Search Results (4,353)

The giant honeybee, Apis dorsata, is an important pollinator native to tropical and subtropical Asia. Understanding its potential response to climate change is important for pollinator conservation, ecological risk assessment, and biosecurity planning. This study used an optimized MaxEnt ecological niche modeling framework to predict the current and future global habitat suitability of A. dorsata. The model was developed using 1060 occurrence records and seven non-collinear bioclimatic variables and projected under three global climate models (IPSL-CM6A-LR, BCC-CSM2-MR, and MPI-ESM1-2-HR) and three Shared Socioeconomic Pathways (SSP126, SSP245, and SSP585) for 2041–2060, centered on 2050. The optimized model used linear, quadratic, and hinge features (LQH) with a regularization multiplier of 0.5 and demonstrated good predictive performance under 5-fold spatial cross-validation (mean AUC = 0.905 ± 0.007; TSS = 0.746 ± 0.010). The results indicate that the potential distribution of A. dorsata is primarily associated with the combined effects of seasonal temperature and moisture availability. Current projections identified high climatic suitability across South and Southeast Asia, while also revealing potentially suitable environments in parts of Africa, the Americas, and northern Australia. Future projections suggest that suitable climatic conditions will largely persist through 2050, with habitat gains generally exceeding losses and increasing under higher climate-forcing scenarios. Continued monitoring and proactive biosecurity are essential to address habitat loss within the native range and prevent establishment in newly suitable regions. This study highlights the potential redistribution of A. dorsata under future climate change. Full article

Flood prediction remains a critical challenge in environmental risk management and disaster preparedness. Accurate river-level forecasting is essential for the development of reliable early warning systems and the mitigation of flood-related risks. However, conventional ensemble approaches, such as averaging and majority voting, often exhibit limited adaptability when individual models respond differently to anomalies or incomplete data. To address this limitation, this study proposes a graph-based synchronization framework that integrates XGBoost and Random Forest models using a Graph Convolutional Network (GCN). The proposed framework represents the outputs of the base prediction models as graph nodes and employs graph message passing to learn context-dependent relationships between their predictions. The framework is evaluated using real-world hydrological observations from the Lepenec River Basin in North Macedonia together with meteorological data obtained from the OpenWeatherMap API. Experimental results demonstrate that the proposed GCN-based synchronization framework outperforms both the standalone prediction models and the previously proposed linear synchronization method, achieving an R² value of 0.91 and a Mean Absolute Error (MAE) of 0.21. The obtained results indicate that graph-based synchronization provides an adaptive approach for integrating heterogeneous machine-learning models and has the potential to support future flood early-warning systems and intelligent environmental monitoring applications. Full article

Surface soil moisture forecasting is important for drought monitoring, irrigation management, and land–atmosphere process analysis but remains challenging because near-surface soil moisture is jointly influenced by antecedent precipitation, atmospheric drying, soil properties, vegetation conditions, and irregular multi-source observations. This study proposes an Antecedent-Precipitation-Informed Surface Soil Water Balance and Time-Aware Mamba–Mixture-of-Experts (API-SWB-Mamba-MoE) framework for forecasting in situ volumetric soil moisture at approximately 5 cm depth using only information available before the target time. The framework combines a process-guided API-SWB physical prior, a time-aware Mamba temporal encoder, a context-conditioned MoE residual decoder, and gated residual fusion. The U.S. source-domain stations were evaluated using five independently repeated station-level random holdout splits, with approximately 70%, 15%, and 15% of the stations assigned to training, validation, and testing in each repetition, respectively. The five resulting U.S.-trained models were further applied without target-domain retraining or fine-tuning to six German and French stations for zero-shot transfer evaluation. Across the U.S. test prediction–observation pairs pooled from the five repetitions, the proposed model achieved a Pearson correlation coefficient (R) of 0.934, a root mean square error (RMSE) of 0.035 cm³ cm⁻³, a Kling–Gupta efficiency (KGE) of 0.922, and a mean bias error (MBE) of −0.001 cm³ cm⁻³. Pooled zero-shot predictions yielded RMSE values of 0.036 and 0.038 cm³ cm⁻³ and KGE values of 0.885 and 0.917 for Germany and France, respectively. The pooled U.S. test RMSE was 7.9–22.2% lower than that of the ablation variants and 20.5–32.7% lower than that of the benchmark models. These results suggest that combining a process-guided prior with time-aware sequence modeling and context-conditioned expert routing offers a promising approach for station-scale soil moisture forecasting under irregular multi-source observations. The external results provide preliminary evidence of zero-shot transferability at the six selected sites, although broader regional validation remains necessary. Full article

Both hive displacement and entrance relocation can challenge honeybee homing navigation, yet the temporal dynamics of colony-level homing behavior following hive entrance displacement remain poorly quantified. To address this, we established a non-invasive automated pipeline integrating YOLO11m-based detection, OC-SORT tracking, and the Homing Rate (HR) to monitor nine honeybee (Apis mellifera) colonies under semi-natural apiary conditions. HR links trajectory endpoints with the experimentally defined valid entrance state and provides a colony-level measure of entrance-targeting accuracy. Horizontal entrance displacement caused a substantial reduction in HR in the treated colonies, whereas the Control Group showed only a small concurrent change. During the subsequent four-day observation period, all six treated colonies displayed a similar dynamic pattern characterized by an initial rapid increase in HR, followed by a slower increase. The asymptotic exponential model provided a better descriptive representation of these temporal dynamics than a linear model. The 14-day observation of Colony A1 further revealed an early increase, a transient decline, and subsequent recovery toward a relatively stable level, indicating that the post-displacement trajectory was not strictly monotonic. Overall, this study provides an automated quantitative pipeline for continuously characterizing colony-level entrance-targeting behavior and its temporal dynamics following hive entrance displacement. Full article

Urbanization is widely recognized as a major driver of biodiversity loss; however, cities can also provide important refuges for pollinators when diverse and continuous floral resources are available. This study assessed the diversity of spontaneous insect-pollinated flora and its associated pollinator communities across five major urban habitat types (grasslands, road verges, artificial surfaces, wooded areas, and wetlands) in the Mediterranean cities of Pisa, Livorno, and Lucca (central Italy). During a three-year field survey, spontaneous flowering plant species were systematically recorded, and pollinator visitation rates were quantified using standardized sampling protocols. Approximately 150 spontaneous insect-pollinated plant species were identified, providing a continuous supply of nectar and pollen throughout the flowering season. Pollinator assemblages included honey bees (Apis mellifera), solitary bees, bumblebees, Diptera, and Lepidoptera, with bees representing the dominant group across all habitats. Pollinator visitation rates differed significantly among habitat types, reaching their highest values in grasslands, road verges, and wetlands, whereas wooded habitats and artificial surfaces supported significantly lower pollinator activity. Numerous plant species belonging to diverse botanical families attracted a broad range of pollinator taxa, revealing complex plant–pollinator interactions and demonstrating the ecological value of spontaneous urban vegetation for sustaining pollinator diversity. In summary, the present study shows that spontaneous flora constitutes a fundamental component of urban green infrastructure, supporting pollination services and enhancing ecological connectivity within cities. Biodiversity-friendly management strategies that conserve and promote spontaneous flowering vegetation should therefore be considered a key element of sustainable urban planning and the ecological resilience of Mediterranean urban landscapes. Full article

Historic commercial buildings are often described as resilient, but the mechanisms that sustain ordinary use remain fragmented across morphology, adaptive reuse, heritage governance, and community research. This Review examines how inherited spatial configurations and actor arrangements jointly shape everyday-use resilience. Using transparent, PRISMA-informed searching, screening, appraisal, and narrative/thematic synthesis, we identified 50 eligible empirical studies from 3314 records. Evidence came from 18 countries or regional settings but remained concentrated in Malaysia (9 studies), Indonesia (8), Türkiye (7), and mainland China (7); no eligible study came from the Americas, and direct African evidence was limited to one Egyptian case. In response to prescreen-sensitivity concerns, all 80 records excluded solely because the deterministic dictionary did not detect a spatial-focus term were independently reassessed; 30 were advanced beyond title/metadata review and 13 ultimately met all eligibility criteria. Four evidence-informed spatial capacities recurred: network permeability, fine-grained and reversible subdivision, climatic and service compatibility, and active public–private interfaces. These capacities did not determine resilience independently. Owners and traders translated capacity into maintenance and incremental adaptation; users sustained social legitimacy; and public, professional, financial, and digital intermediaries enabled or constrained adaptation through regulation, investment, programming, and visibility. Heritage retention was assessed in all 50 studies; use vitality was observed in 40 and assessed prospectively in 10; community benefit was observed or partly assessed in 37, prospectively in 10, and not assessed in 3. The geographically concentrated, English-language, public-API, and full-text corpus limits transferability, while the targeted validation demonstrates why deterministic terminology rules require human sensitivity checks. Everyday-use resilience is therefore presented as an evidence-informed, actor-mediated proposition concerning the conversion of inherited spatial capacity into heritage retention, use vitality, and community benefit. Full article

Clinical adoption of large language models (LLMs) in public healthcare faces a structural impasse: the capital expenditure of centralised high-performance computing on one side and the privacy risk of routing patient data through third-party cloud interfaces on the other. Italian local health authorities (Aziende Sanitarie Locali, ASL) operate large fleets of clinical workstations that remain idle outside peak administrative hours. We present OmniMed Federated, a browser-native architecture using the WebGPU application programming interface (API) and the WebLLM framework to distribute LLM inference tasks across these existing workstations. The system federates task allocation rather than model training or partitioned inference: each query executes in full on one node, selected under a data residency constraint. A five-tier escalation model, coordinated by a metadata-only PHP back end, ranks tiers by data exposure rather than capability, with commercial cloud fallback disabled by default. In a pilot three-node testbed (50 queries), federated throughput reached 19.5 versus 8.2 tokens/second standalone, peak per-node memory fell 62%, and node discovery took 140 ms; query content remained within the institutional perimeter throughout. These figures establish infrastructural feasibility at pilot scale. Clinical output quality, security hardening, and scalability remain unevaluated. Full article

Elevated salinity can significantly affect the rheological behavior, consistency development, and early strength development of API Class G cement slurries. This study experimentally investigated the effect of NaCl additions of 0, 4, and 8 wt.% BWOC on cement slurries with densities of 1.5 and 1.9 g/cm³ at temperatures of 25, 50, 75, and 90 °C. Rheological parameters, the times required to attain selected UCA-estimated compressive strength thresholds, and the ultrasonic cement analyzer (UCA)-estimated compressive strength after 12 and 24 h were evaluated. The results indicated that the effect of NaCl was not proportional to its concentration but varied with temperature, slurry formulation, and the property being evaluated. NaCl generally reduced plastic viscosity and yield stress, whereas gel strength exhibited a non-uniform response with a pronounced dependence on temperature. Its influence on atmospheric consistency development and early strength development was likewise non-monotonic. Under certain test conditions, higher NaCl concentrations accelerated the attainment of the initial strength thresholds; however, they did not consistently increase the UCA-estimated compressive strength after 24 h or improve all evaluated operational properties. Among the three investigated NaCl concentrations, the 4 wt.% formulation frequently exhibited favorable responses across several evaluated parameters; however, no single NaCl concentration consistently provided the most favorable response for every property or under all investigated conditions. The findings provide a basis for the further optimization of cement slurry formulations intended for use in high-salinity environments. Since the experimental program did not include independent repetitions, the observed differences should be interpreted as descriptive trends that require statistical validation in future studies. Full article

Background/Objectives: Oseltamivir is a widely used antiviral agent indicated for the treatment of influenza and plays a central role in pandemic preparedness strategies. Although an extemporaneously prepared 15 mg/mL oral solution is recognized in formularies, preparing large volumes during a pandemic requires extended storage options. However, no stability studies are currently available for an oseltamivir 15 mg/mL oral solution stored under freezing and subsequent refrigerated conditions for the duration of a standard treatment. The aim of this study was to evaluate the physicochemical and microbiological stability of a 15 mg/mL oseltamivir oral solution prepared from the active pharmaceutical ingredient (API) and packaged in amber glass containers. Methods: The oral solution was formulated using oseltamivir phosphate API, sodium benzoate as a preservative, and purified water and then packaged in 125 mL Type II amber glass bottles, allowing for thermal expansion. The samples were stored at −20 ± 2 °C for up to 75 days, followed by refrigerated storage (5 ± 3 °C) after thawing for up to 10 days. Chemical stability was assessed using a validated HPLC method in accordance with ICH guidelines and was defined as 90–110% recovery of the initial concentration. Physical stability (color, pH, particulate matter, crystallization, and homogeneity) and microbiological stability were also evaluated. Results: The HPLC method demonstrated excellent linearity, precision, and accuracy. Oseltamivir concentrations remained within the predefined acceptance limits throughout the 75-day study period under freezing conditions, with no significant changes in pH, color, or particulate formation. After thawing, the drug concentration continued to remain fully stable and within the required limits for up to 10 days under refrigerated conditions. Despite the prolonged storage and phase changes, no significant changes in physical parameters were observed, and microbiological testing confirmed the absence of aerobic, anaerobic, and fungal microorganisms on the final day of the study. Conclusions: Oseltamivir 15 mg/mL oral solution in amber glass bottles is physicochemically and microbiologically stable for up to 85 days (75 days under frozen conditions plus 10 days under refrigeration after thawing). Full article

Artificial intelligence (AI) models are increasingly deployed in edge-deployed cyber-physical security systems for tasks encompassing monitoring, threat classification, and automated decision-making. While these models offer robust performance, their deployment through open or semi-open Machine Learning as a Service (MLaaS) interfaces exposes them to severe security threats, prominently model extraction attacks. In such attacks, an adversary systematically queries a target API to replicate the victim model’s behavior. This study proposes a novel hybrid defense framework combining Adaptive Misinformation (AM) and Trigger-Based Watermarking (WM) to protect AI models against black-box extraction. Utilizing a LeNet architecture, the victim model was trained on the MNIST dataset, while a simulated attack utilized 50,000 EMNIST samples to train a clone model. The framework employs Maximum Softmax Probability (MSP) for out-of-distribution (OOD) detection to identify suspicious queries and strategically inject misleading responses, alongside a fine-tuned embedded watermark for ownership verification. Experimental evaluations using 10-fold cross-validation reveal that the baseline extraction attack yielded a clone model accuracy of 96.32%. Upon implementing the AM + WM framework, clone model accuracy degraded significantly to 53.67%, while the victim model maintained an accuracy of 98.97%. Furthermore, the protected model achieved a 100% Trigger Match Rate (TMR), ensuring reliable intellectual property verification. The proposed framework provides a prototype validation for lightweight edge architectures to balance security, model utility, and ownership protection in cyber-physical deployments. Full article

Accurate and scalable driver performance assessment is critical for improving road safety and reducing traffic-related injuries and fatalities, particularly in low- and middle-income countries where the majority of global road deaths occur. This paper presents an exploratory proof-of-concept framework for automated driver evaluation that combines real-world dashcam footage, YOLOv8-based object detection, and multimodal large language models (MLLMs), specifically Gemini 1.5 Flash. Two prompting strategies, narrative and rule-based, were designed to assess driver behavior against standardized licensing criteria derived from the California Department of Motor Vehicles (DMV) driving performance evaluation score sheet. The framework was evaluated across 11 manually curated driving scenarios covering intersections, pedestrian crossings, stop signs, cyclists, and emergency vehicles. Ground-truth labels were established through consensus between two traffic engineering experts cross-referencing official California DMV evaluation criteria. In this preliminary evaluation, the rule-based prompt achieved higher agreement with ground-truth assessments (10/11 scenarios, 90.9%) compared to the narrative prompt (7/11 scenarios, 63.6%), particularly in detecting clear rule violations. The narrative approach demonstrated greater contextual flexibility in ambiguous situations. These results should be interpreted as preliminary, given the small sample size, manually curated dataset, and absence of large-scale statistical validation. Nonetheless, the findings illustrate how combining visual detection with structured language-model prompting may support interpretable, policy-aligned driver evaluation. Key limitations include dependence on video quality, limited scenario diversity, absence of temporal behavioral modeling, and reproducibility constraints tied to proprietary API behavior. Future work should expand validation to larger annotated datasets, incorporate temporal sequence modeling, and explore region-specific regulatory adaptation. Full article

Agentic AI is changing enterprise cybersecurity as AI systems move beyond passive content generation toward autonomous planning, tool use, delegated execution, and operational action. As agents connect to email, code repositories, security operations center (SOC) platforms, finance workflows, cloud services, and enterprise application programming interfaces (APIs), they increasingly function as dynamic non-human identities rather than conventional software tools or service accounts. Existing identity and access management (IAM), Zero Trust, machine identity, and AI-governance approaches remain fragmented in their treatment of delegated authority, task intent, autonomy, runtime tool use, and auditable organizational consequences. This paper addresses these gaps by proposing the AIGATE (Agentic Identity Governance, Authority, Tool-Control and Evidence) Framework. AIGATE integrates eight governance layers: agent identity registration, lifecycle governance, delegated authority mapping, intent-bound access, least agency and least privilege, runtime tool-call control, audit evidence and accountability, and revocation and resilience. The framework treats agents as governed non-human enterprise identities whose actions remain attributable to designated human and organizational roles. AIGATE is developed through a structured critical synthesis of the recent literature on agentic AI security, machine identity, Zero Trust, runtime enforcement, and AI governance, with literature-derived governance requirements mapped explicitly to the eight framework layers. Three SOC, DevOps, and finance scenarios are used as illustrative applications rather than empirical validation. The contribution is an integrated governance architecture connecting identity, delegated authority, autonomy, runtime enforcement, evidence, and revocation across the agent lifecycle. Full article

The manual creation of personalized travel itineraries remains a labor-intensive process that requires travel agents to consolidate heterogeneous information from multiple sources, including natural language interactions, booking confirmations, screenshots, and reservation documents. Although recent advances in multimodal artificial intelligence have significantly improved language understanding and content generation, existing solutions typically address isolated tasks rather than providing an integrated workflow capable of automating the complete travel planning process. This paper presents the development and validation of an AI-powered conversational agent designed to automate itinerary generation and generative video synthesis through natural language processing (NLP) and multimodal data extraction. The system, evaluated at Technology Readiness Level 4 (TRL4), employs an agentic workflow in which a single orchestrator model delegates subtasks to specialized models through tool calling, integrating specialized large language models (Gemini 2.5/2.0 Flash, GPT-4o-mini) with optical character recognition (OCR) and Latent Diffusion Models to interpret user requests, extract structured data from images and PDFs, and produce comprehensive travel packages formatted as professional PDF deliverables alongside promotional videos assembled programmatically from curated and AI-generated visual assets. Validation in a controlled laboratory environment demonstrated over 90% intent recognition accuracy, successful data extraction from non-standard formats, and the automated production of both client-ready documents and coherent visual narratives from static itinerary data. The system achieved an average video generation latency of approximately $4.2$ min while maintaining structural consistency in the generated PDF itineraries. The system represents a viable proof-of-concept for intelligent travel planning automation, with implications for enhancing operational efficiency in the tourism industry. Future work will advance the prototype to TRL5 through integration with external booking APIs and real-user testing scenarios. Full article

Generalized category discovery (GCD) aims to classify known categories while discovering novel ones in unlabeled data, yet existing methods lack mechanisms to correct boundary-ambiguous samples that receive noisy pseudo-labels, as they primarily rely on visual feature learning without external semantic guidance. Vision-language models (VLMs) offer a natural source of cross-modal semantic correction. However, applying VLM-guided contrastive signals directly within the GCD training loop proves counterproductive because the locally-oriented InfoNCE loss conflicts geometrically with the globally oriented K-means objective in the shared backbone space. We identify the root cause as a dual resource allocation problem: the VLM-derived signal must be allocated to the correct feature subspace to avoid geometric conflict with K-means clustering (space allocation), and the limited VLM inference budget must be allocated to the correct samples to maximize discriminative return (budget allocation). These two decisions are coupled; failure on either renders the other ineffective. To resolve this, we propose DualGLEAN, a framework that addresses the dual allocation challenge through two coupled mechanisms: decoupled contrastive alignment (DCA), which routes the VLM-guided neighbor contrastive loss to a dedicated projector space while preserving the backbone space for global clustering, and compound uncertainty querying (CUQ), a three-stage filtering metric that jointly evaluates predictive entropy, boundary proximity, and local label inconsistency to direct VLM queries exclusively to truly boundary-critical samples. Extensive experiments on the AID and RSSDIVCS datasets demonstrate that DualGLEAN achieves strong performance, improves four diverse GCD baselines as a plug-in module, generalizes across seven VLM backbones, introduces zero additional trainable parameters to the base GCD network, and incurs a total VLM API cost of only CNY 2.45 per full training run on the AID dataset under the default search-scope configuration, with the cost scaling linearly with the query budget. Full article

Memory-safety vulnerabilities remain the dominant class of security defects in C/C++ software underpinning Internet infrastructure. Rust offers a structural solution through its ownership system, yet migrating existing codebases remains costly. This paper defines a structured methodology for LLM-assisted porting of security-critical C libraries to idiomatic Rust and applies it to cJSON (∼3200 LOC, 14 CVEs). A manual expert porting serves as the baseline; five LLMs (Claude Opus 4.6, Gemini 3 Pro, GPT-5.4, Kimi K2.7-Code, and Qwen3.5-27B) produce independent portings in agentic mode. Verification uses an end-to-end pipeline: CVE-specific tests, coverage-guided and differential fuzzing (>1.7 billion executions), Miri analysis, and comparative benchmarking. All six portings eliminate all in-scope CVE classes by construction, with zero unsafe blocks and zero memory-safety crashes. In this case study, structural safety holds consistently across all five evaluated models and across all five Kimi repetitions, whereas code quality varies widely (0–9 residual bugs). Because only Kimi was repeated ($N=5$), its variance bounds run-to-run noise at the 95% confidence level, against which some but not all between-model differences are distinguishable from chance; a single porting attempt costs approximately $3 in API usage. Differential fuzzing reveals complementary bugs in the manual and LLM portings, supporting a hybrid workflow, and we translate these findings into concrete practical guidance for teams planning a similar migration. These results are scoped to one compact, single-threaded C library. The entire codebase and evaluation pipeline are publicly released. Full article