Search Results (579)

Many confined-space accidents have happened in wastewater treatment tanks, mainly caused by hazard gases. To identify the factors affecting the distribution of toxic and harmful gases in wastewater treatment tanks, in this study, we collected data on confined-space accidents occurring in wastewater treatment tanks in China and analyzed accident types, the substances that caused the accidents and the purpose of entry. We carried out field tests to detect the concentrations of oxygen, hydrogen sulfide, combustible gas and carbon monoxide in 222 wastewater treatment tanks from 28 industrial enterprises and investigated the influence of wastewater treatment tank type, cover type and industry type on gas distribution. Through continuous monitoring, the concentrations of hydrogen sulfide and carbon monoxide in the regulating tanks of two industrial enterprises were monitored for a few days. The mechanism of harmful gas generation and control approaches were explored and analyzed. The results showed that more than 90% of confined-space accidents in wastewater treatment tanks were poisoning accidents, and the levels of harmful gas in wastewater collection tanks, regulating tanks, hydrolysis acidification tanks, sedimentation tanks and sludge tanks were high, qualifying them as high-risk wastewater treatment tanks prone to accidents. Without disturbance, there is basically no harmful gas in wastewater treatment tanks with completely uncovered tops. In addition, the concentration of toxic and hazardous gases in wastewater treatment tanks is not always stable, instead fluctuating greatly with time. The main purposes of this study are to identify the factors affecting the concentration of toxic and harmful gases in wastewater treatment tanks and to assess the risks of using wastewater treatment tanks. Full article

Text-to-image (T2I) generation, a core component of generative artificial intelligence(AI), is increasingly important for creative industries and human–computer interaction. Despite impressive progress in realism and diversity, diffusion models still exhibit critical security blind spots particularly in the Transformer key-value mapping mechanism that underpins cross-modal alignment. Existing backdoor attacks often rely on large-scale data poisoning or extensive fine-tuning, leading to low efficiency and limited stealth. To address these challenges, we propose two efficient backdoor attack methods AttnBackdoor and SemBackdoor grounded in the Transformer’s key-value storage principle. AttnBackdoor injects precise mappings between trigger prompts and target instances by fine-tuning the key-value projection matrices in U-Net cross-attention layers (≈5% of parameters). SemBackdoor establishes semantic-level mappings by editing the text encoder’s MLP projection matrix (≈0.3% of parameters). Both approaches achieve high attack success rates (>90%), with SemBackdoor reaching 98.6% and AttnBackdoor 97.2%. They also reduce parameter updates and training time by 1–2 orders of magnitude compared to prior work while preserving benign generation quality. Our findings reveal dual vulnerabilities at visual and semantic levels and provide a foundation for developing next generation defenses for secure generative AI. Full article

Converter gas is highly susceptible to leakage during the recovery and utilization process, which threatens personnel security and sustainable industrial development. To address this issue, a numerical model was established based on an actual converter fan room, and the accuracy of the simulation was verified through comparison with actual measurement data. In this study, the gas leakage flow field, diffusion trajectories, and hazard zone gradations were analyzed. Results showed that the gas contamination was significantly influenced by the leakage direction, leakage location, and structural boundary. The jet dominated the gas dispersion near the leakage source, with similar initial diffusion characteristics across different scenarios. Then, the diffusion velocity decayed rapidly within a distance of 0.6 m. Obstacles can significantly promote vortex formation, restrict the gas dispersion path, and reduce the extent of the hazardous area. In addition, it can be found that the far-field velocity under downward leakage was the highest, presenting the greatest risk of poisoning. At a height of 1.6 m, a lethal zone with a radius of 0.8 m was formed directly beneath the leakage hole. This work can guide the optimization of the monitoring program and emergency planning for converter gas leakage accidents. Full article

Non-intrusive load monitoring (NILM), as a key technology in smart-grid advanced metering infrastructure, aims to disaggregate mains power from smart meters into individual load-level power consumption. Traditional NILM methods require centralizing sensitive measurement data from users, which poses significant privacy risks. Federated learning (FL) enables collaborative training without centralized measurement data, effectively preserving privacy. However, FL-based NILM systems face serious threats from attacks such as model inversion and parameter poisoning, and rely heavily on the availability of a central server, whose failure may compromise measurement robustness. This paper proposes a hybrid FL framework that dynamically switches between centralized FL (CFL) and decentralized FL (DFL) modes, enhancing measurement privacy and system robustness simultaneously. In CFL mode, layer-sensitive pruning and robust parameter aggregation methods are developed to defend against model inversion and parameter poisoning attacks; even with 30% malicious clients, the proposed defense limits the increases in key error metrics to under 15.4%. In DFL mode, a graph attention network (GAT)-based dynamic topology adapts to mitigate topology poisoning attacks, achieving an approximately 17.2% reduction in MAE after an attack and rapidly restoring model performance. Extensive evaluations using public datasets demonstrate that the proposed framework significantly enhances the robustness of smart-grid measurements and effectively safeguards measurement privacy. Full article

While Federated Learning (FL) and Split Learning (SL) aim to uphold data confidentiality by localized training, they remain susceptible to adversarial threats such as model poisoning and sophisticated inference attacks. To mitigate these vulnerabilities, we propose $\mathbf{SplitML}$, a secure and privacy-preserving framework for Federated Split Learning (FSL). By integrating $IND-CP{A}^D$ secure Fully Homomorphic Encryption (FHE) with Differential Privacy (DP), $\mathbf{SplitML}$ establishes a defense-in-depth strategy that minimizes information leakage and thwarts reconstructive inference attempts. The framework accommodates heterogeneous model architectures by allowing clients to collaboratively train only the common top layers while keeping their bottom layers exclusive to each participant. This partitioning strategy ensures that the layers closest to the sensitive input data are never exposed to the centralized server. During the training phase, participants utilize multi-key CKKS FHE to facilitate secure weight aggregation, which ensures that no single entity can access individual updates in plaintext. For collaborative inference, clients exchange activations protected by single-key CKKS FHE to achieve a consensus derived from Total Labels (TL) or Total Predictions (TP). This consensus mechanism enhances decision reliability by aggregating decentralized insights while obfuscating soft-label confidence scores that could be exploited by attackers. Our empirical evaluation demonstrates that $\mathbf{SplitML}$ provides substantial defense against Membership Inference (MI) attacks, reduces temporal training costs compared to standard encrypted FL, and improves inference precision via its consensus mechanism, all while maintaining a negligible impact on federation overhead. Full article

Saxitoxin (STX) is one of the most potent marine neurotoxins, produced by several species of freshwater cyanobacteria and marine dinoflagellates. Although omics-based approaches have advanced our understanding of STX biosynthesis in recent decades, the origin, regulation, and ecological drivers of STX in dinoflagellates remain poorly resolved. Specifically, dinoflagellate STX biosynthetic genes (sxt) are extremely fragmented, inconsistently expressed, and unevenly distributed between toxic and non-toxic taxa. Environmental studies further report inconsistent relationships between abiotic factors and STX production, suggesting regulation across multiple genomic, transcriptional, post-transcriptional, and epigenetic levels. These gaps prevent a comprehensive understanding of STX biosynthesis in dinoflagellates and limit the development of accurate predictive models for harmful algal blooms (HABs) and paralytic shellfish poisoning (PSP). Artificial intelligence (AI), including machine learning and deep learning, offers new opportunities in ecological pattern recognition, molecular annotation, and data-driven prediction. This review explores the current state of knowledge and persistent knowledge gaps in dinoflagellate STX research and proposes an AI-integrated multi-omics framework highlighting recommended models for sxt gene identification (e.g., DeepFRI, ProtTrans, ESM-2), evolutionary reconstruction (e.g., PhyloGAN, GNN, PhyloVAE, NeuralNJ), molecular regulation (e.g., MOFA+, LSTM, GRU, DeepMF), and toxin prediction (e.g., XGBoost, LightGBM, LSTM, ConvLSTM). By integrating AI with diverse biological datasets, this novel framework outlines how AI can advance fundamental understanding of STX biosynthesis and inform future applications in HAB monitoring, seafood safety, and PSP risk management in aquaculture and fisheries. Full article

Salt stress causes harm to plants through multiple aspects, such as osmotic pressure, ion poisoning, nutrient imbalance, and oxidative damage. Zinc finger proteins harboring two B-box domains, known as double B-box (DBB) proteins, constitute the DBB family. While DBB genes have been implicated in regulating circadian rhythms and stress responses in various plant species, their functions in cotton remain largely unexplored. The present study characterized the DBB gene family across the genomes of Gossypium hirsutum L., Gossypium raimondii L., and Gossypium arboreum L., revealing a complement of 58 members. These DBB genes were assigned to three separate clades based on phylogenetic analysis. Members possessing close phylogenetic relationships have similar conserved protein motifs and gene structures. All DBB proteins were predicted to be nuclear-localized, consistent with their roles as transcription factors. Furthermore, the presence of multiple cis-acting elements related to light, hormone, and stress responses in the promoters implies that GhDBBs are integral to cotton’s environmental stress adaptation. Expression pattern analysis indicated that the expression of GhDBB genes was associated with the plant’s response to multiple abiotic stresses, such as salt, drought, heat (37 °C), and cold (4 °C). The reliability of the expression data was confirmed by qPCR analysis of eight selected GhDBBs. Under 200 mM NaCl, Arabidopsis plants overexpressing GhDBB22 displayed longer roots and healthier true leaves than the wild-type controls. Conversely, VIGS-mediated silencing of GhDBB22 in G. hirsutum led to significantly reduced salt tolerance, accompanied by exacerbated oxidative damage. Taken together, the findings from our integrated genomic and functional analyses provide a foundational understanding of the molecular mechanisms through which proteins encoded by DBB genes are involved in the plant’s response to salt stress. Full article

Semantic communication systems leverage deep neural networks to extract and transmit essential information, achieving superior performance in bandwidth-constrained wireless environments. However, their vulnerability to backdoor attacks poses critical security threats, where adversaries can inject malicious triggers during training to manipulate system behavior. This paper introduces Selective Communication Unlearning (SCU), a novel defense mechanism based on Variational Information Bottleneck (VIB) principles. SCU employs a two-stage approach: (1) joint unlearning to remove backdoor knowledge from both encoder and decoder while preserving legitimate data representations, and (2) contrastive compensation to maximize feature separation between poisoned and clean samples. Extensive experiments on the RML2016.10a wireless signal dataset demonstrate that SCU achieves 629.5 ± 191.2% backdoor mitigation (5-seed average; 95% CI: [364.1%, 895.0%]), with peak performance of 1486% under optimal conditions, while maintaining only 11.5% clean performance degradation. This represents an order-of-magnitude improvement over detection-based defenses and fundamentally outperforms existing unlearning approaches that achieve near-zero or negative mitigation. We validate SCU across seven signal processing domains, four adaptive backdoor types, and varying SNR conditions, demonstrating unprecedented robustness and generalizability. The framework achieves a 243 s unlearning time, making it practical for resource-constrained edge deployments in 6G networks. Full article

In recent years, the convergence of advanced technologies has enabled real-time data access and sharing across diverse devices and networks, significantly amplifying cybersecurity risks. For organizations with digital infrastructures, network security is crucial for mitigating potential cyber-attacks. They establish security policies to protect systems and data, but employees may intentionally or unintentionally bypass these policies, rendering the network vulnerable to internal and external threats. Detecting these policy violations is challenging, requiring frequent manual system checks for compliance. This paper addresses key challenges in safeguarding digital assets against evolving threats, including rogue access points, man-in-the-middle attacks, denial-of-service (DoS) incidents, unpatched vulnerabilities, and AI-driven automated exploits. We propose a Blockchain-Enhanced Network Scanning and Monitoring (BENSAM) Framework, a multi-layered system that integrates advanced network scanning with a structured database for asset management, policy-driven vulnerability detection, and remediation planning. Key enhancements include device profiling, user activity monitoring, network forensics, intrusion detection capabilities, and multi-format report generation. By incorporating blockchain technology, and leveraging immutable ledgers and smart contracts, the framework ensures tamper-proof audit trails, decentralized verification of policy compliance, and automated real-time responses to violations such as alerts; actual device isolation is performed by external controllers like SDN or NAC systems. The research provides a detailed literature review on blockchain applications in domains like IoT, healthcare, and vehicular networks. A working prototype of the proposed BENSAM framework was developed that demonstrates end-to-end network scanning, device profiling, traffic monitoring, policy enforcement, and blockchain-based immutable logging. This implementation is publicly released and is available on GitHub. It analyzes common network vulnerabilities (e.g., open ports, remote access, and disabled firewalls), attacks (including spoofing, flooding, and DDoS), and outlines policy enforcement methods. Moreover, the framework anticipates emerging challenges from AI-driven attacks such as adversarial evasion, data poisoning, and transformer-based threats, positioning the system for the future integration of adaptive mechanisms to counter these advanced intrusions. This blockchain-enhanced approach streamlines security analysis, extends the framework for AI threat detection with improved accuracy, and reduces administrative overhead by integrating multiple security tools into a cohesive, trustworthy, reliable solution. Full article

With the accelerated expansion of the Internet of Things (IoT), massive distributed and heterogeneous devices are increasingly exposed to severe security threats. Traditional centralized intrusion detection systems (IDS) suffer from significant limitations in terms of privacy preservation and communication overhead. Federated Learning (FL) offers an effective paradigm for building the next generation of distributed IDS; however, it remains vulnerable to poisoning attacks in open environments, and existing client selection strategies generally lack robustness and security awareness. To address these challenges, this paper proposes an Uncertainty-Guided Evolutionary Game-Theoretic (UEGT) Client Selection mechanism. Built upon evolutionary game theory, UEGT integrates Shapley value, gradient similarity, and data quality to construct a multidimensional payoff function and employs a replicator dynamics mechanism to adaptively optimize client participation probabilities. Furthermore, uncertainty modeling is introduced to enhance strategic exploration and improve the identification accuracy of potentially high-value clients. Experimental results under adversarial scenarios demonstrate that UEGT maintains stable convergence even under a high fraction of malicious participating clients, achieving an average accuracy exceeding 89%, which outperforms several mainstream client selection and robust aggregation methods. Full article

Intensive Care Units (ICUs) face unprecedented challenges in resource allocation, particularly during health crises in which algorithmic systems may be exposed to adversarial manipulation. A transformer-based expert system, ICU-Transformer, is presented to optimize resource allocation across 200 ICUs in Physionet while maintaining robustness against data poisoning attacks. The framework incorporates a Robust Multi-Head Attention mechanism that achieves an AUC-ROC of 0.891 in mortality prediction under 20% data contamination, outperforming conventional baselines. The system is trained and evaluated using data from the MIMIC-IV and eICU Collaborative Research Database and is deployed to manage more than 50,000 ICU admissions annually. A Resource Optimization Engine (ROE) is introduced to dynamically allocate ventilators, Extracorporeal Membrane Oxygenation (ECMO) machines, and specialized clinical staff based on predicted deterioration risk, resulting in an 18% reduction in preventable deaths. A Surge Capacity Planner (SCP) is further employed to simulate disaster scenarios and optimize cross-hospital resource distribution. Deployment across the Physionet ICU Network demonstrates improvements, including a 2.1-day reduction in average ICU bed turnover time, a 31% decrease in unnecessary admissions, and an estimated USD 142 million in annual operational savings. During the observation period, 234 algorithmic manipulation attempts were detected, with targeted disparities identified and mitigated through enhanced auditing protocols. Full article

Machine learning (ML) has become a cornerstone of critical applications, but its vulnerability to data poisoning attacks threatens system reliability and trustworthiness. Prior studies have begun to investigate the impact of data poisoning and proposed various defense or evaluation methods; however, most efforts remain limited to quantifying performance degradation, with little systematic comparison of internal behaviors across model architectures under attack and insufficient attention to interpretability for revealing model vulnerabilities. To tackle this issue, we build a reproducible evaluation pipeline and emphasize the importance of integrating robustness with interpretability in the design of secure and trustworthy ML systems. To be specific, we propose a unified poisoning evaluation framework that systematically compares traditional ML models, deep neural networks, and large language models under three representative attack strategies including label flipping, random corruption, and adversarial insertion, at escalating severity levels of 30%, 50%, and 75%, and integrate LIME-based explanations to trace the evolution of model reasoning. Experimental results demonstrate that traditional models collapse rapidly under label noise, whereas Bayesian LSTM hybrids and large language models maintain stronger resilience. Further interpretability analysis uncovers attribution failure patterns, such as over-reliance on neutral tokens or misinterpretation of adversarial cues, providing insights beyond accuracy metrics. Full article

The integration of artificial intelligence into remote patient monitoring (RPM) offers significant benefits for proactive and continuous healthcare, but also raises critical concerns regarding privacy, integrity, and robustness. Federated Learning (FL) provides a decentralized approach to model training that preserves data locality, yet most existing solutions address only isolated security aspects and lack contextual adaptability for clinical use. This paper presents MedGuard-FL, a context-aware FL framework tailored to e-healthcare environments. Spanning device, edge, and cloud layers, it integrates encryption, adaptive differential privacy, anomaly detection, and Byzantine-resilient aggregation. At its core, a policy engine dynamically adjusts privacy and robustness parameters based on the patient’s status and the system’s risk. Evaluations on real-world clinical datasets show MedGuard-FL maintains high model accuracy while neutralizing various adversarial attacks (e.g., label-flip, poisoning, backdoor, membership inference), all with manageable latency. Compared to static defenses, it offers improved trade-offs between privacy, utility, and responsiveness. Additional edge-level privacy analyses confirm its resilience, with attack effectiveness near random. By embedding clinical risk awareness into adaptive defense mechanisms, MedGuard-FL lays a foundation for secure, real-time federated intelligence in RPM. Full article

Background/Objectives: The delayed effects of organophosphate poisoning may manifest years after exposure, often masked by age-related diseases. The aim of this retrospective cohort study was to identify the biochemical “trace” that could remain in patients decades after poisoning. We determined a wide range of biochemical parameters, along with the spectrum of esterified and non-esterified fatty acids (EFAs and NEFAs, respectively), in the blood plasma of a cohort of elderly patients diagnosed with occupational pathology (OP) due to (sub)chronic exposure to organophosphates in the 1980s. Methods: Elderly patients with and without a history of exposure to organophosphates were retrospectively divided into two groups: controls (n = 59, aged 73 ± 4, men 29% and women 71%) and those with OP (n = 84, aged 74 ± 4, men 29% and women 71%). The period of neurological examination and blood sampling for subsequent analysis was from mid-2022 to the end of 2023. Determination of the content of biomarkers of metabolic syndrome, NEFAs, and EFAs in blood plasma was performed by HPLC-MS/MS and GC-MS. Results: The medical histories of the examined elderly individuals with OP and the aged control group included common age-related diseases. However, patients with OP more often had hepatitis, gastrointestinal diseases, polyneuropathy, and an increased BMI. Analysis of metabolic biomarkers revealed, in the OP group, a decrease in the concentrations of 3-hydroxybutyrate (p < 0.05), 2-hydroxybutyrate (p < 0.0001), and acetyl-L-carnitine (p < 0.001) and the activity of butyrylcholinesterase (BChE) (p < 0.05), but an increase in the esterase activity of albumin (p < 0.05). Correlation analysis revealed significant relationships between albumin esterase activity and arachidonic acid concentrations in the OP group (0.64, p < 0.0001). A study of a wide range of fatty acids in patients with OP revealed reciprocal relationships between EFAs and NEFAs. A statistically significant decrease in concentration was shown for esters of margaric, stearic, eicosadienoic, eicosatrienoic, arachidonic, eicosapentaenoic, and docosahexaenoic fatty acids. A statistically significant increase in concentration was shown for non-esterified heptadecenoic, eicosapentaenoic, eicosatrienoic, docosahexaenoic, γ-linolenic, myristic, eicosenoic, arachidonic, eicosadienoic, oleic, linoleic, palmitic, linoelaidic, stearic, palmitoleic, pentadecanoic, and margaric acids. Decreases in the ratios of omega-3 to other unsaturated fatty acids were observed only for the esterified forms. Conclusions: The data obtained allow us to consider an increased level of NEFAs as one of the main cytotoxic factors for the vascular endothelium. Modification of albumin properties and decreased bioavailability of docosahexaenoic acid could be molecular links that cause specific manifestations of organophosphate-induced pathology at late stages after exposure. Full article

Toxic Shock Syndrome Toxin-1 (TSST-1) is produced by Staphylococcus aureus strains encoded by the tst gene. Toxic shock syndrome (TSS) is a severe disease caused by TSST-1 toxin and associated with staphylococcal food poisoning (SFP). The aim of the present review was to present data on the occurrence of S. aureus TSST-1 in foods published in various countries. PCR-based assays are most frequently used for the detection of S. aureus TSST-1 in foods. S. aureus TSST-1 is predominantly detected in foods of animal origin. The highest occurrence has been observed in mastitic ruminants’ milk, indicating that mastitis is a risk of milk contamination with the pathogen. High occurrence rates of S. aureus TSST-1 have also been identified in raw milk and artisanal cheeses. Various occurrence levels have also been reported in beef, pork, lamb, and chicken meat. Low occurrence levels have also been reported for fish or other seafood products. The tst gene was also found in combination with other toxigenic genes in S. aureus TSST-1 isolates (e.g., MRSA or Panton-Valentine Leukocidin, PVL). Monitoring S. aureus TSST-1 in food is important for public health because food can be a vehicle for transmitting the antibiotic-resistant pathogen to humans. Full article