Search Results (2,597)

The electrocardiogram (ECG) is essential for cardiovascular diagnosis but limited by inter-observer variability, low sensitivity for subclinical disease, and labor-intensive telemonitoring analysis. Artificial intelligence (AI), particularly deep learning, addresses these constraints by extracting high-dimensional patterns that correlate with arrhythmias, structural abnormalities, and systemic conditions. This integrative review synthesizes recent advances in AI-enabled ECG, covering technical foundations—including foundation models and validation strategies—and clinical applications, such as arrhythmia detection, structural heart disease identification, and digital biomarker derivation. We discuss emerging trends like self-supervised learning, multimodal integration, generative models, and explainability techniques. Furthermore, we tackle critical challenges regarding generalizability, algorithmic bias, privacy, and regulatory systems. Finally, we outline research priorities, including curated open datasets, and deployment in resource-constrained settings. With stringent validation, transparent governance, and human-centered design, AI-ECG has the potential to enhance cardiovascular diagnostics and clinical outcomes across a variety of healthcare settings. Full article

Neural machine translation (NMT) systems are widely used, but their performance remains strongly dependent on the availability of large-scale digital corpora, making translation for low-resource languages a persistent challenge. In parallel, large language models (LLMs) have recently emerged as a promising paradigm for multilingual text generation and translation; however, their behavior in low-resource settings remains largely underexplored. The challenge becomes even more acute for historical languages. Chagatai, a historical Turkic literary language of Central Asia with no native speakers, unstable orthography, and parallel data, represents an extreme case of such a condition. This study investigates whether transliteration significantly affects translation performance and how LLM-based and NMT-based systems compare under an extremely low-resource setting. To address these questions, we evaluated four source-text configurations (original Arabic script, expert manual transliteration, LLM-based transliteration, and rule-based Uroman transliteration) for translation into six target languages: Kazakh, English, Uzbek, Uyghur, Turkish, Russian, and Arabic. The results show that manual transliteration consistently yields the best translation performance, while noisy automatic romanization reduces these gains. For model comparison, GPT-4o was assessed alongside two fine-tuned NMT baselines, NLLB and TranslateGemma. The findings further show that LLM-based translation can be competitive with, and in some settings outperform, fine-tuned NMT systems, although this advantage comes with lower interpretability. Overall, these findings show that, for extremely low-resource historical languages written in non-Latin scripts, source-side representation is a decisive factor and may be as important as the choice of translation model itself. Full article

Amid escalating climate change and China’s “dual carbon” policy imperative, resource-based listed firms face urgent yet complex green transition challenges that single-factor approaches fail to adequately explain. Grounded in complex adaptive systems theory and the Technology-Organization-Environment (TOE) framework, this study employs dynamic fuzzy-set Qualitative Comparative Analysis (fsQCA) to analyze a panel of 160 Chinese resource-based A-share listed firms (2019–2024) to identify configurational pathways driving Environmental, Social, and Governance (ESG) performance. The results reveal eight high-performance pathways categorized into three archetypal patterns and six low-performance pathways reflecting distinct systemic misalignment traps. High- and low-performance pathways are causally asymmetric. Dynamic analysis further uncovers strong path-locking effects and divergent regional transition strategies. These findings advance configurational theory of corporate sustainability and offer strategic guidance for resource-dependent firms and policymakers. Full article

The growing deployment of Internet of Things (IoT) monitoring systems has resulted in demands for low-latency, secure, and energy-efficient intelligence on embedded chips. But most cloud-based and edge-assisted solutions are prone to high communication latency, lack adaptability, consume more energy, and lack decision security under resource-limited conditions. This paper introduces an Edge-Intelligent IoT Framework for Real-Time Adaptive Monitoring and Trust-Aware Secure Decision Validation with Resource-Aware Artificial Intelligence and Machine Learning (AI/ML) on embedded chips. Unlike conventional TinyML or Edge AI deployments that use a fixed inference model, the proposed framework introduces a validation-calibrated adaptive inference mechanism that jointly considers chip resources, input complexity, and sensor trust before accepting an embedded decision. The main scientific contribution is the unified coupling of resource-aware model selection with trust-aware decision validation for low-power embedded IoT inference. The framework dynamically selects the inference path and validates sensor trust before decision acceptance. Through experimentation, the proposed framework is demonstrated with 97.2% accuracy, 96.4% F1-score, and 98.1% AUROC, and 40.4% lower inference latency (31.2 ms to 18.6 ms) and 39.6% lower energy (9.6 mJ to 5.8 mJ) compared with traditional TinyML deployment. These results were obtained using the MHEALTH wearable IoT dataset with a leakage-safe 70:15:15 split and were statistically validated across five independent runs. The findings demonstrate a promising resource-aware TinyML-style embedded inference pipeline for wearable IoT monitoring, with improved latency-energy efficiency and trust-aware decision validation under the evaluated settings. Full article

Background: Preterm birth affects about 10% of births globally, often resulting in neurodevelopmental delays and disrupted parent-infant bonding. In low-resource settings, such as South Africa, neonatal intensive care unit (NICU) interventions require contextual adaptation. The NeuroSense PremmieEd Parenting Educational Intervention (PremmieSense) was developed to strengthen parent-infant bonding and promote neuroprotective care during NICU admission. Objectives: To describe the development, components, and pilot delivery of the PremmieSense intervention using the Template for Intervention Description and Replication (TIDieR) framework and document contextual adaptations and implementation lessons in low-resource NICUs. Approach: This project report outlines PremmieSense according to TIDieR. The programme comprises a picture-based booklet in English and Setswana, supplemented by facilitator-led group sessions delivered by trained healthcare professionals. It was piloted at two public-sector NICUs in the North West province of South Africa (N = 60 mothers; 30 per site). Parent knowledge was measured using the Knowledge of Preterm Infant Behavior (KPIB) scale, and stress was measured using the Parental Stress Scale: NICU (PSS:NICU tool). Quantitative outcomes are reported separately in a companion paper. Findings: PremmieSense was feasible and acceptable in low-resource NICUs. Logistical challenges including early discharges, staff constraints and language needs required pragmatic adaptations. The modular, multilingual design supported flexible delivery. TIDieR reporting facilitates replication and contextual adaptation. Conclusions and Recommendations: PremmieSense shows promise as a culturally appropriate and adaptable intervention for resource-constrained NICUs. Future work should tailor content to gestational age, prior parenting experience, and literacy, expand implementation, and assess long-term outcomes. Full article

The Upper Oligocene–Lower Miocene Alibonca Formation retains an essential record of intricate relationships among carbonate platform evolution, volcanic–sedimentary inflow, and high-purity silica deposition. This study examines the stratigraphic structure, paleoenvironmental development, and industrial viability of the Ağın diatomite deposits using comprehensive sedimentological, mineralogical, and geochemical investigations. Stratigraphic evidence indicates that the formation commenced with Early Miocene alluvial fan and shallow restricted marine sub-basin sedimentation prior to evolving into a significant marine incursion. This marine phase created a resilient carbonate platform structure consisting of reef-core, fore-reef, and back-reef sub-environments. Simultaneously, vigorous regional synsedimentary volcanism introduced high-flux silica pulses into the basin, acting as a major catalyst for diatom proliferation and high biological productivity within a restricted sub-basin setting. Geochemical analyses indicate that these bright white, diatomite deposits formed in conjunction with potassium-rich clays in a relatively deep, low-energy, and confined sub-basin of the Alibonca Sea. The high concentration of bulk SiO₂ and low trace element baselines are consistent with a high-purity deposional system and a low total rare earth element (ΣREE) abundance. However, their relatively high Al₂O₃ and K₂O contents indicate significant volcanic and terrigenous detrital input together with authigenic clay mineral formation during diatomite deposition, classifying the deposits as clay-bearing (argillaceous) diatomites rather than exceptionally pure diatomites. Chemical Index of Alteration (CIA) values indicate moderate continental chemical weathering under mostly hot and humid paleoclimatic conditions. The rapid terrestrial runoff and nutrient influx stimulated significant diatom growth before the ultimate late Early Miocene marine regression, transforming the area into a subaerial, volcanically influenced terrestrial environment. The Ağın deposits exemplify intra-platform marine silica sinks, demonstrating how tectonic–magmatic influences can surpass typical carbonate factory conditions to provide economically valuable biogenic mineral resources. Full article

This paper reports on the pilot implementation of ChakapBot, a generative AI-supported chatbot developed to support both the revitalisation and documentation of Baba Malay, an endangered heritage language in Singapore. A six-week pilot was conducted with 26 participants, the majority of whom were adult learners with limited prior exposure to the language and disrupted intergenerational transmission. Drawing on surveys, the study examines participant engagement, perceived usefulness, and the role of the chatbot within a blended learning model combining in-person instruction and digital practice. ChakapBot was trained on a community-curated corpus comprising a dictionary, a textbook, and a grammatical description of Baba Malay, positioning it as both a learning support tool and a means of operationalising existing documentation resources for community use. Findings indicate that participants valued the chatbot for its flexibility, accessibility, and low-pressure environment for independent practice, and perceived it as a complement to human-led instruction rather than a replacement. Participants also reported using Baba Malay beyond classroom settings, including in the home, workplace, and community events. The study highlights how ethically designed, community-curated AI tools can bridge language documentation and revitalisation by transforming documented linguistic knowledge into accessible, everyday practice for learners. Full article

Complex cyber-physical systems in the automotive and avionics domains increasingly consolidate safety-critical and non-critical functions onto shared multicore platforms. A central design challenge in these environments is ensuring strict timing guarantees while allowing graceful degradation under resource contention. In partitioned multiprocessor mixed-criticality systems, a mode switch on one processor forces all low-criticality (LC) tasks on that processor to degrade to their mandatory execution budgets, even though neighboring processors may have spare capacity. Existing approaches either focus on uniprocessor systems or address only the offline partitioning problem without considering the runtime survivability of LC tasks. This paper proposes IMC-PALM (Imprecise Mixed-Criticality via Partitioning and Adaptive Lightweight Migration), a two-phase framework for imprecise mixed-criticality (IMC) multiprocessor systems. The offline phase combines a tightened EDF-VD-IMC schedulability test with Mode-Balanced Partitioning (MBP). The tightened test identifies high-criticality tasks whose HI-mode utilization yields a tighter bound, while MBP allocates tasks based on mode-specific residual capacity. The runtime phase migrates LC tasks from a mode-switched processor to other processors that remain in LO mode, exploiting the per-processor isolation property of partitioned scheduling. Simulation results with 2, 4, and 8 processors show that MBP with the tightened test improves the acceptance ratio by 12.3 %p over existing algorithms under standard conditions. Furthermore, runtime migration reduces the degraded job ratio by 28.1 %p compared to the no-migration baseline under these standard settings, of which 4.6 %p is attributable to home-processor recovery. The benefit grows with the number of processors and remains robust under realistic migration overheads. Full article

Unmanned Aerial Vehicle (UAV)-based small object detection is a challenging computer vision task. It is constrained by two primary factors: UAV platforms have limited onboard computational resources, and high-altitude objects often have weak features that are easily overwhelmed by complex backgrounds. To address these challenges, we propose LUIM-YOLO. First, a Lightweight Multi-Scale Feature Enhancement (LMSFE) module integrates parallel multi-scale convolutions with attention to strengthen small and low-contrast object feature extraction. Second, an Adaptive Multi-Scale Bottleneck (AMSB) module enhances key semantic features of small objects and spatial correlation of medium-scale objects. Third, an Enhanced Cross-layer Compensation Feature Pyramid Network (ECC-FPN) constructs cross-level interaction pathways to improve small object position and scale perception. Experimental results on VisDrone2019 show that compared with YOLOv8n, LUIM-YOLO reduces parameters by 57% and improves mAP@50 by 12.9%. Additional full-validation-set PyTorch inference tests on NVIDIA Jetson Orin show that LUIM-YOLO achieves 88.19 ms/image in FP32, indicating a parameter-efficient accuracy-oriented design with edge deployment potential. Full article

Vehicular Ad Hoc Networks (VANETs) are mobile networks that offer new services and communication between moving vehicles, roadside infrastructure, and a trusted authority. With the development of autonomous and connected vehicles, the issue of authentication in VANETs has become increasingly prominent due to the lack of mutual trust among network entities. However, standard authentication models for VANETs must account for total computational and communication overhead, regardless of the timing of authentication message generation. To address this limitation, this work proposes an advanced authentication paradigm for VANETs called the online/offline VANET framework, and formalizes this novel framework to realize lightweight authentication by shifting heavy computational overhead to the offline phase. The proposed model is divided into an offline phase and an online phase. In the offline phase of the free time before the message becomes available, it allows more powerful trusted authority to pre-compute, and in the online phase, resource-constrained devices only execute a small set of residual operations. Based on this model and a new identity-based signature, we give an efficient instantiation and use a mobile platform to evaluate it. The experimental results demonstrate that our construction achieves low online computational and communication overhead. Full article

Background/Objectives: In resource-limited settings, where access to advanced imaging modalities such as magnetic resonance imaging (MRI) and histopathological confirmation may be limited, differentiating between hepatic cysts and metastatic lesions based solely on computed tomography (CT) remains challenging. This limitation may affect diagnostic confidence and increase the risk of misclassification, potentially impacting clinical decision making and patient management. In this study, we aimed to explore a more direct and automated approach for classifying hepatic lesions from CT images. Methods: We developed a deep learning-based framework combining transfer learning, decision fusion, and a stacking strategy by integrating five CNN architectures. The study included 100 patients, equally divided between metastatic liver tumors and pathological hepatic cysts. The dataset was built from both public data (LiTS) and internal clinical cases, and then split into training and testing sets. Results: The proposed stacking model provided the most consistent results, reaching an accuracy of 0.98, with high precision and sensitivity. The improvements in individual models, although moderate, were observed across all evaluation metrics. Conclusions: Overall, this approach offers a practical and reliable way to classify hepatic lesions with minimal manual intervention. It may help improve consistency in diagnosis and could be integrated into clinical workflows to support decision making in low-resource areas. Full article

The large-scale integration of renewable energy has significantly improved the low-carbon performance of power systems, but has also increased operational uncertainty. Under extreme weather conditions, wind and solar power may experience consecutive days of simultaneous output shortfalls—referred to as “renewable energy drought”—leading to persistently high net load and severe challenges to supply guarantee. To address this issue, this paper proposes a multi-source coordinated supply-guarantee dispatch strategy for consecutive-day renewable energy drought scenarios. First, net load is defined as the total system load minus the available wind and solar output. Based on magnitude and duration thresholds, renewable energy drought events are extracted from historical data to generate representative scarcity scenarios. Second, a multi-source coordinated optimization dispatch model is constructed, incorporating wind power, solar power, thermal units, battery energy storage, and pumped-storage hydro. The objective is to minimize the total system operating cost, which includes thermal fuel cost, start-up/shut-down costs, storage cycling cost, wind/solar curtailment penalty cost, and load shedding penalty cost. The load shedding penalty coefficient is set to a magnitude much higher than conventional costs to highlight the priority of supply guarantee. The model accounts for operational constraints such as minimum up/down times, deep regulation capability, ramping limits of thermal units, and charge/discharge power limits of storage. Taking a provincial power system in China for the year 2030 as a case study, a dispatch case covering four consecutive days (96 time periods) is designed. Based on a baseline scenario, eight groups of sensitivity analyses are conducted to comprehensively investigate the impacts of key factors on the supply-guarantee strategy, including: the minimum up/down time of thermal units, deep regulation capability, load shedding penalty cost, load level, rated energy capacity and charge/discharge efficiency of battery energy storage, rated energy capacity and pumping/generating efficiency of pumped-storage hydro, thermal fuel cost coefficient, and renewable energy capacity. Simulation results show that the proposed strategy can effectively coordinate multiple resources under consecutive-day drought conditions; reducing the minimum up/down time of thermal units improves supply flexibility but increases start-up/shut-down costs; enhancing deep regulation capability optimizes storage utilization and reduces total system cost; the load shedding penalty cost directly determines the trade-off between supply guarantee and economic efficiency; and as load level decreases by 5%, 10%, and 15%, the total system operating cost reduces by approximately 6.3%, 12.5%, and 18.8%, respectively. This study provides a quantitative method and technical support for supply-guarantee dispatch decisions and resource allocation in high-renewable power systems under persistent drought conditions. Full article

Background: The nursing process provides a structured framework for delivering safe, holistic, and patient-centered care; however, its implementation in primary healthcare settings, particularly in low-resource systems, remains inconsistent due to limited training and institutional support. Objectives: This study aimed to evaluate the effectiveness of a structured educational intervention in improving nurses’ knowledge and practical competencies in applying the nursing process in primary healthcare centers in Vlora, Albania. Methods: A quasi-experimental study was conducted with 32 nurses from five public primary healthcare centers. Sixteen nurses received a structured training program consisting of theoretical instruction and case-based practice, while sixteen nurses served as a control group. Pre- and post-intervention assessments were performed using standardized questionnaires and skill evaluation tools, and differences were analyzed using nonparametric statistical tests. Results: The results showed clear improvements in the intervention group across all domains, after the training. The reported use of the nursing process increased from 62.5% to 100%, while the use of Gordon’s Functional Health Patterns increased from 6.3% to 93.7%. The use of NANDA nursing diagnosis increased from 62.5% to 100%. The proportion of nurses reporting written nursing care plans increased from 62.5% to 93.7%, and the implementation and evaluation of care plans increased from 62.5% to 100%. The control group showed no comparable progress. Nurses who participated in the training also reported increased confidence and consistency in applying the nursing process in daily practice. Conclusions: These findings suggest that structured, competency-based training programs may improve immediate nurses’ theoretical knowledge and practical skills. Such training may contribute to improving the quality of nursing care, but further studies and longer follow-up and patient-related results are needed. Full article

Acute Lymphoblastic Leukemia (ALL) is a rapidly progressing blood cancer that demands timely and accurate diagnosis, particularly in pediatric patients. Conventional diagnosis relies on manual examination of peripheral blood smear images by hematologists, which is time-consuming, subjective, and prone to error. Although deep learning approaches have demonstrated high accuracy in medical imaging, many existing models are computationally intensive, limiting their use in resource-constrained clinical settings. To address these challenges, we propose LeukSNN, a lightweight spiking neural network (SNN) for automated ALL detection from peripheral blood smear images. Unlike conventional CNNs, SNNs employ sparse event-driven computations that can substantially reduce computational and energy requirements, making them attractive for deployment in low-resource healthcare environments. The proposed architecture combines depthwise separable convolutions, residual connections, and attention mechanisms within an SNN framework to achieve high classification performance with reduced computational cost. Experiments on three publicly available datasets demonstrate accuracies of 99.91–100%, while requiring only 8% and 28% of the multiplication and addition operations, respectively, of current state-of-the-art efficient methods. These results demonstrate that LeukSNN can achieve highly accurate and computationally efficient ALL classification on benchmark datasets, highlighting the potential of SNN-based approaches for future investigation in resource-constrained healthcare environments. Full article

Variations in mechanical load constitute a dominant source of domain shift in data-driven fault diagnosis of rotating machinery, causing models trained under one operating condition to degrade sharply when deployed under another. This work addresses the problem through unsupervised domain adaptation (UDA)—which transfers diagnostic knowledge from a labeled source condition to an unlabeled target condition by aligning their feature distributions—and introduces Padé Approximant Neural Networks (PadéNets) as compact yet highly expressive feature extractors. One-dimensional PadéNet encoders are embedded into three established adaptation frameworks—Deep CORAL, Domain-Adversarial Neural Networks (DANNs), and Conditional Domain-Adversarial Networks (CDANs)—to learn load-invariant representations without any labeled target data. On the Case Western Reserve University benchmark, where the models operate directly on raw time-domain vibration signals, replacing conventional convolutional encoders with PadéNets consistently improves cross-load diagnostic accuracy, reaching up to 99.28% average target-domain accuracy at a low parameter count. To assess generalization to a more demanding setting, the CDAN–PadéNet configuration is further evaluated on frequency-domain representations of the Paderborn University dataset, where domain shift arises from simultaneous variation of load torque and radial force on bearings with real accelerated-lifetime damage, attaining 99.84% average accuracy across six cross-condition transfer tasks while requiring fewer parameters than competing methods. These results establish PadéNet-enhanced UDA as an accurate, broadly applicable approach for robust bearing fault diagnosis under varying operating conditions, with a reduced parameter count suited to resource-constrained embedded platforms. Full article