Search Results (713)

Power-system safety is fundamental to protecting lives and ensuring reliable grid operation. Yet, hierarchical text classification (HTC) methods struggle with domain-dense accident narratives that require cross-sentence reasoning, often yielding limited fine-grained recognition, inconsistent label paths, and weak evidence traceability. We propose EG-HPT (Evidence-Grounded Hierarchy-Aware Prompt Tuning), which augments hierarchical prompt tuning with Global Pointer-based nested-entity recognition and a sentence–entity heterogeneous graph to aggregate cross-sentence cues; label-aware attention selects Top-k evidence nodes and a weighted InfoNCE objective aligns label and evidence representations, while a hierarchical separation loss and an ancestor-completeness constraint regularize the taxonomy. On a HFACS-based power-accident corpus, EG-HPT consistently outperforms strong baselines in Micro-F1, Macro-F1, and path-constrained Micro-F1 (C-Micro-F1), with ablations confirming the contributions of entity evidence and graph aggregation. These results indicate a deployable, interpretable solution for automated risk factor analysis, enabling auditable evidence chains and supporting multi-granularity accident intelligence in safety-critical operations. Full article

We present a novel deep generative semi-supervised framework for credit card fraud detection, formulated as a time series classification task. As financial transaction data streams grow in scale and complexity, traditional methods often require large labeled datasets and struggle with time series of irregular sampling frequencies and varying sequence lengths. To address these challenges, we extend conditional Generative Adversarial Networks (GANs) for targeted data augmentation, integrate Bayesian inference to obtain predictive distributions and quantify uncertainty, and leverage log-signatures for robust feature encoding of transaction histories. We propose a composite Wasserstein distance-based loss to align generated and real unlabeled samples while simultaneously maximizing classification accuracy on labeled data. Our approach is evaluated on the BankSim dataset, a widely used simulator for credit card transaction data, under varying proportions of labeled samples, demonstrating consistent improvements over benchmarks in both global statistical and domain-specific metrics. These findings highlight the effectiveness of GAN-driven semi-supervised learning with log-signatures for irregularly sampled time series and emphasize the importance of uncertainty-aware predictions. Full article

Cardiac arrhythmias are among the leading causes of morbidity and mortality worldwide, and accurate classification of electrocardiogram (ECG) beats is critical for early diagnosis and follow-up. Supervised deep learning is effective but requires abundant labels and substantial computation, limiting practicality. We propose a simple, efficient framework that learns self-supervised ECG representations with SimCLR and uses a lightweight Multi-Layer Perceptron (MLP) for classification. Beat-centered 300-sample segments from MIT-BIH Arrhythmia are used, and imbalance is mitigated via SMOTE. Framed from a symmetry/asymmetry perspective, we exploit a symmetric beat window (150 pre- and 150 post-samples) to encourage approximate translation invariance around the R-peak, while SimCLR jitter/scale augmentations further promote invariance in the learned space; conversely, arrhythmic beats are interpreted as symmetry-breaking departures that aid discrimination. The proposed approach achieves robust performance: 97.2% overall test accuracy, 97.2% macro-average F1-score, and AUC > 0.997 across five beat classes. Notably, the challenging atrial premature beat (A) attains 94.1% F1, indicating effective minority-class characterization with low computation. These results show that combining SMOTE with SimCLR-based representations yields discriminative features and strong generalization under symmetry-consistent perturbations, highlighting potential for real-time or embedded healthcare systems. Full article

Artificial intelligence (AI) is an important tool for improving agricultural tasks. In particular, object detection methods based on convolutional neural networks (CNNs) enable the detection and classification of objects directly in the field. Combined with unmanned aerial vehicles (UAVs, drones), these methods allow efficient crop monitoring. The primary challenge is to develop models that are both accurate and feasible under real-world conditions. This study addresses this challenge by evaluating marigold flower detection using three groups of CNN detectors: canonical models, including YOLOv2, Faster R-CNN, and SSD with their original backbones; modified versions of these detectors using DarkNet-53; and modern architectures, including YOLOv11, YOLOv12, and the RT-DETR. The dataset consisted of 392 images from marigold fields, which were manually labeled and augmented to a total of 940 images. The results showed that YOLOv2 with DarkNet-53 achieved the best performance, with 98.8% mean average precision (mAP) and 97.9% F1-score (F1). SSD and Faster R-CNN also improved, reaching 63.1% and 52.8%, respectively. Modern models obtained strong results: YOLOv11 and YOLOv12 reached 96–97%, and RT-DETR 93.5%. The modification of YOLOv2 allowed this classical detector to compete directly with, and even surpass, recent models. Precision–recall (PR) curves, F1-scores, and complexity analysis confirmed the trade-offs between accuracy and efficiency. These findings demonstrate that while modern detectors are efficient baselines, classical models with updated backbones can still deliver state-of-the-art results for UAV-based crop monitoring. Full article

Coastal debris is a global environmental issue that requires systematic monitoring strategies based on reliable statistical data. Recent advances in remote sensing and deep learning-based object detection have enabled the development of efficient coastal debris monitoring systems. In this study, two state-of-the-art object detection models—RT-DETR and YOLOv10—were applied to UAV-acquired images for coastal debris detection. Their false positive characteristics were analyzed to provide guidance on model selection under different coastal environmental conditions. Quantitative evaluation using mean average precision (mAP@0.5) showed comparable performance between the two models (RT-DETR: 0.945, YOLOv10: 0.957). However, bounding box label accuracy revealed a significant gap, with RT-DETR achieving 80.18% and YOLOv10 only 53.74%. Class-specific analysis indicated that both models failed to detect Metal and Glass and showed low accuracy for fragmented debris, while buoy-type objects with high structural integrity (Styrofoam Buoy, Plastic Buoy) were consistently identified. Error analysis further revealed that RT-DETR tended to overgeneralize by misclassifying untrained objects as similar classes, whereas YOLOv10 exhibited pronounced intra-class confusion in fragment-type objects. These findings demonstrate that mAP alone is insufficient to evaluate model performance in real-world coastal monitoring. Instead, model assessment should account for training data balance, coastal environmental characteristics, and UAV imaging conditions. Future studies should incorporate diverse coastal environments and apply dataset augmentation to establish statistically robust and standardized monitoring protocols for coastal debris. Full article

Fetal movement monitoring (FMM) is crucial for assessing fetal well-being, traditionally relying on clinical assessments or maternal perception, each with inherent limitations. This study presents a novel lightweight deep learning framework for real-time FMM on edge devices. Data were collected from 120 participants using a wearable device equipped with an inertial measurement unit, which captured both accelerometer and gyroscope data, coupled with a rigorous two-stage labeling protocol integrating maternal perception and ultrasound validation. We addressed class imbalance using virtual-rotation-based augmentation and adaptive clustering-based undersampling. The data were transformed into spectrograms using the Short-Time Fourier Transform, serving as input for deep learning models. To ensure model efficiency suitable for resource-constrained microcontrollers, we employed knowledge distillation, transferring knowledge from larger, high-performing teacher models to compact student architectures. Post-training integer quantization further optimized the models, reducing the memory footprint by 74.8%. The final optimized model achieved a sensitivity (SEN) of 90.05%, a precision (PRE) of 87.29%, and an F1-score (F1) of 88.64%. Practical energy assessments showed continuous operation capability for approximately 25 h on a single battery charge. Our approach offers a practical framework adaptable to other medical monitoring tasks on edge devices, paving the way for improved prenatal care, especially in resource-limited settings. Full article

Edge intelligence plays an increasingly vital role in ensuring the reliability of distributed microservice-based applications, which are widely used in domains such as e-commerce, industrial IoT, and cloud-edge collaborative platforms. However, anomaly detection in these systems encounters a critical challenge: labeled anomaly data are scarce. This scarcity leads to severe class asymmetry and compromised detection performance, particularly under the resource constraints of edge environments. Recent approaches based on Graph Neural Networks (GNNs)—often integrated with DeepSVDD and regularization techniques—have shown potential, but they rarely address this asymmetry in an adaptive, scenario-specific way. This work proposes Heterogeneous Graph Adaptive Augmentation (HGAA), a framework tailored for edge intelligence scenarios. HGAA dynamically optimizes graph data augmentation by leveraging feedback from online anomaly detection. To enhance detection accuracy while adhering to resource constraints, the framework incorporates a selective bias toward underrepresented anomaly types. It uses knowledge distillation to model dataset-dependent distributions and adaptively adjusts augmentation probabilities, thus avoiding excessive computational overhead in edge environments. Additionally, a dynamic adjustment mechanism evaluates augmentation success rates in real time, refining the selection processes to maintain model robustness. Experiments were conducted on two real-world datasets (TraceLog and FlowGraph) under simulated edge scenarios. Results show that HGAA consistently outperforms competitive baseline methods. Specifically, compared with the best non-adaptive augmentation strategies, HGAA achieves an average improvement of 4.5% in AUC and 4.6% in AP. Even larger gains are observed in challenging cases: for example, when using the HGT model on the TraceLog dataset, AUC improves by 14.6% and AP by 18.1%. Beyond accuracy, HGAA also significantly enhances efficiency: compared with filter-based methods, training time is reduced by up to 71% on TraceLog and 8.6% on FlowGraph, confirming its suitability for resource-constrained edge environments. These results highlight the potential of adaptive, edge-aware augmentation techniques in improving microservice anomaly detection within heterogeneous, resource-limited environments. Full article

Slope stability prediction is a critical task in geotechnical engineering, but machine learning (ML) models require large datasets, which are often costly and time-consuming to obtain. This study proposes a domain-driven teacher–student framework to overcome data limitations for predicting the dry factor of safety (FS dry). The teacher model, XGBoost, was trained on the original dataset to capture nonlinear relationships among key site-specific features (unit weight, cohesion, friction angle) and assign pseudo-labels to synthetic samples generated via domain-driven simulations. Six student models, random forest (RF), decision tree (DT), shallow artificial neural network (SNN), linear regression (LR), support vector regression (SVR), and K-nearest neighbors (KNN), were trained on the augmented dataset to approximate the teacher’s predictions. Models were evaluated using a train–test split and five-fold cross-validation. RF achieved the highest predictive accuracy, with an R² of up to 0.9663 and low error metrics (MAE = 0.0233, RMSE = 0.0531), outperforming other student models. Integrating domain knowledge and synthetic data improved prediction reliability despite limited experimental datasets. The framework provides a robust and interpretable tool for slope stability assessment, supporting infrastructure safety in regions with sparse geotechnical data. Future work will expand the dataset with additional field and laboratory tests to further improve model performance. Full article

Accurate classification of optical communication signal quality is crucial for maintaining the reliability and performance of high-speed communication networks. While existing supervised learning approaches achieve high accuracy on laboratory-collected datasets, they often face difficulties in generalizing to real-world conditions due to the lack of variability and noise in controlled experimental data. In this study, we propose a targeted data augmentation framework designed to improve the robustness and generalization of binary optical signal quality classifiers. Using the OptiCom Signal Quality Dataset, we systematically inject controlled perturbations into the training data including label boundary flipping, Gaussian noise addition, and missing-value simulation. To further approximate real-world deployment scenarios, the test set is subjected to additional distribution shifts, including feature drift and scaling. Experiments are conducted under 5-fold cross-validation to evaluate the individual and combined impacts of augmentation strategies. Results show that the optimal augmentation setting (flip_rate = 0.10, noise_level = 0.50, missing_rate = 0.20) substantially improve robustness to unseen distributions, raising accuracy from 0.863 to 0.950, precision from 0.384 to 0.632, F1 from 0.551 to 0.771, and ROC-AUC from 0.926 to 0.999 compared to model without augmentation. Our research provides an example for balancing data augmentation intensity to optimize generalization without over-compromising accuracy on clean data. Full article

Assessing teaching behavior is essential for improving instructional quality, particularly in Physical Education, where classroom interactions are fast-paced and complex. Traditional evaluation methods such as questionnaires, expert observations, and manual discourse analysis are often limited by subjectivity, high labor costs, and poor scalability. These challenges underscore the need for automated, objective tools to support pedagogical assessment. This study explores and compares the use of Transformer-based language models for the automatic classification of teaching behaviors from real classroom transcriptions. A dataset of over 1300 utterances was compiled and annotated according to the teaching styles proposed in the circumplex approach (Autonomy Support, Structure, Control, and Chaos), along with an additional category for messages in which no style could be identified (Unidentified Style). To address class imbalance and enhance linguistic variability, data augmentation techniques were applied. Eight pretrained BERT-based Transformer architectures were evaluated, including several pretraining strategies and architectural structures. BETO achieved the highest performance, with an accuracy of 0.78, a macro-averaged F1-score of 0.72, and a weighted F1-score of 0.77. It showed strength in identifying challenging utterances labeled as Chaos and Autonomy Support. Furthermore, other BERT-based models purely trained with a Spanish text corpus like DistilBERT also present competitive performance, achieving accuracy metrics over 0.73 and and F1-score of 0.68. These results demonstrate the potential of leveraging Transformer-based models for objective and scalable teacher behavior classification. The findings support the feasibility of leveraging pretrained language models to develop scalable, AI-driven systems for classroom behavior classification and pedagogical feedback. Full article

Electroencephalography signal generation remains a challenging task due to its non-stationarity, multi-scale oscillations, and strong spatiotemporal coupling. Conventional generative models, including VAEs and GAN variants such as DCGAN, WGAN, and WGAN-GP, often yield blurred waveforms, unstable spectral distributions, or lack semantic controllability, limiting their effectiveness in emotion-related applications. To address these challenges, this research proposes a Transformer-based conditional variational autoencoder–generative adversarial network (Trans-cVAE-GAN) that combines Transformer-driven temporal modeling, label-conditioned latent inference, and adversarial learning. A multi-dimensional structural loss further constrains generation by preserving temporal correlation, frequency-domain consistency, and statistical distribution. Experiments on three SEED-family datasets—SEED, SEED-FRA, and SEED-GER—demonstrate high similarity to real EEG, with representative mean ± SD correlations of Pearson ≈ 0.84 ± 0.08/0.74 ± 0.12/0.84 ± 0.07 and Spearman ≈ 0.82 ± 0.07/0.72 ± 0.12/0.83 ± 0.08, together with low spectral divergence (KL ≈ 0.39 ± 0.15/0.41 ± 0.20/0.37 ± 0.18). Comparative analyses show consistent gains over classical GAN baselines, while ablations verify the indispensable roles of the Transformer encoder, label conditioning, and cVAE module. In downstream emotion recognition, augmentation with generated EEG raises accuracy from 86.9% to 91.8% on SEED (with analogous gains on SEED-FRA and SEED-GER), underscoring enhanced generalization and robustness. These results confirm that the proposed approach simultaneously ensures fidelity, stability, and controllability across cohorts, offering a scalable solution for affective computing and brain–computer interface applications. Full article

Ensuring the surface quality of fresh soybeans is critical for maintaining their commercial value and consumer confidence. However, traditional manual inspection remains labor-intensive, subjective, and inadequate for real-time, high-throughput sorting. In this study, we present a multi-scale attention-augmented You Only Look Once version 8 (YOLOv8) framework tailored for real-time surface defect detection in fresh soybeans. The proposed model integrates two complementary attention mechanisms—Squeeze-and-Excitation (SE) and Multi-Scale Dilated Attention (MSDA)—to enhance the detection of small, irregular, and low-contrast defects under complex backgrounds. Rather than relying on cross-model comparisons, we perform systematic ablation studies to evaluate the individual and combined contributions of SE and MSDA across diverse defect categories. Experimental results from a custom-labeled soybean dataset demonstrate that the integrated SE+MSDA model achieves superior performance in terms of precision, recall, and Mean Average Precision (mAP), particularly for challenging categories such as wormholes and speckles. The proposed framework provides a lightweight, interpretable, and deployment-ready solution for intelligent agricultural inspection, with potential applicability to broader food quality control tasks. Full article

Supervised deep learning methods have been widely utilized in hyperspectral image (HSI) classification tasks. However, acquiring a large number of reliably labeled samples to train deep networks is not always possible in practical HSI applications due to the time-consuming and laborious labeling process. Semi-supervised learning is commonly used in scenarios with insufficient labeled samples. However, semi-supervised models based on a pseudo-label strategy often suffer from error accumulation. To address this issue and improve HSI classification performance with few labeled samples, a semi-supervised deep learning approach is proposed. First, a multi-scale convolutional neural network with accurate discriminative capability is constructed to reduce pseudo-label errors. Then, a new pseudo-label generation strategy based on Dropout is presented, in which feature-level data augmentation is applied by considering multiple predictions of the unlabeled samples to mitigate the error accumulation problem. Finally, the multi-scale CNN and the new pseudo-label strategy are integrated into a unified model to improve HSI classification performance. The experimental results demonstrate that the proposed approach outperforms other semi-supervised methods in the literature on four real HSI datasets with limited labeled samples. Full article

Gait recognition, as a biometric technology with great potential, has been widely applied in numerous fields due to its unique advantages. Through in-depth research and the creation of in-the-wild gait datasets, gait recognition technology is progressively extending from laboratory settings to complex real-world scenarios, achieving notable advancements. However, the complexity of annotating gait data inevitably leads to labeling errors, known as noisy labels, which are one of the reasons for the suboptimal performance of in-the-wild gait recognition. To address these issues, this paper explores noisy label learning for in-the-wild gait recognition for the first time. We propose a plug-and-play gait recognition framework named Dynamic Noise Label Correction Network (DNLC). Specifically, it consists of two main parts: the dynamic class-center feature library and the label correction module, which can automatically identify and correct noisy labels based on the class-center feature library. In addition, we introduce the two-stage augmentation strategy to increase the diversity of the data and help reduce the impact of noisy labels. We integrated our proposed framework into five existing gait models and conducted extensive experiments on two widely used gait recognition datasets: Gait3D and CCPG. The results show that our framework increased the average Rank-1 accuracy of five methods by 10.03% and 6.45% on the Gait3D and CCPG datasets, respectively. These findings demonstrate the superior performance of our method. Full article

In real-world visual recognition tasks, long-tailed distributions pose a widespread challenge, with extreme class imbalance severely limiting the representational learning capability of deep models. In practice, due to this imbalance, deep models often exhibit poor generalization performance on tail classes. To address this issue, data augmentation through the synthesis of new tail-class samples has become an effective method. One popular approach is CutMix, which explicitly mixes images from tail and other classes, constructing labels based on the ratio of the regions cropped from both images. However, region-based labels completely ignore the inherent semantic information of the augmented samples. To overcome this problem, we propose a saliency-guided local semantic mixing (LSM) method, which uses differentiable block decoupling and semantic-aware local mixing techniques. This method integrates head-class backgrounds while preserving the key discriminative features of tail classes and dynamically assigns labels to effectively augment tail-class samples. This results in efficient balancing of long-tailed data distributions and significant improvements in classification performance. The experimental validation shows that this method demonstrates significant advantages across three long-tailed benchmark datasets, improving classification accuracy by 5.0%, 7.3%, and 6.1%, respectively. Notably, the LSM framework is highly compatible, seamlessly integrating with existing classification models and providing significant performance gains, validating its broad applicability. Full article