Search Results (31,511)

As organizations embed artificial intelligence (AI) in everyday work, understanding how employees respond has become a pressing question for human resource management. Drawing on the cognitive appraisal theory of stress, this study examines how AI application relates to employees’ proactive learning behavior through two parallel primary appraisals: opportunity perception and threat perception. It also tests proactive personality both as a direct predictor and as a moderator at two theoretically distinct locations: the formation of appraisals and the translation of appraisals into behavior. Using a four-wave panel design among employees from multiple industries, we find that AI application is positively associated with proactive learning behavior overall. This association operates through two opposing pathways: opportunity perception strengthens proactive learning, whereas threat perception suppresses it. Additional analyses indicate that the opportunity pathway is the more robust of the two across longitudinal specifications. The two appraisals are statistically separable and jointly elevated in a subgroup of employees, indicating parallel rather than strictly bipolar appraisal. Proactive personality is positively associated with opportunity perception and proactive learning, negatively associated with threat perception, and moderates the appraisal-formation stage rather than the appraisal-to-behavior stage. These findings extend the cognitive appraisal theory of stress to the AI context and suggest that dispositional proactivity operates mainly by shaping how employees read AI, rather than by changing how they act on an appraisal once it has formed. Full article

The rapid evolution of sophisticated cyber threats has drastically increased the cybersecurity risks in Industrial Internet of Things environments due to the massive interconnection of industrial devices, sensors, programmable logic controllers, gateways, and edge computing infrastructures. Traditional intrusion detection systems are insufficient for modern IIoT networks due to challenges with dynamic attack behaviors, class imbalance, concept drift, and computational limitations of resource-constrained edge devices. Although machine learning and deep learning have significantly improved intrusion detection performance, current studies usually deal with these challenges separately and do not provide a holistic view of adaptive and real-time industrial internet of things (IIoT) cybersecurity. In this survey, we provide a structured narrative review of adaptive intrusion detection techniques, focusing on three emerging research directions, including GAN-based data augmentation, drift-aware learning, and Edge Intelligence. It provides a structured narrative review of machine learning, deep learning, and hybrid IDS models, benchmark IIoT datasets, and representative techniques addressing data imbalance, concept drift, and low-latency edge deployment. The survey further provides a comparative study of existing approaches in terms of detection capability, adaptability, computational efficiency, scalability, and deployment suitability. To fill the gap between those complementary research directions, the survey combines the literature into a unified reference architecture that merges GAN-based data augmentation, drift-aware learning, and Edge Intelligence to enable adaptive, real-time IIoT intrusion detection. Finally, we discuss key research challenges and future opportunities in autonomous, collaborative, and trustworthy IIoT cybersecurity, thus providing a practical roadmap for the development of next-generation intelligent intrusion detection systems. Full article

Strong earthquake-induced geological hazards are characterized by extensive affected areas, severe consequences, and significant long-term cascading effects, posing serious threats to human life, property safety, and sustainable socio-economic development in earthquake-affected regions. Existing landslide susceptibility assessments mainly focus on regional-scale predictions, while slope-scale co-seismic responses and dynamic hazard management remain insufficiently studied. This study focuses on Hejiapo, Longchi Town, Dujiangyan City, located in the strong earthquake-affected area of the Wenchuan earthquake. A comprehensive approach integrating multi-temporal remote sensing interpretation, field geological surveys, topographic mapping, statistical modeling, and numerical simulation was employed to identify major controlling factors and evaluate landslide susceptibility. Six conditioning factors were selected, and an area-based Information Value (IV) model was applied for landslide susceptibility assessment. Based on the susceptibility assessment results and field investigation, potential unstable zones were identified. Numerical simulations were further conducted to analyze the potential failure modes, movement processes, and affected areas of these unstable zones. The results indicate that: (1) A total of 28 post-earthquake landslides were identified, with a cumulative area of 364,932 m², accounting for 19.64% of the study area. These landslides exhibit characteristics of being spatially clustered and having relatively small individual scales and a single dominant failure type. (2) The susceptibility assessment results show that very-high- and high-susceptibility zones are negatively correlated with road distance and fault distance, and are mainly distributed along both sides of the roads and steep ridge areas. (3) Two potential unstable zones were delineated, covering a total area of approximately 0.051 km², mainly distributed along the southwestern ridges of Hejiapo. (4) The identified unstable zones have the potential to generate high-elevation landslide debris flow hazards. Numerical simulations reveal that the debris materials would mainly migrate downward along slope channels. The channel–road intersections and channel outlets are identified as the principal potential impact zones, which should be considered a priority area for geological hazard prevention. This study provides a methodological framework beyond the traditional static approach of “hazard investigation” by establishing a dynamic risk management concept (susceptibility zones–risk points–potential disaster chains). The proposed framework enhances the understanding of the entire disaster prevention and mitigation process and provides scientific support for precise regional hazard mitigation planning and territorial spatial management. Full article

Breast milk is the optimal nutrition for infants. However, a number of viruses can accumulate in breast milk and be transmitted to a newborn during breastfeeding. Holder pasteurization is a gold-standard method for breast milk treatment, achieving an acceptable compromise between microbiological safety and biological quality. We have evaluated the virucidal activity of Holder pasteurization of breast milk contaminated with RNA and DNA viruses in a portable home-use pasteurizer. Viruses were selected on basis of their ability to infect humans, contaminate breast milk, and pose a threat for a newborn. The efficacy was evaluated in experimental conditions using breast milk samples from four different donors. Infectious virus and viral genome content were assessed. Holder pasteurization resulted in a decrease in infectivity below the detection limit for enveloped RNA viruses (HIV-1, CHIKV, and SARS-CoV-2), but not RSV, and a partial decrease in infectious virus titer for non-enveloped RNA enteroviruses (EV-A71, E30, and PV1), and DNA viruses (hAdV5 and HSV-1). Moreover, infectivity reduction in E30 and PV1 in different samples of breast milk significantly differed. Therefore, Holder pasteurization can be a tool in the novel pandemic preparedness toolbox or probably help mothers manage breastfeeding during common infections. Full article

With the large-scale integration of electric vehicles (EVs), distributed charging scheduling has become a key enabler for coordinated charging management. However, its reliance on information exchange makes it susceptible to cyberattacks, including False Data Injection (FDI), Denial of Service (DoS), and replay attacks. Such attacks may compromise privacy and corrupt or interrupt communication, leading to incorrect consensus prices and degraded scheduling performance. To mitigate these threats, this paper proposes a distributed resilient charging scheduling strategy. Specifically, anomalous nodes are detected through neighbor-based observations, while a belief-degree-based trust mechanism is employed to isolate low-trust nodes and suppress attack propagation. In addition, an individual price resetting mechanism is developed to restore convergence to the optimal price of the remaining EVs following node isolation. Simulations on communication networks with 5 to 100 EVs show that, under all three attacks, the compromised node is detected at the second and isolated at the third consensus iteration after attack onset, no healthy node is falsely isolated, and the remaining fleet converges to the optimum of the reduced scheduling problem with a price deviation below $8.2\times {10}^{-3}$. A buffered detection envelope extends these guarantees to asynchronous communication, heterogeneous time-varying delays, packet losses, and intermittent attacks: in 245 randomized stress runs on 5-to-100-EV networks, every attacker is isolated, no healthy node is isolated outside the harshest composite scenario, and the final price deviation remains below $1.1\times {10}^{-2}$. Extensive simulations under representative cyberattack scenarios verify the effectiveness and robustness of the proposed strategy within the stated assumptions. Full article

Flash floods are sudden and highly destructive, posing serious threats to human life. Conventional annual-scale or static susceptibility assessments cannot adequately capture intra-annual dynamics, while susceptibility alone does not directly represent potential population risk. This study integrated a 1990–2016 flash-flood inventory, monthly precipitation, multisource topographic and environmental factors, and population exposure to develop an XGBoost- and SHAP-based monthly flash-flood susceptibility (FFS) model for Hunan Province, China. As a retrospective application, population exposure was then incorporated to map population-exposure-based flash-flood risk (FFR) from January to December 2024. Independent temporal testing for 2013–2016 showed good model performance (AUC = 0.82). SHAP identified monthly maximum 1-day precipitation (M1P) as the most important factor (23.0%), followed by elevation (21.1%) and the topographic wetness index (14.7%). FFR increased markedly in June–July, peaking in July, when high- and very-high-risk areas covered 26.86% of the province, and declined sharply in August, remaining relatively stable thereafter. High-risk areas were mainly concentrated in eastern, southeastern, and parts of central Hunan. Jaccard similarity between high-to-very-high FFS and FFR ranged from 0.478 to 0.687, indicating that population exposure reshaped the spatial priority of susceptibility hotspots. The framework provides insights into seasonal variations in population-exposure-based flash-flood risk and supports improved understanding of spatial risk patterns. Full article

Drones have become a significant threat in both modern warfare and civilian environments, presenting unprecedented security challenges. While hard kill (physical destruction) defense methods exist to counter threat drones, they remain prohibitively expensive and often result in collateral damage. This study focuses on software-based peaceful neutralization as an anti-drone countermeasure, with particular emphasis on optimizing Wi-Fi credential recovery by leveraging exploitable vulnerabilities in threat drone wireless communications. The methodology included extensive performance evaluation experiments to assess vulnerabilities in drone Wi-Fi networks through GPU-accelerated credential recovery. Across the tested GPU platforms, the proposed implementation achieved an average throughput improvement of 8.78% over the Hashcat baseline, with platform-specific gains of 5.0% on H100, 8.77% on A100, 6.53% on RTX PRO 6000, and 14.8% on RTX 4060. The proposed approach represents a preliminary step toward a cost-effective and damage-minimizing alternative, applicable to Wi-Fi-dependent drone platforms under specific operational conditions. Our experimental results suggest the potential feasibility of credential-recovery-based approaches as an enabling step toward software-based neutralization of Wi-Fi-dependent drone systems. This paper contributes to the emerging field of counter-unmanned aerial systems (C-UAS) by establishing a methodology for credential recovery against Wi-Fi-based drone communications, which may contribute to reducing reliance on physical intervention in applicable scenarios. Full article

Sustainable transport policy shifts traffic from road to rail, which presupposes safe and reliable railroad-road crossings. Migration of railway communications to the Future Railway Mobile Communication System (FRMCS) makes that safety depend on digital resilience: an attack on the communication layer translates directly into the physical state of a crossing. This paper develops a threat model for FRMCS at level crossings covering three attack classes that act on that state—repetition and injection of Euroradio telegrams, functional identity takeover, and abuse of the railway emergency call—and evaluates their detection by machine learning. Operational FRMCS traffic does not yet exist, so detectors were trained on 400,000 records generated from the 3GPP and UIC specifications in four phases of increasing difficulty. Over 20 independent replications, the random forest and XGBoost achieved a macro F1 score of 0.920 ± 0.003, were statistically indistinguishable, and inferred in under 1 ms per packet, meeting the real-time requirement for a crossing. Detection capability varies far more between attack classes than between models, and the gravest attack class is the hardest to detect. Costing the operational and environmental impact of an attack-induced safe state links detection to Sustainable Development Goal targets 9.1 and 11.2 and to NIS2 obligations. Full article

Municipal solid waste transfer stations generate substantial odorous emissions during waste collection and compression, posing threats to the surrounding environment and public health. Existing odor control technologies predominantly employ single-process configurations, ill-suited to the operational complexity of transfer stations. An efficient odor mitigation process was designed and implemented for a large-scale waste transfer station in Shanghai. The project established a combined system integrating source control and end-of-pipe treatment and verified its performance through full-scale actual operation. The results show that the odors are reduced effectively in situ by source control using a plant extract spray coupled with an ion oxidation fresh air system. At the end-of-pipe stage, a hybrid process combining a non-thermal plasma with a two-stage plant extract scrubber is employed to enhance the synergistic degradation of key odorants, including H₂S and NH₃. The monitoring results indicate that the post-renovation concentrations of H₂S, NH₃, and odor concentration in the unloading and transfer hall and at the emission stack remain well below the associated limits of the local Emission Standards for Odorous Pollutants (DB31/1025-2016). This study demonstrates that for waste transfer stations in environmentally sensitive residential areas, a systematic approach integrating source control and end-of-pipe treatment can achieve effective odor abatement performance, thereby improving the environmental quality of surrounding communities. Full article

Side-channel analysis (SCA) remains a significant threat to embedded cryptographic implementations, yet hardware security evaluation often lacks a systematic workflow that links leakage detection, exploitability validation, trace-efficiency estimation, and robustness assessment. Rather than proposing a new leakage distinguisher, this work presents an evidence-based framework that integrates established SCA techniques into a reproducible hardware security evaluation methodology for embedded AES implementations. The proposed framework follows a structured screen–validate–quantify–stress-test pipeline. Power traces are acquired from an AES-128 implementation on a CW312/SAM4S target using a shunt-based measurement path and firmware-triggered ChipWhisperer–Husky acquisition. Fixed-versus-random test vector leakage assessment (TVLA) using 5000 fixed and 5000 random traces reveals strong first-order leakage, with a dominant peak of |t| = 210.61 at sample index 1916, defining a leakage window of [1666:2167]. Exploitability is validated through timing-aware per-byte correlation power analysis (CPA) under a Hamming-weight leakage model, successfully recovering the full AES-128 key and revealing staggered byte-wise leakage timing for point-of-interest selection. Operational feasibility is quantified through 100 independent subsampling trials over trace budgets M ∈ {25, 50, 75, 100, 150, 200, 300, 400, 700, 1000}, achieving 97% full-key recovery with 25 attack traces and 100% recovery from 50 attack traces onward when using pre-established byte-specific POIs. Independent split-data point-of-interest validation further confirms the stability and reproducibility of the selected leakage locations. Finally, controlled synthetic timing-misalignment experiments evaluate robustness by measuring recovery degradation, weakest-byte confidence behaviour, and the effect of local peak-search compensation under non-ideal analysis conditions. Collectively, this case study demonstrates a reproducible SCA evaluation workflow on the investigated CW312/SAM4S AES-128 implementation, integrating leakage screening, exploitability validation, trace-efficiency analysis, independent POI validation, and controlled robustness testing. The present results establish the workflow on this experimental platform, but evaluation across additional devices, firmware implementations, keys, acquisition sessions, and protected implementations is required before broader generalisation can be claimed. Full article

With the growing complexity of mountain tunnel construction and the increasing engineering demand for rapid detection, ground-penetrating radar (GPR) has become a core technique for engineering-scale fracture detection. Fracture zones are not only a major obstacle to tunneling through complex geological sections but also a key trigger of tunnel hazards, posing a serious threat to construction safety. Consequently, the effective identification of fracture zones and the investigation of their development characteristics remain central challenges in advance geological prediction for mountain tunnels. Owing to the complex morphology of fracture zones, existing approaches—including simple model simulation, single-parameter identification, and integrated prediction methods—cannot adequately characterize core attributes such as connectivity and development degree. To address this challenge, this study constructed a fracture attribute model and established a multidimensional collaborative identification system covering fracture scale, connectivity, and density, integrating time-domain wave-frequency morphology with the two instantaneous attributes in the spatial domain, namely instantaneous frequency and instantaneous amplitude. A collaborative analysis scheme based on multiple GPR statistical attributes was adopted to perform a qualitative comparison of fracture development characteristics by tracking multidimensional parameter trends. The results reveal that fracture zones in different development states exhibit certain correlation trends between their internal structural features (e.g., connectivity and compactness) and the frequency-related physical attributes of electromagnetic waves (e.g., wave-frequency morphology and instantaneous frequency). Based on the simulation results, the analysis of waveform characteristics, spatial variations of the two instantaneous attributes, and multi-attribute evolution trends clarified the correlation trends between the wave-frequency response patterns and the fracture development degree. These findings provide scientific and theoretical guidance for geological prediction in tunnel engineering in Hunan, China, and lay a foundation for the GPR-based identification of fracture zones in tunnels. Full article

The rapid workplace diffusion of conversational artificial intelligence (CAI) introduces a new social actor, yet existing reviews often neglect the intrapsychic processes through which employees construe these systems. This systematic review synthesizes the emerging psychological mechanisms suggested by the literature through which CAI reshapes the employee experience. Following PRISMA guidelines, we searched the Web of Science database (up to February 2026) and analyzed 84 SSCI-indexed studies based on the sociotechnical systems theory, the Computers as Social Actors paradigm, and the Job Demands–Resources model; the corpus quality was appraised using the Mixed Methods Appraisal Tool. Findings reveal that while instrumental framing remains dominant, a meaningful theoretical turn has emerged among a minority of studies, where conversational artificial intelligence is increasingly conceptualized as a partner or supervisor fundamentally driven by anthropomorphic cues. The synthesis identifies a double-edged reshaping of work: cognitively, human intelligence augmentation competes with skill threat; emotionally, constant support contrasts with social fabric erosion; and career-wise, inclusion coexists with work alienation and generative AI loafing. These dimensions are psychologically coupled, with patterns indicating that cognitive threats co-occur with emotional anxiety and moral expediency. While limited by cross-sectional primary evidence, the review provides an integrated conceptual model illustrating proposed conceptual pathways that connect multi-layered antecedents to employee outcomes, which could be adopted to guide future longitudinal validation. By shifting the focus from productivity to psychological experience, this study offers theoretical foundations for human–AI collaboration and outlines a future research agenda on trust dynamics and algorithmic fairness. Full article

Ocean acidification, driven by the absorption of anthropogenic carbon dioxide into seawater, poses a significant threat to marine environments and biodiversity. This comprehensive review examines the specific impacts of acidification on the reproductive physiology of fish, integrating recent advances in our understanding of gonadal development, gametogenesis, and early embryonic development. We synthesize current findings on how altered carbonate chemistry disturbs the reproductive system from gametogenesis and fertilization to larval development and beyond. Our analysis reveals that ocean acidification affects multiple physiological pathways, including disruption of the hypothalamic–pituitary–gonadal (HPG) axis, impairment of calcium signaling, and alterations in sex hormone synthesis. We propose a mechanistic model in which pH reduction may weaken Dax1-mediated repression of P450arom transcription, thereby dysregulating aromatase activity and leading to sex hormone imbalance that impairs gonadal development; however, this proposed mechanism requires direct experimental validation under ocean acidification, as the net outcome may be modulated by additional regulatory layers such as epigenetic modifications and other transcription factors. Furthermore, we discuss the synergistic effects of ocean acidification with temperature elevation and other stressors, which often exacerbate reproductive dysfunction. This review concludes that no multi-generational study has yet evaluated HPG axis disruption to population-level recruitment failure, a critical gap for fishery resource assessment. We emphasize the critical importance of multi-generational and field-based studies to fully understand these impacts, which is essential for forecasting consequences for global marine biodiversity and the long-term sustainability of fisheries. Full article

Background/Objectives: Antimicrobial resistance (AMR) poses a global threat, exacerbated by the overuse and misuse of antimicrobials. This study evaluated antimicrobial use and prescribing patterns among inpatients in three tertiary hospitals in Nepal. Methods: A cross-sectional point prevalence survey of the inpatient module of the standardized Global-PPS methodology was conducted across three tertiary hospitals in Nepal during the 2022 Global-PPS Period 2 (P2) survey, carried out between May and August 2022. Data were collected via a paper-based form (University of Antwerp;) and analyzed descriptively. Antimicrobial use was defined as the proportion of inpatients receiving at least one systemic antimicrobial on the day of the survey. Results: Of 379 patients eligible for the study, 233 (61.5%) received at least one antimicrobial, with the highest prevalence observed in adult and neonatal intensive care units (ICUs). Out of 371 antibiotic prescriptions, third-generation cephalosporins (30.2%) and carbapenems (8.1%) were the most frequently prescribed, and 61.0% of overall antibiotics belonged to the Watch group of WHO AWaRe classification. Surgical prophylaxis often exceeded recommended durations, with 91.5% of prophylactic antibiotics prescribed for more than one day. Empirical therapy dominated 56.33% of all prescriptions, with targeted therapy limited to 4.04%. Conclusions: The high overall prevalence of antimicrobial use, the reliance on empirical therapy, and over-reliance on Watch-group antibiotics and Unclassified AWaRe antibiotics highlights the urgent need for robust antimicrobial stewardship programs in Nepal. Full article