compounds tentatively identified by full-scan UHPLC-QtoF-HRMS were docked against XO (PDB ID: 3NVY), with phytosphingosine (−8.5 kcal/mol) and rutin (−8.3 kcal/mol) exhibiting the strongest binding affinities, forming favorable predicted interactions with critical catalytic residues, followed by 6‴-feruloylspinosin, 3′,5′-di-C-β-glucopyranosylphloretin and hexadecasphinganine (ranging from −7.8 to −7.6 kcal/mol). Predictive structure–activity relationships were also observed. These results provide insights into the antioxidant potential of Z. lotus phytochemicals and highlight the value of this extremophile plant as sustainable resource for phytotherapy and the management of oxidative stress-related diseases. Full article

The rapid deployment of Internet of Things (IoT) devices has increased exposure to a diverse array of evolving cyberattacks, motivating the need for accurate and interpretable intrusion detection systems (IDS). In this work, we develop an explainable hybrid Convolutional Neural Network–Extreme Gradient Boosting (CNN–XGBoost) framework for multi-class IoT attack classification using the CIC IoT-DIAD 2024 dataset. Network-traffic records are preprocessed and standardized using a scalable, chunk-wise workflow, after which a compact top-k subset of features is selected via Random Forest importance ranking. To reduce selection bias, a leakage-prone feature-ranking strategy is compared with a leakage-aware strategy in which features are ranked using only the training data within each split. Subsequently, a one-dimensional Convolutional Neural Network (CNN) learns a 128-dimensional representation from the selected predictors, and XGBoost performs the final multi-class classification. Under the leakage-aware protocol, the proposed model achieves 0.9324 accuracy with 0.5910 macro-F1. Results indicate that leakage-aware selection provides a more defensible estimate of generalization while maintaining competitive detection performance. Finally, SHapley Additive exPlanations (SHAP) is used to interpret the model’s decisions in the learned latent space. The analysis shows that only a small number of embedding dimensions contribute most of the decision evidence, which can aid analyst triage, although the explanations remain indirect with respect to the original traffic features. Full article

Objective: The objective of this study is to conduct a systematic review of the evidence on the use of remotely administered walking tests (RaWTs) in patients with chronic pulmonary diseases (CPDs) and heart failure (HF), focusing on agreement, reliability, feasibility, and clinical utility as outcomes. Methods: This study followed the Preferred Reporting Items for Systematic Reviews and was registered on the International Prospective Register of Systematic Reviews platform (ID: CRD420251180996). The PubMed, Web of Science, CENTRAL, Scopus, and ACM databases were comprehensively searched from inception up to October 2025. Observational, randomized and non-randomized control studies assessing the agreement, reliability, feasibility, and clinical utility of RaWTs in people with CPDs and HF and reporting quantitative outcomes were eligible. Two reviewers independently conducted study selection, data extraction, and risk of bias assessment using the COSMIN Risk of Bias tool for the reliability studies, the Risk of Bias in Non-Randomized Studies—of Interventions (ROBINS-I) tool for non-randomized studies, and the Quality in Prognosis Studies (QUIPS) tool for the prognostic studies. Results: Eleven studies met the inclusion criteria. Five studies included patients with HF, five with pulmonary hypertension (PH), and one study included candidates for lung transplantation due to advanced CPD. All studies used the 6 min walk test (6MWT); one also included the incremental shuttle walk test. Agreement with face-to-face in-clinic testing (in five studies) is setting-dependent and influenced by the testing setup. Reliability (in eight studies), derived from variable statistical indices in both patient populations, showed that RaWTs are reliable. Adherence and safety were used as the main feasibility indicators. Eight studies concluded that remote assessment is feasible, acceptable, and safe. Clinical utility was examined in only one HF study, showing that remotely administered 6MWT can predict all-cause mortality and HF hospitalization. According to COSMIN, the overall methodological quality of nine studies ranged from very good to inadequate. One study was rated as having a serious risk of bias according to ROBINS-I, and one study as having a high risk of bias according to QUIPS. Conclusions: Although the evidence is limited and heterogeneous, RaWTs demonstrate robust reliability across repeated measurements while agreement with in-clinic testing is context-dependent and strongly influenced by test setup and environmental conditions. RaWTs appear to be acceptable to patients; however, further high-quality studies are needed to confirm these findings and determine the clinical utility of RaWTs on specific clinical outcomes in these populations. Full article

Background: Embodied artificial intelligence (EAI), integrating advanced AI algorithms with robotic platforms capable of sensing, planning, and acting, has emerged as a transformative approach in healthcare delivery. This systematic review synthesizes evidence on robotic perception, decision-making, and clinical impact of EAI systems in healthcare settings. Methods: Following PRISMA 2020 guidelines, we searched PubMed/MEDLINE, Scopus, Web of Science, IEEE Xplore, and ACM Digital Library for studies published between January 2020 and August 2025. Seventeen studies met eligibility criteria, spanning four domains: surgical assistance, rehabilitation, hospital logistics, and telepresence. The protocol was prospectively registered in PROSPERO under ID: CRD420261285936. Results: Perception architectures predominantly employed multimodal sensor fusion, combining vision with force/torque, depth, and physiological signals. Decision-making approaches included imitation learning, reinforcement learning, and hybrid symbolic-neural control. Key findings indicate that surgical robots demonstrated consistency advantages in specific experimental tasks, rehabilitation robotics produced statistically significant improvements (SMD = 0.29) across 396 randomized controlled trials, and both logistics and telepresence systems achieved very high operational success levels. Nonetheless, important barriers remain, including limited external validation, small sample sizes, and insufficient cost-effectiveness data. Conclusions: Future research should prioritize standardized benchmarks, prospective multicenter trials, and patient-centered outcome measures to facilitate clinical translation of EAI technologies. Full article

In contemporary 5G network environments, intrusion detection systems must balance detection accuracy with operational efficiency, as improvements in one dimension are often achieved at the expense of the other. This study addresses this trade-off by proposing a lightweight two-stage intrusion detection architecture that augments a standard decision-tree classifier with a conditional counter-inspection mechanism. At inference time, a global decision tree produces an initial classification for each traffic record, which is selectively validated by a small set of class-biased expert trees trained under controlled minority exposure. Only experts associated with the opposite class of the initial prediction are activated, and decision revision is governed by a unanimous-dissent rule, ensuring conservative and deterministic correction while avoiding over-correction. Experiments conducted on the 5G-NIDD dataset in a binary benign/malicious setting show that the proposed architecture consistently improves upon the standalone decision tree, reducing false negatives from 51 to 27 (−47.1%) and false positives from 48 to 30 (−37.5%), and achieving an F1-score of 0.99981 on a held-out test set. Ablation and paired statistical tests confirm that these gains arise from selective validation and the unanimous-dissent mechanism rather than from uniform ensembling. The complete pipeline operates in the microsecond inference regime per record, evaluates fewer models on average than flat voting strategies, and preserves full interpretability through deterministic decision paths, making it suitable for practical and resource-constrained 5G intrusion detection deployments. Full article

As networks expand in size and complexity, coupled with an exponential increase in intrusions on network and IoT systems, this leads to traditional models failing to capture increasingly intricate correlations among network components accurately. Graph Convolution Networks (GCNs) have recently acquired prominence for their capacity to represent nodes, edges, or entire graphs by aggregating information from adjacent nodes. However, the correlations between nodes and their neighbours, as well as related edges, differ. Assigning higher weights to nodes and edges with high similarity improves model accuracy and expressiveness. In this paper, we propose the GCN-DQN model, which integrates GCN with a multi-head attention mechanism and DQN (Deep Q Network) to adaptively adjust attention weights optimizing its performance in intrusion detection tasks. After extensive experiments using the UNSW NB15 and CIC-IDS2017 dataset, the proposed GCN-DQN outperformed the baseline model in classification accuracy. We also applied LIME and SHAP techniques to provide explainability to our proposed intrusion detection model. Full article

Public transportation is the core of easing urban traffic congestion, reducing pollution and advancing smart city transportation intellectualization. Its refined operation relies heavily on accurate, real-time passenger origin–destination (OD) data. However, traditional manual surveys are costly with low sampling rates, while smart card big data lacks alighting information and has deviations, failing to reflect real travel behaviors and becoming a bottleneck for intelligent public transportation development. To address this, this paper proposes a bus passenger boarding/alighting detection and recognition study based on video images and the YOLO algorithm. Aiming at traditional YOLO’s shortcomings in on-vehicle scenarios (insufficient feature extraction, inefficient feature fusion, slow convergence), the baseline YOLOv8n is improved for bus scenarios’ high-density, high-occlusion and variable-target scales: (1) DAC2f structure (deformable attention + C2f) captures occluded passengers’ core features and suppresses background interference; (2) SWD-PAN enables bidirectional cross-scale feature interaction to adapt to scale differences; and (3) WIoUv3 balances sample weights for small targets and non-standard posture passengers. Experiments show that precision, recall and mAP increase by 3.68%, 5.12% and 6.26%, respectively, meeting real-time requirements. The improved YOLOv8 is deeply integrated with DeepSORT to enhance tracking stability. Tests show that MOTA reaches 31.24% (2.6% higher than YOLOv8n, 16.4% higher than YOLO-X) and MOTP reaches 88.06%, solving trajectory breakage and ID switching. This addresses traditional OD data collection pain points, providing technical support for intelligent public transportation refined management and smart city transportation optimization. Full article