Search Results (2,546)

Full-physics reservoir simulation for CO₂ storage becomes computationally expensive when many operational schedules must be screened, motivating machine-learning (ML) surrogates that reduce simulation burden while preserving the essential physics-driven response. We propose an activity-based partitioning methodology that produces an interpretable applicability map, identifying regions where surrogate substitution is expected to be reliable and regions where highly active dynamics make it unsafe. In this work, we focus exclusively on the slow-varying region and develop proxy models for pressure and saturation time derivatives in that domain. The fast-varying region is intentionally excluded, and no fully coupled hybrid simulator is claimed at this stage. The partition is constructed from temporal changes in derivative signals and aggregated across multiple schedules to obtain a conservative, scenario-robust delineation. For slow cells, local stencil-based neural proxies leverage overlapping time windows and features describing the local state, schedule forcing, and injector influence. Because saturation derivatives in the slow region are strongly zero-inflated, with many cells remaining outside the advancing CO₂ plume for long periods, a two-stage strategy is adopted: first detecting whether meaningful change occurs and then predicting the derivative magnitude only when active, with additional smoothing to suppress near-zero artifacts. The framework also supports selective surrogate deployment over user-selected time windows. The objective is therefore to establish a conservative zone of applicability for derivative-based ML updates, rather than to demonstrate full simulator replacement or end-to-end coupled acceleration. In the case study, 5914 of the 8243 grid blocks evaluated by the proxy workflow were classified as slow-varying, corresponding to 71.7% of the evaluated proxy-analysis domain. For the blind schedule, full-rollout pressure reconstruction produced mean absolute errors of 5.34, 3.69, and 2.80 psi over early, middle, and late time-window groups, respectively. In a future coupled implementation using the same partition, these 5914 cells could be advanced by the ML proxy, while the remaining dynamically active or unsupported cells would remain under full-physics treatment. This would reduce the full-physics active-cell count from 9212 to 3298 in the future coupled setting, although direct wall-clock acceleration remains to be quantified after simulator integration. Full article

Modeling thermal environments across scales is crucial for climate-adaptive design and energy management. Traditional numerical methods (e.g., CFD) offer high accuracy and physical consistency, but they are computationally expensive. In contrast, purely data-driven models, though efficient, lack physical consistency and generalization capability. This review systematically examines Physics-Informed Neural Networks (PINNs), a hybrid paradigm in which physical prior knowledge is embedded directly into the neural network training process. A structured keyword search of the Web of Science Core Collection was performed, and 94 peer-reviewed journal articles were analyzed. The evolution from numerical simulations and data-driven surrogate models to PINNs is outlined. PINN methods are classified according to the stage at which physical prior information is integrated (i.e., dataset development, model construction, or loss function formulation). Current research remains heavily focused on loss function constraints, whereas systematic integration into data augmentation and model construction remains limited. Application domains span indoor environments, outdoor environments, and building systems, with each domain exhibiting unique prior integration strategies tailored to specific problems. Future PINN modeling should evolve toward multi-physics coupling, adaptive loss balancing, cross-scenario transfer learning, and unified evaluation benchmarks. PINNs in this field are promising but remain at an early stage, especially for complex urban-scale deployment. This review synthesizes existing research around the three stages of dataset development, model construction, and loss function formulation, summarizes the prior integration strategies adopted in the domain of building thermal environments, and provides a practical workflow for embedding physical prior knowledge at different stages of model development. Full article

Accurate recognition of partially occluded faces remains challenging in unconstrained and real-time environments, especially under masks, partial occlusions, pose variation, and illumination changes. This study presents a controlled comparison of three hybrid multi-face recognition pipelines for robust occluded face recognition. For fair evaluation, all pipelines use the same SCRFD face detector, preprocessing protocol, Linear SVM classifier, and 60% unknown rejection threshold, while varying only the feature extractor: ResNet29, ConvNeXt, and ResNet100 with ArcFace embeddings. To reduce data leakage, models are trained only on normal, non-occluded faces and tested on unseen partially occluded faces. Evaluation is performed on a custom dataset and the public Real-World Occluded Faces dataset, alongside three existing paper methods with publicly available code tested under the same experimental protocol. The SCRFD with ArcFace ResNet100 and Linear SVM pipeline achieved the best results compared to existing papers and our other pipelines, reaching 97.475% real-time accuracy for five faces and over 99% confusion-matrix-based accuracy on the custom dataset. On the ROF dataset, it also achieved closed-set accuracies of 98.66% for sunglasses and 97.92% for masks, with threshold-based accuracies of 96.35% for the sunglass test and 95.14% for the mask test. Furthermore, it obtained EER values below 0.007 and AUC values above 99%. In real-time testing, it achieved 29.25 FPS with 34.18 ms/frame latency on a GPU-enabled laptop and approximately 5 FPS with 273.4 ms/frame latency on a Raspberry Pi 4. Full article

Background: Breast cancer remains a leading cause of cancer-related mortality, and reliable computational decision support is increasingly viewed as a complement to expert pathological assessment rather than a replacement for it. Variational quantum classifiers (VQCs) and Quantum Support Vector Machines (QSVMs) have recently been promoted as candidate models for medical classification, yet most published comparisons rely on internal hold-out validation alone and report only a single point estimate of discrimination, omitting calibration, decision-analytic value, and explainability—three ingredients that any clinically credible model must furnish. Methods: We assembled a complete quantum–classical machine learning pipeline and evaluated it under a deliberately stringent protocol designed to expose, rather than conceal, the limitations of current Noisy Intermediate-Scale Quantum (NISQ)-era models. The analytical hypothesis was conservative and stated in advance; in light of saturated classical baselines on this benchmark, we did not anticipate a quantum advantage in raw discrimination, and we framed the study as a methodological probe rather than as a competition. Using the Wisconsin Diagnostic Breast Cancer (WDBC) dataset (n = 569) for development and an independent Wisconsin Original (WBC) cohort (n = 683) for external validation, we benchmarked five classical learners (XGBoost, LightGBM, CatBoost, RandomForest, RBF-SVM), two quantum models (an eight-qubit VQC implemented in PennyLane and a ZZ-feature-map QSVM implemented in Qiskit), and a stacked hybrid ensemble. The evaluation framework combined Optuna-driven hyperparameter optimisation, internal–external cross-validation, and external validation on the independent WBC cohort. Robustness and interpretability were then probed through circuit depth and embedding rotation ablation, depolarising noise stress tests, learning curve and feature stability analysis, decision curve analysis, and dual SHAP-based explanations covering both a direct tree-based explanation and a quantum surrogate. Reporting followed the TRIPOD + AI guideline. Results: On the internal test partition, RBF-SVM achieved the highest discrimination (AUC = 0.998), with XGBoost, LightGBM, CatBoost, the hybrid ensemble, and the VQC clustering between 0.992 and 0.996; the QSVM with a ZZ-fidelity kernel underperformed substantially (AUC = 0.727). Pairwise tests for correlated ROC curves indicated that most differences among top models were not statistically significant. On the external WBC cohort, model rankings reorganised, as RBF-SVM (AUC = 0.986, 95% CI 0.946–0.997), RandomForest (0.985, 95% CI 0.945–0.996), VQC (0.983, 95% CI 0.942–0.995), and the hybrid ensemble (0.982, 95% CI 0.941–0.995) all retained near-ceiling discrimination with extensively overlapping confidence intervals. Ablation analysis demonstrated that the choice of embedding rotation is decisive—Z-rotation embeddings collapsed VQC performance to chance levels (AUC ≈ 0.50), whereas X- and Y-rotations preserved it. Depolarising noise up to p = 0.10 had a negligible effect on the VQC, and SHAP analyses converged on worst concave points, mean concave points, and worst area as the dominant predictors across both classical and quantum models. Decision curve analysis showed positive net benefit for both classical and hybrid models across the clinically meaningful threshold range, exceeding both the treat-all and treat-none reference strategies throughout. Conclusions: In the present regime, the principal contribution of QML is not raw discrimination—modern classical learners are already at the data ceiling—but the construction of a rigorous, reproducible, externally validated, and interpretable benchmarking framework in which quantum models can be fairly compared with their classical counterparts. Because evaluation was confined to curated benchmark datasets rather than real-world clinical populations, the interpretability and net benefit findings reported here should be read as benchmark-level evidence and not as a demonstration of readiness for clinical deployment. Full article

Gene duplication and subsequent selection are central to genome evolution and transcription factor diversification, but the conservation and divergence of the basic helix–loop–helix (bHLH) family in pear remain unclear from a pan-genome perspective. Here, we performed a pan-genome and transcriptome-guided analysis across 15 pear genome assemblies, including Asian pear, European pear, and hybrid/haplotype assemblies. Genome-wide duplicated gene pairs were classified into different duplication types, and Ka, Ks, and Ka/Ks values were calculated to establish an evolutionary background for duplicated pear genes. Based on this framework, 3222 bHLH were identified and grouped into evolutionary clades and orthologous gene groups. The pear bHLH family contained conserved core members and variable dispensable members, indicating both functional conservation and genome diversification. Duplication and Ka/Ks analyses showed that WGD/segmental duplication contributed to bHLH expansion and that most duplicated PbrbHLH gene pairs were constrained by purifying selection. By integrating 17-tissue and fruit-development transcriptomes from three pear cultivars, 39 fruit-development-associated PbrbHLHs were selected. Co-expression analysis with 185 PbrMYBs identified candidate MYB–bHLH co-expression modules from the available pear fruit-development transcriptomes. These results provide an evolutionary framework for pear bHLH diversification and candidate regulatory modules for future functional studies. Full article

As global populations grow and technology advances, daily life is increasingly shaped by digital systems such as computers and smart devices. However, prolonged device use has contributed to increasing physical and mental health concerns, particularly those associated with poor sitting posture. Posture-related strain is frequently overlooked and contributes to musculoskeletal discomfort, including back, neck, shoulder, and wrist pain, and may also be associated with sleep disturbances and elevated stress levels. To the best of our knowledge and based on the existing literature, this is the first study to introduce a machine learning-based framework for advanced muscle strain severity classification using Internet of Things (IoT) devices that integrates posture monitoring and muscle strain detection into a unified low-cost framework ($23 hardware cost). The primary objective of this work is accurate classification of muscle strain severity, while real-time alerts serve as a secondary ergonomic feedback mechanism. Specifically, this study makes four major contributions. First, we created a novel dataset through real-time acquisition of electromyography (EMG) and posture signals from participants in hospital and industrial environments, capturing diverse muscle strain patterns validated against clinical assessment procedures. Second, we designed a two-part hardware architecture consisting of posture detection (PD) and strain detection (SD) modules using a NodeMCU ESP8266, HC-SR04 ultrasonic sensor, EMG sensor, and buzzer for real-time physiological monitoring, incorporating EMG-specific preprocessing including band-pass filtering, rectification, and RMS smoothing. Third, we proposed and evaluated a hybrid machine learning framework integrating Vision Transformer (ViT) and XGBoost to classify strain severity into three study-specific categories: baseline (EMG RMS < 40 µV), compensatory strain (40–59 µV), and overload (≥60 µV). These categories were used as reproducible severity proxies for machine learning annotation and should not be interpreted as universal biomarkers of structural tissue damage. Finally, the proposed framework achieved a classification accuracy of 99.0% (95% CI: 98.5–99.5%) with an inference latency of 15.2 ms. Full article

Background: Cell division plays a crucial role in plant growth and development. Cyclin-dependent kinase inhibitors (ICK/KRP) negatively regulate the cell cycle, thereby affecting cell elongation and organ development. This study aimed to systematically identify and characterize the ICK/KRP gene family in pepper, and to explore their roles in growth, development, and stress responses. Methods: Bioinformatics approaches were used for genome-wide identification, chromosomal localization, collinearity analysis, sequence characterization, promoter element prediction, and tissue-specific expression profiling of pepper ICK genes. Phylogenetic analysis was performed with homologs from Arabidopsis, tomato, maize, and rice. Quantitative real-time PCR and virus-induced gene silencing (VIGS) were applied to validate gene expression patterns and gene function, respectively. Subcellular localization assays were also conducted. Results: A total of six ICK genes were identified in pepper. They were classified into three subfamilies and distributed on different chromosomes, with one pair showing evidence of duplication. All ICK/KRPs contain the conserved Motif 1 (amino acid sequence: KIPTTREIEEFFATAEKQQQRRFIEKYNFDPVNEKPL) and were predicted to localize to the nucleus. Promoter analysis revealed cis-acting elements associated with plant development, stress responses, and hormone signaling. Expression pattern analysis indicated tissue-specific divergence and significant induction/repression under temperature stress. qRT-PCR results were consistent with transcriptome data, and expression differences were observed in materials with different stigma lengths. Subcellular localization confirmed that Caz03g38750.1 and Caz12g03790.1 proteins localize to both the nucleus and plasma membrane. Silencing of CazICK1 significantly repressed stigma elongation and altered stigma morphogenesis. Conclusions: The six pepper ICK/KRP genes display distinct diversity in distribution, structure and expression, and function in plant growth, development and stress adaptation. This work not only lays a solid basis for exploring the cell cycle regulatory network of pepper and contributes to relevant theoretical research, but it also identifies key gene resources for improving stigma traits. It has great potential for application in molecular breeding to promote high yield and efficient hybrid seed production in pepper. Full article

Diagnosing tilapia diseases in complex aquaculture environments is severely hindered by noisy backgrounds and limited high-quality pathological data. To overcome these bottlenecks, this study presents a two-stage diagnostic framework integrating an enhanced Segment Anything Model (SAM) with an Adaptive Neuro-Fuzzy Inference System (ANFIS). In the first stage, SAM is augmented with a Convolutional Block Attention Module (CBAM) feature adapter and a Region Proposal Network (RPN)-based prompt encoder. This design enables the automated and precise extraction of irregular disease lesions by self-generating spatial prompts, thereby isolating water background noise. In the second stage, clinical color features extracted from the lesion masks are classified using ANFIS. To optimize performance on small-scale datasets, ANFIS parameters are trained via Particle Swarm Optimization (PSO) under a numerically stable One-vs-Rest (OvR) binary cross-entropy loss. Validated on the public dataset “Enhancing Disease Detection in Nile Tilapia”, our method delivers an average segmentation Dice coefficient of 86.2% and a classification accuracy of 93.5%. This hybrid approach demonstrates strong potential as a foundational baseline for the automated monitoring of aquaculture diseases. Full article

Detecting and identifying faults in industrial chemical plants is essential for safe and stable operation, and modern monitoring systems increasingly rely on deep learning to classify faults from multivariate sensor data. A practical obstacle to adoption is trust: most deep-learning diagnosers reach their decisions through a classification layer that operators cannot inspect, making it hard to see how the model maps process signals to a particular fault. This study targets fault diagnosis on the Tennessee Eastman (TE) process, a standard benchmark of simulated chemical-plant sensor data, and asks whether this final decision stage can be made directly inspectable without sacrificing accuracy. We propose CNN–Transformer–KAN (CTKAN), a hybrid model that learns local temporal patterns with a one-dimensional convolutional encoder, captures global inter-time-step dependencies with a Transformer encoder, and classifies faults with a Kolmogorov–Arnold Network (KAN) head whose learnable B-spline activations can be plotted and examined individually, in place of a conventional multi-layer perceptron (MLP). On the TE benchmark, CTKAN attains a Macro-F1 of 91.38 ± 0.26% over ten independent runs, comparable to a CNN + Transformer + MLP ablation (91.21 ± 0.32%) and a capacity-matched MLP-head variant (91.43 ± 0.37%) within seed-to-seed variability. The main finding is therefore not a higher score: at matched capacity the KAN and MLP heads are statistically indistinguishable in accuracy, so the KAN head’s value is to add a directly inspectable view of the classification stage at no measurable accuracy cost, helping process engineers sanity-check how the diagnoser separates faults in safety-critical settings. Full article

Medulloblastoma is a heterogeneous solid tumor, and its molecular characteristics are the most important prognostic factors for this neoplasm. Unfortunately, the molecular classification of MB-G3 and MB-G-4 medulloblastoma is very complex because of molecular similarity. Therefore, in this work, through unsupervised machine learning-based gene expression profiling, we identified a low molecular profile associated with four molecular groups of medulloblastoma. We performed medulloblastoma expression microarray data mining via the Partek Genomics Suite and Transcriptome Analysis Console (TAC), and we included a total of 25 fresh medulloblastoma tumors that were obtained and hybridized into HG U133 Plus 2.0 Array microarrays. To identify the molecular groups of the 25 patients, we compared them against classified patients, which were obtained from free repositories, and through data mining based on gene expression, compared the expression profiles of our patients. To do so, we performed an analysis via the least squares method via PCA. The molecular groups MB-WNT and MB-SHH were confirmed via immunohistochemistry via β-catenin, YAP1 and GAB1 antibodies in tissue fixed in formalin and embedded in paraffin, and another tissue section was placed on a Visium Spatial slide to perform spatial RNA-seq via Illumina NextSeq 2000 platform sequencers. The data obtained were analyzed with R. We identified the expression profiles associated with the four molecular groups and formed a reference set. Through unsupervised analysis via the least squares method, we assigned the molecular profiles of 25 patients with medulloblastoma, via the integration of bulk and spatial tumor molecular gene expression profiling analysis and with immunohistochemical findings, this strategy was fast and accurate. We observed correlations in three of the trials carried out and, in part, in one study, a patient who presented two tumor strains and two molecular signatures (SHH and G4), which led us to believe that this patient presented mixed phenotypic characteristics. Multigene expression profile analysis of medulloblastoma represents a significant advance in precision medicine; integrating different layers of transcriptomic information allows us to demonstrate underlying molecular changes in the four molecular groups that are essential for personalized therapy. Full article

Hybrid RF-FSO-THz communication systems are promising candidates for future Internet of Things (IoT) and 6G networks because they combine the robustness of radio frequency links, the high-capacity potential of Free-Space Optical communications, and the ultra-wideband capabilities of terahertz transmission. Adaptive channel selection in such systems depends on multiple correlated environmental and physical-layer variables, including distance, rain intensity, humidity, visibility, turbulence strength, signal-to-noise ratio, channel capacity, and energy-efficiency metrics. This paper presents a physics-supported benchmark framework for supervised adaptive channel selection in hybrid RF-FSO-THz systems and systematically investigates the impact of linear and nonlinear dimensionality-reduction techniques on predictive performance, statistical robustness, computational complexity, and physical interpretability. A multi-scenario dataset comprising 5000 samples was generated using calibrated RF, FSO, and THz propagation models under clear, rain, fog, and worst-case environmental conditions. Principal Component Analysis (PCA) and Kernel PCA were evaluated together with Random Forest, Support Vector Machines (SVMs), XGBoost, Gradient Boosting (GB), Multi-Layer Perceptron (MLP), Logistic Regression, and Decision Trees. The results demonstrate that PCA preserves nearly all predictive capabilities while reducing the original 33-dimensional feature space by approximately 81.8%, maintaining accuracies close to 97–98% with the best-performing classifiers. Statistical significance analysis confirms that PCA introduces only modest degradations, whereas Kernel PCA consistently reduces the predictive performance while increasing memory requirements and inference latency. Additional environmental-only validation experiments indicate that adaptive channel selection remains highly learnable even when only pre-selection environmental descriptors are available, partially mitigating concerns regarding self-consistency bias. Overall, the results suggest that PCA provides an advantageous compromise among predictive accuracy, computational efficiency, statistical robustness, and physical interpretability for supervised adaptive channel selection in physics-supported hybrid wireless communication systems. Full article

Non-Intrusive Load Monitoring (NILM) enables estimation of the energy use of individual appliances in smart buildings from a single aggregate meter. In practice, however, this task is not straightforward. Signals from different appliances can overlap, and the measured data may also include distortions such as spikes, drops, and noise. To address these issues, this study presents a multi-task triple-hybrid deep learning framework that handles appliance classification and anomaly detection together. The model brings together 1D-CNN, BiLSTM, and Transformer Attention so that local patterns, temporal dependencies, and wider contextual information can be learned within the same structure. It also uses a dual-output design to classify appliance categories and detect anomaly types simultaneously. Experiments were carried out on Building 1 of the UK-DALE dataset with four appliances: kettle, microwave, washer dryer, and fridge freezer. For the anomaly task, synthetic disturbances were added to segmented signal windows and grouped as normal, spike, drop, and noise. To check how well the proposed framework handled different scenarios, it was tested on both the UK-DALE and REDD datasets. Looking at the main UK-DALE results, the model correctly identified appliances 99.48% of the time and spotted anomalies with 98.80% accuracy. A secondary test on the REDD dataset yielded an 86.44% classification score. This proves the architecture can adjust to completely new power grid environments without losing its edge. On top of that, when pitted against standard benchmark models like Seq2Point, this triple-hybrid design clearly does a better job of mapping out complex signal changes. As a result, it yields much stronger anomaly detection metrics. Full article

Ship self-sinking accidents threaten maritime safety, human life, property, and the marine environment, and understanding their causal-factor associations is essential for developing effective preventive measures. This study aims to identify the multi-level factors, recurrent association patterns, and critical structural nodes involved in ship self-sinking accidents. A hybrid framework integrating grounded theory, the Human Factors Analysis and Classification System (HFACS), FP-growth association rule mining, and complex network analysis was applied to 150 accident investigation reports released by the China Maritime Safety Administration between 2014 and 2024. Findings suggest that adverse weather and sea conditions, inadequate ship safety management, and crew incompetence are the most frequent factors. Thirty causal factors were identified and classified into four HFACS levels, and 229 association rules were generated to construct a directed weighted causal-factor association network with 19 nodes and 229 edges. Network results indicate that inadequate ship safety management, crew incompetence, ship unseaworthiness, insufficient maintenance of hull weathertight integrity, and improper or untimely emergency measures occupy critical positions in the association structure. This research offers insight into ship self-sinking accidents and identifies priority intervention points for more targeted maritime supervision, safety management and accident prevention. Full article

Urban environments increasingly rely on interconnected digital infrastructures like IoT devices, SDN-enabled networks, and cloud platforms to support essential municipal services. Ensuring the resilience of these systems requires advanced, data-driven mechanisms capable of detecting and mitigating cyber disruptions. This study presents Persian Eagle, a hybrid machine learning and deep learning framework designed to enhance the cyber-resilience of urban digital infrastructures by providing high-precision detection of Distributed Denial of Service (DDoS) attacks. DDoS attacks disrupt service availability by flooding targets with massive malicious traffic orchestrated through botnets, and in critical infrastructures, disruptions can be life-threatening. The proposed framework integrates multi-stage data preprocessing, SMOTE-based class balancing, and a four-phase feature-selection pipeline combining filtering, statistical ranking, PCA, and XGBoost. Seven complementary classifiers, including Random Forest, SVM, Gaussian Naive Bayes, XGBoost, MLP, LSTM, and Autoencoder, are bonded through a stacking cooperative with a Gradient Boosting meta-learner. The framework was evaluated on CICDDoS2019 and CICIDS2017 datasets, and achieved near-perfect performance up to 99.9998% accuracy, demonstrating strong generalization across diverse attack scenarios. By offering a scalable, transparent, and data-driven detection mechanism, Persian Eagle maintains urban digital-risk management and supports the continuity and resilience of critical smart-city services. Full article

Reliable carbon emission data are essential for the effective operation of emissions trading systems (ETS), especially as China’s ETS expands to include energy-intensive industries. This study proposes a hybrid, risk-based anomaly detection framework for high-frequency CO₂ emission data by cross-validating material-based emissions with flue gas-based monitoring data. Under normal operating conditions, the ratio of material-based to flue gas-based emissions is expected to remain within a relatively stable distribution. Potential high-risk periods can therefore be identified when this relationship is distorted or when local temporal patterns deviate from expected behavior. The framework combines Hartigan’s dip test with a window-based Random Forest (RF) classifier, which is suitable for continuous monitoring data that may exhibit temporal dependence. The framework was evaluated using 15-min CO₂ emission data from a cement production facility, with simulations of anomaly magnitude, duration, and mode. Results show that the dip test performs well for long-lasting or strong anomalies, whereas the RF model is more sensitive to subtle, short-term deviations. In the integrated framework, 94.7% of anomalous periods were detected by at least one method and flagged as potential data-quality risks, whereas normal periods were not flagged, supporting its use to prioritize verification efforts. Full article