Search Results (1,989)

Translational neuroscience relies on both in vitro slice recordings and in vivo recordings. Their spontaneous population dynamics are observed under decisively different conditions, and across independent experiments, there is typically no clear neuron-to-neuron correspondence. Here, we formulate a one-step-ahead, 1 ms binned, bidirectional transfer task between in vitro and in vivo multineuronal spike trains and provide a standardized evaluation procedure for generation across markedly different recording preparations. We train an autoregressive transformer on 1 ms binned, 128-unit binary sequences and introduce Dice loss to directly optimize spike-event overlap under extreme class imbalance, comparing it with Binary Focal Cross-Entropy (γ = 2.0). Across 12 mouse datasets (6 in vitro HD-MEA sessions and 6 in vivo Neuropixels sessions), the method achieves strong within-domain performance and remains above chance for cross-domain generation (ROC-AUC 0.70 ± 0.09 for in vitro → in vivo; 0.80 ± 0.10 for in vivo → in vitro). Because spike events are rare, we report Precision–Recall curves and PR-AUC alongside ROC-AUC to reflect minority-event quality. The present results should be interpreted as predictive generation under preparation/domain shift rather than as direct evidence of preserved causal biological dynamics; whether the framework also reflects features such as E/I balance or oscillatory structure remains an important question for future validation. To our knowledge, this is the first demonstration of bidirectional, time-resolved generation between unpaired in vitro and in vivo population spike trains without assuming cell correspondence, and the framework can be adapted to other sparse neural event data and related event-based datasets when domain-specific validation criteria are defined. Full article

Passive Optical Networks (PONs) play a vital role in providing high-speed broadband access in the 5G and F5G generation. However, their shared nature makes them vulnerable to physical-layer attacks like fiber bending, tapping and fiber cut. The problem is more serious in high-density PONs, where high split ratios result in high optical loss and overlapping back-scattered light, making it difficult to distinguish small attacks from background noise. Contrary to most existing works that neglect class imbalance and signal interference in high-density networks, this paper proposes a robust hierarchical two-stage attack detection scheme. First, we employ a binary classifier to distinguish eavesdropping attacks from normal traffic. Then, a second stage focuses on the specific eavesdropping categories (C1–C4). To address the small amount of attack samples, SMOTE is utilized for oversampling the minority class, and PCA-SVM is used to refine feature selection. Finally, the output of both stages is combined using probability score to obtain reliable decision. The experimental results show the effectiveness of our approach, achieving a classification accuracy of 89.07%. When evaluated on the same data, it has shown superior results in comparison to conventional machine learning algorithms, including decision tree (86.3%), k-nearest neighbors (79%), logistic regression (60%), and Naïve Bayes (52.6%). Full article

Coral reefs are among the most biodiverse marine ecosystems, playing irreplaceable roles in maintaining marine ecological balance and coastal services. Under dual pressures of global climate change and human activities, coral bleaching and degradation have become increasingly frequent, creating an urgent need for large-scale, long-term, and highly automated monitoring technologies. In recent years, advances in underwater imaging and deep learning have made visual recognition a core approach for coral classification and health assessment. However, most studies only focus on isolated model accuracy optimization, lacking systematic full-chain analysis integrating datasets, model evolution, cross-domain generalization, engineering constraints, and ecological adaptation, which severely hinders large-scale cross-regional and long-term application. This paper systematically reviews coral visual recognition technologies. It summarizes underwater image acquisition, public dataset characteristics, and annotation system evolution, then compares traditional feature engineering and deep learning in key tasks, highlighting their differences in feature representation and generalization. Four core challenges are identified: class imbalance, poor underwater image quality, weak cross-device/region generalization, and mismatched algorithm metrics with ecological needs. Finally, feasible solutions based on self-supervised pre-training, domain adaptation, and multimodal fusion are discussed to enhance model robustness and ecological interpretability, providing methodological support for intelligent coral reef monitoring systems. Full article

Accurate and rapid detection of anatomical structures, such as the glottis, is critical during tracheal intubation (TI) to ensure patient safety and procedural success. However, it remains a challenge due to the limited field of view and computational resources of video laryngoscopy, especially for difficult airway situations. Existing deep learning (DL) models struggle to balance high accuracy and real-time clinical deployment. To address these issues, we propose TI-YOLO (TI-You Only Look Once), a lightweight and efficient object detection model built upon the YOLOv11 architecture. TI-YOLO introduces the Bidirectional Feature Pyramid Network (BiFPN) module for multi-scale feature fusion, effectively enhancing the ability to detect anatomical structures of different sizes. TI-YOLO integrates the Deformable Attention Transformer (DAT) module to enhance the perception of crucial regions, improving detection accuracy and robustness. To further reduce the consumption of computational resources while maintaining efficiency, TI-YOLO is optimized by reconstructing the backbone based on MobileNetV4. Furthermore, TI-YOLO employs the Slide Weight Function (SWF) as a loss function during model training to mitigate the class imbalance within the dataset. One self-built dataset is used to validate the effectiveness of TI-YOLO. Compared to the original YOLOv11, TI-YOLO achieves mean Average Precision at IoU 0.50 (mAP50) scores of 0.902, with improvements of 3.8%. Meanwhile, TI-YOLO balances detection accuracy and computational efficiency with a 10.5% reduction in floating-point operations (FLOPs) and a 28.9% reduction in parameters, and the model weight is only 4.6 MB. Additionally, to evaluate TI-YOLO real-time inference capability, we quantize and deploy it on a low-cost embedded OrangePi 5 platform. The inference speed reaches over 50 frames per second (FPS), meeting real-time clinical requirements. Full article

Brucellosis poses a persistent threat to livestock health in high-altitude pastoral regions of China, where harsh environments and semi-nomadic grazing increase transmission risk. Existing surveillance systems rely mainly on periodic serological testing and lack effective early warning capability. This study proposes an ensemble learning-based early warning framework integrating veterinary epidemiological indicators with environmental and herd-movement data. A total of 4826 herd-level records collected over five years (2019–2024) were analyzed, with an overall positivity rate of 11.4%. Multi-source data, including serological, clinical, reproductive, vaccination, meteorological, pasture-management, and herd-movement information (from GPS tracking and structured surveys), were integrated through epidemiology-guided feature engineering. To address class imbalance and temporal dynamics, Synthetic Minority Over-sampling Technique (SMOTE) resampling and sliding time-window features were applied. The proposed ensemble model combines Random Forest, XGBoost, and LightGBM using a soft-voting strategy, with logistic regression as a baseline. Results show that the ensemble model outperforms single models, achieving an AUC of 0.86 and a PR-AUC of 0.65. After threshold optimization, sensitivity increased from 0.78 to 0.87. Under field conditions, the system provided herd-level early warnings with an average lead time of approximately 12 days before confirmed outbreaks, demonstrating its feasibility and practical value for proactive brucellosis surveillance in high-altitude pastoral systems. Full article

Earthquake-induced landslides in active orogens such as the Nepal Himalaya pose severe threats to lives, infrastructure, and post-disaster recovery. While machine learning (ML) and deep learning (DL) approaches to coseismic landslide susceptibility mapping have advanced considerably, spaceborne interferometric synthetic aperture radar (InSAR) products, particularly line-of-sight (LOS) displacement and coherence-based damage proxy maps (DPMs), remain underutilized in event-based frameworks. This study develops and evaluates a multi-factor coseismic landslide probability model that integrates InSAR-derived deformation metrics with geomorphic and hydrologic predictors to support rapid post-earthquake hazard assessment. Using the 25 April 2015 Mw 7.8 Gorkha earthquake as a case study, LOS displacement was derived from ALOS-2 PALSAR-2 ScanSAR interferometry, and the normalized channel steepness index (K_sn) was computed from a digital elevation model. Fourteen conditioning factors were used to train five architectures: Random Forest (RF), XGBoost, CNN, U-Net, and DeepLabV3. Spatial autocorrelation was mitigated using a leave-one-basin-out three-fold spatial cross-validation strategy, with models evaluated on a patch-based domain comprising 655,360 pixels at a positive-class prevalence of 6.35%, establishing a no-skill AUC-PR baseline of 0.0635. InSAR integration consistently improved model performance under high class imbalance, increasing AUC-PR across all models by 7.8% to 17.3%. Random Forest achieved the highest AUC-PR (0.7940, nearly 12.5 times the baseline) and CSI (0.3027), providing the best balance between landslide recall (88.09%) and non-landslide specificity (88.68%) with the lowest false alarm rate (11.32%). XGBoost attained the highest AUC-ROC (0.9501) but exhibited lower recall (83.73%) and poorer calibration (Brier = 0.1397). Among DL models, DeepLabV3 produced the best-calibrated probabilities (Brier = 0.0693) and the highest CSI (0.2307), while U-Net offered the most balanced DL performance and CNN achieved the highest recall (92.40%) at the expense of elevated false alarms. Permutation feature importance identified K_sn as the dominant predictor, highlighting the strong tectono-geomorphic control on coseismic landslide occurrence. These results demonstrate that integrating InSAR-derived products substantially enhances landslide hazard assessment and supports more reliable rapid response in the Nepal Himalaya. Full article

Sentiment analysis of online game reviews has increasingly become important in understanding player experiences and supporting data-driven game development. However, research in this domain has continuously faced two unresolved challenges: (1) the extreme imbalance between positive and negative feedback, and (2) the inefficiency of existing feature-weighting schemes in capturing sentiment signals embedded in informal gaming discourses. Prior works demonstrated that negative feedback—though a few in number are highly influential—usually contain richer emotional content and longer textual structures; yet, prevailing classification models often perform poorly for these minorities (i.e., negative feedback). Numerous studies explored multimodal imbalance issues, class imbalance in cross-lingual ABSA (Aspect-Based Sentiment Analysis), reinforcement-learning-based architectures for imbalanced extraction tasks, and oversampling strategies like SMOTE (Synthetic Minority Over-sampling Technique) variants. Few investigations specifically addressed imbalanced sentiment classification in the contexts of online game reviews, where user-generated content exhibits unique lexical, structural, and emotional characteristics. To address these gaps, this study integrated TF-IDF (Term Frequency-Inverse Document Frequency), VADER (Valence Aware Dictionary and Sentiment Reasoner) lexicon features, and IGM (Inverse Gravity Moment) weightings with advanced oversampling methods such as ADASYN (Adaptive Synthetic Sampling Approach for Imbalanced Learning) and Borderline-SMOTE to improve the detection of minority sentiment classes. Ensemble models, including XGBoost (Extreme Gradient Boosting) and LightGBM (Light Gradient-Boosting Machine), were further employed to enhance the robustness of imbalance. Using a large-scale dataset of Steam game reviews, the proposed framework demonstrated substantial improvement in identifying negative sentiments, addressing a critical limitation in the existing computational game-analysis literature, and advancing the modeling for detecting the emotion-rich but imbalance-prone user feedback. Full article

To address face occlusion, low detection rates of small-scale faces, and sample imbalance in dense visual scenarios, we propose a YOLOv7-based detector with four key improvements: (1) an optimized MPConv module to enhance feature extraction; (2) a novel CFPM to boost sensitivity to occluded samples; (3) an integration of the DyHead block in IDetect to mitigate feature loss from sample imbalance; (4) an SW-SCE loss function with a dual-input network to better detect small faces. Experiments on the WiderFace dataset show that our method improves detection performance by 1.2%, 1.8%, and 3% on the easy, medium, and hard subsets over the baseline. These gains strengthen face detection in dense, challenging environments with heavy occlusion and small-scale targets. Full article

With the explosive growth of the Internet of Things (IoT) and cloud-computing traffic, Intrusion Detection Systems (IDSs) have become a cornerstone of network security. However, modern traffic data often exhibits extreme class imbalance and long-tailed distributions, leading to persistently high miss rates for minority attack categories in Machine Learning (ML)-based IDSs. Conventional oversampling may introduce decision noise, whereas standard Generative Adversarial Networks (GANs) can suffer from training instability and mode collapse when modeling high-dimensional tabular traffic features. To address these challenges, we propose a high-fidelity traffic augmentation framework based on an Adaptive Hybrid-loss Conditional GAN (AH-CGAN). Specifically, AH-CGAN introduces an iteration-dependent adaptive gradient penalty (AGP) schedule to enforce the Lipschitz continuity constraint more effectively during training and incorporates a feature-matching objective to align intermediate critic representations between real and synthetic traffic. Experiments on the CIC-IDS2017 benchmark show that AH-CGAN generates distribution-consistent synthetic samples and that augmentation improves downstream detection across multiple classifiers. In particular, the weighted F1-score of Logistic Regression increases from 0.8237 to 0.8697 (Δ = +0.0460, i.e., +4.6%). Overall, the proposed approach enhances minority coverage in the feature space and can improve class separability, providing a practical solution for long-tailed IDS. Full article

The rapid integration of artificial intelligence (AI) into educational systems is transforming how student performance is analysed and how educational policies are informed by large-scale data. Within this context, machine learning techniques are increasingly used to identify patterns associated with academic success and educational inequality. However, the use of predictive algorithms in education also raises important questions regarding transparency, fairness, and potential algorithmic bias. This study examines the predictive performance and fairness implications of machine learning models used to identify academically resilient students using data from the Programme for International Student Assessment (PISA) 2022. The analysis is based on a dataset containing more than 600,000 student observations across multiple national education systems. Academic resilience is operationalised following the OECD framework, identifying students who belong to the lowest quartile of the socioeconomic status index (ESCS) within their country while simultaneously achieving mathematics performance in the top quartile (PV1MATH). A predictive framework incorporating six supervised learning algorithms—Logistic Regression, Random Forest, Gradient Boosting, XGBoost, LightGBM, and CatBoost—was implemented. The modelling pipeline includes data preprocessing, missing value imputation, class imbalance correction using SMOTE, and model evaluation through multiple classification metrics, including accuracy, F1-score, and the area under the ROC curve (AUC). In addition, fairness diagnostics are conducted to examine potential disparities in prediction outcomes across gender groups, while feature importance analysis and SHAP-based explanations are used to interpret the contribution of key predictors. The results indicate that ensemble-based models achieve the highest predictive performance, particularly those based on gradient boosting techniques. At the same time, the analysis reveals that socioeconomic status, migration background, and school repetition constitute the most influential predictors of academic resilience. Although gender displays relatively low predictive importance, measurable differences in positive prediction rates across gender groups suggest the presence of potential algorithmic disparities. These findings highlight the importance of integrating fairness evaluation, transparency, and interpretability into educational data science workflows. The study contributes to ongoing discussions on the responsible use of artificial intelligence in education by emphasising the need for governance frameworks capable of ensuring that algorithmic systems support equity-oriented educational policies. Full article

Sleep plays a fundamental role in human health and cognitive functioning, motivating the development of reliable and scalable methodologies for sleep stage classification (SSC). Recent advances in non-invasive and economically sustainable sensing technologies enable continuous sleep monitoring beyond laboratory settings. However, SSC remains a challenging data analytics task due to the intrinsic class imbalance among sleep stages. This study investigates the effectiveness of different imbalanced data management strategies within a machine learning framework for non-invasive SSC. The proposed approach relies exclusively on heart rate and motion signals, which can be acquired through wearable devices or contactless under-mattress sensors, making it suitable for longitudinal monitoring scenarios. Using the PhysioNet DREAMT dataset, 32 experimental scenarios are defined by combining data-level techniques (ADASYN oversampling with different balancing weights), algorithm-level strategies (cost-sensitive learning), and hybrid solutions. Four model families are evaluated—Decision Tree, k-Nearest Neighbors, Ensemble Classifiers, and Artificial Neural Networks—across classification tasks involving 2, 3, 4, and 5 sleep stages. The experimental results show that ensemble-based models provide robust and consistent performance under severe class imbalance, achieving macro accuracies of 82% for sleep–wake detection, 73% for 3-stage classification, 72% for 4-stage classification, and 64% for 5-stage classification. These findings confirm the relevance of imbalance-aware analytics and demonstrate the feasibility of accurate, minimally invasive SSC within big data and cognitive computing paradigms. Full article

Synthetic Aperture Radar (SAR) ship classification plays a critical role in maritime surveillance, addressing challenges such as the similarity between ship categories, as well as scarcity of annotated datasets and data imbalance. In this paper, a lightweight and real-time dual-branch architecture is proposed to effectively address the SAR ship classification task. The proposed approach integrates dual-polarization data within a hybrid convolution-transformer framework to improve classification performance. The model fuses dual-polarization modes, combining convolutional layers for local feature extraction with transformer blocks for global contextual understanding. Evaluations on the OpenSARShip 2.0 dataset show that the proposed model achieves 97.50% accuracy in the 3-class configuration and 93.28% in the 6-class configuration. For the FUSAR-Ship dataset, which does not provide dual-polarization data for the same ship target, the single branch model achieved an accuracy of 94.92% for the 7-class configuration. Despite its dual-branch design, the model maintains computational efficiency, making it suitable for real-time maritime monitoring applications. The results demonstrate the effectiveness of polarization-aware hybrid models for scalable and robust SAR ship classification. Full article

Extreme class imbalance is a persistent obstacle for machine learning-driven intrusion detection, as rare but high-impact cyberattacks occur far less frequently than benign traffic in training data. In many real-world cybersecurity datasets, this imbalance becomes extreme, with certain attack types containing a handful of samples, effectively placing the problem in a few-shot learning regime. This paper presents a controlled benchmarking study of Generative Adversarial Network (GAN) objectives for synthesizing minority-class cyberattack data. Using the UWF-ZeekData22 network traffic dataset, each MITRE ATT&CK tactic is framed as a separate binary detection task, and tactic-specific GANs are trained solely on minority samples to generate synthetic attack records. Four widely used GAN variants—Vanilla GAN, Conditional GAN (cGAN), Wasserstein GAN (WGAN), and Wasserstein GAN with Gradient Penalty (WGAN-GP)—are compared under unified training steps and fixed augmentation conditions. The utility of generated data is assessed by evaluating downstream detection performance using five traditional classifiers: Logistic Regression, Support Vector Machine, k-Nearest Neighbors, Decision Tree, and Random Forest. The results indicate that GAN augmentation generally strengthens minority-class detection across tactics and models, reducing false negatives and improving recall consistency, while not systematically harming majority-class performance. However, the effectiveness of each GAN objective varies significantly with data sparsity. Specifically, simpler adversarial objectives often outperform more complex architectures by preserving discriminative feature structure, while heavily regularized models may overly smooth minority-class distributions and reduce separability. Wasserstein-based objectives provide improved training stability, but additional regularization does not consistently translate to better detection performance. Overall, the results demonstrate that in extreme-imbalance settings, GAN effectiveness is governed more by data sparsity and structure preservation than by architectural complexity. These findings establish class-specific generative augmentation as a practical strategy for intrusion detection and provide empirical guidance for selecting appropriate GAN objectives for tabular cybersecurity data under highly imbalanced conditions. Full article

Accurate seamount identification is important for understanding submarine tectonic and magmatic processes and for supporting deep-sea geomorphological analysis. However, seamount recognition faces a severe class imbalance as abyssal plains constitute the majority of deep-sea topography while seamounts occupy only a minimal portion, which makes accurate segmentation difficult. To address this issue, this study proposes an improved U-Net architecture, termed Spatial–Channel Reconstruction U-Net (RSCU-Net), built upon a Residual Spatial–Channel Reconstruction Convolution (Res-SCConv) module. The Res-SCConv module is embedded into each skip connection of the U-Net architecture. The model combines a Spatial Reconstruction Unit (SRU) and a Channel Reconstruction Unit (CRU) to suppress dominant background interference and reduce channel redundancy, and further introduces a Selective Kernel-based Multi-scale Gradient Module (SK-MGM) to improve boundary refinement. Experiments on the GEBCO 2023 bathymetric dataset, including 696 training samples and 88 independent test samples, show that RSCU-Net achieves an Accuracy of 0.938, Recall of 0.833, F1-score of 0.720, and IoU of 0.563. Compared with the baseline U-Net, Recall improves from 0.741 to 0.833 and IoU from 0.405 to 0.563. Additional validation on the Suda Seamount dataset yields an Accuracy of 0.987, F1-score of 0.958, and IoU of 0.920, demonstrating the robustness and generalization capability of the proposed method. Full article

Class imbalance is a persistent challenge in supervised machine learning, particularly in biological datasets where minority classes represent functionally critical categories. Synthetic data generation has emerged as a principal strategy for mitigating this problem, yet systematic comparisons of classical and modern deep generative approaches remain limited. This study presents a comprehensive benchmark evaluation of four synthetic data generation methods—SMOTE, CTGAN, TVAE, and TabDDPM—across two well-established biological datasets from the UCI Machine Learning Repository: the E. coli protein localization dataset (307 samples, 6 features, 4 classes) and the yeast protein localization dataset (1299 samples, 8 features, 4 classes). Synthetic data quality was rigorously assessed using a multi-dimensional evaluation framework encompassing distributional fidelity (Fréchet Distance, Wasserstein Distance), machine learning utility (Train-on-Synthetic-Test-on-Real and Train-on-Real-Test-on-Real protocols using XGBoost version 3.2.0, Logistic Regression, Support Vector Machines, and Random Forest), and distinguishability (Classifier Two-Sample Test). The datasets are rather imbalanced. During the experiments, the dataset size increased to three times its original size while preserving the imbalanced class-sample ratio. To evaluate the quality of synthetic data, the max(AUC,1−AUC) score is proposed. This score is inversely proportional to classification performance, indicating that synthetic data are not easily distinguishable from real data. Per-class analysis reveals that minority classes remain the primary challenge across all generative methods. SMOTE and TabDDPM obtained the highest predictive utility F1-scores across both datasets. TVAE offers the strongest distributional fidelity among deep generative models, producing synthetic samples that are most difficult to distinguish from real data (lowest C2ST scores). CTGAN exhibits significant performance degradation on both small- and medium-scale datasets, with F1 utility ratios below 0.50. Full article