Search Results (1,203)

Early and accurate analysis of breast cancer is critical for improving patient outcomes. Ultrasound imaging is widely used for breast tumor screening due to its safety, accessibility, and low cost. We propose DGTN (Diffused Graph-Transformer Network), a lightweight, end-to-end deep learning model that jointly performs breast tumor segmentation and multi-class classification (benign, malignant, and normal) from ultrasound images. DGTN integrates Graph Convolutional Networks (GCNs) and Transformer encoders through a bidirectional diffusive attention mechanism and a learnable gating strategy, enabling structured spatial information and global contextual features to co-evolve. We evaluate DGTN on the public BUSI breast ultrasound dataset using balanced sampling and a joint cross-entropy and Dice loss. The model achieves 68.8% classification accuracy and a Dice score of 0.6227. While its performance remains below that of recent state-of-the-art pipelines, DGTN offers a favorable trade-off between accuracy and computational efficiency within a single unified framework. Ablation experiments indicate that both diffusive attention and gating contribute meaningfully to performance (paired t-test across five cross-validation folds, p < 0.05, with large paired effect sizes, d ≈ 1.0–1.4); because this test is based on only five folds, the result should be interpreted as indicative rather than conclusive, and we report it alongside fold-level effect sizes rather than as a stand-alone confirmation of significance. To the best of our knowledge, this work represents one of the first applications of diffusive graph-transformer co-learning to breast ultrasound imaging, demonstrating the potential of graph-enhanced attention for efficient multi-task medical image analysis. Full article

Intrusion Detection Systems (IDSs) based on traditional machine learning treat network flows as independent tabular samples, ignoring the relational and topological structure that characterizes modern distributed attacks. Graph Neural Networks (GNNs) overcome this limitation by modeling network topology, which in turn raise the need to make their predictions transparent. This work develops and compares traditional classifiers against a GNN-based IDS on the UNSW-NB15 dataset, for both binary and multiclass classification. A novel graph construction is proposed in which each node is an individual flow and edges are defined by temporal proximity through three complementary strategies (conversation chains and temporal k-NN by source and destination IP). Three GNN backbones—GraphSAGE, Graph Convolutional Network (GCN) and Graph Attention Network (GAT)—are trained under an identical, matched pipeline and a chronological, inductive evaluation protocol, so that any difference is attributable to the backbone alone. A two-stage classifier then separates detection from attack-type categorisation, with GNNExplainer providing interpretability, and SHAP applied to the traditional models. In binary classification, GraphSAGE achieves an Accuracy of 0.9906, Precision of 0.9856, Recall of 0.9998, F1-Score of 0.9927 and ROC-AUC of 0.9965, exceeding the traditional baselines in their conventional evaluation setting, while GCN and GAT reach comparable detection (F1 ≈ 0.99), showing that the temporal graph rather than the specific backbone drives detection. The explainability analysis identifies TTL-related and connection-state variables as dominant predictors and reveals attack-specific structural patterns, confirming that temporally structured GNNs improve detection while providing interpretable predictions. Full article

Hepatocellular carcinoma (HCC) is one of the most prevalent cancers worldwide and exhibits considerable biological heterogeneity in both molecular and clinical characteristics. The diverse molecular alterations and clinical manifestations of HCC indicate substantial heterogeneity across patient subgroups. This study aimed to identify novel therapeutic targets and predictive biomarkers associated with HCC using an integrative bioinformatics approach. High-throughput genomic datasets were obtained from the UCSC Xena browser to retrieve mRNA HTSeq-count data from the TCGA-HCC cohort. Gene co-expression network (GCN), protein–protein interaction network (PPIN), and enrichment analyses were performed to identify key dysregulated genes and their biological significance. Integrated network analyses identified three dysregulated hub genes, namely CXCR2, TLR2, and TLR4. Genomic alterations in these genes were further evaluated across tumor samples in the TCGA-HCC cohort. Kaplan–Meier (KM) survival analysis demonstrated that lower CXCR2 mRNA expression was significantly associated with poorer overall survival (OS) and recurrence-free survival (RFS). Furthermore, TIMER and UALCAN analyses revealed significant associations between CXCR2 expression and tumor purity, as well as immune cell infiltration levels, including T cells, macrophages, dendritic cells (DCs), and neutrophils. These findings suggest that CXCR2 is significantly associated with the immune microenvironment of HCC and represents a potential prognostic biomarker whose biological role warrants further mechanistic investigation. Full article

In this research, we propose a graph-based approach to classify ECG signals. We model the ECG signal as a time-varying graph and examine its dynamics at two levels: intra-beat and inter-beat. The proposed framework employs a two-level graph representation of the ECG. At the intra-beat level, each beat is modeled as a graph in which the P, Q, R, S, and T waves serve as nodes, and temporal distance information is incorporated into the node feature vector. At the inter-beat level, the ECG signal is represented as a graph of beats. We investigated two node representations: learnable MLP embeddings derived from handcrafted features and reduced features obtained via PCA. Unlike conventional temporal GCNs, graph attention networks (GATs), and transformers, which predominantly rely on single-level representations of ECG signals, the proposed approach builds hierarchical graph structures directly from ECG signals. The dual model was trained on 1024 ECG segments from the PTB Diagnostic ECG Database (PTBDB) and tested on 398 ECG segments from the MIT-BIH Arrhythmia Database to evaluate its performance. The results indicate that the best-performing configuration, which employs PCA-reduced node features, outperforms the other configurations tested. The model achieves a mean accuracy of $97.67\%$ and a mean F1-score of $97.18\%$ over five runs. Full article

Complex inter-subject variability induces severe distribution shifts in the physiological features of electroencephalography (EEG) for air traffic controllers (ATCOs). These inter-subject shifts limit the generalization and interpretability of passive brain–computer interfaces (pBCIs) during cognitive workload decoding. To address this, a Dynamic Microstate-Guided Graph Convolutional Network (DMG-GCN) is proposed for robust cross-subject workload recognition. This approach utilizes a Dynamic Selective Kernel Temporal Convolutional Block (DSK-TCB) to adaptively extract multi-scale temporal–spectral dynamics, while concurrently constructing a time-evolving adjacency matrix via a Microstate-Guided Dynamic Graph Block (MG-DGB) to disentangle topological sub-networks. A spatiotemporal graph convolution module then aggregates these representations, and a temporal self-attention mechanism focuses on task-critical transition moments. Extensive experiments on simulated multi-level air traffic control tasks demonstrate that the proposed model achieves an overall average accuracy of 80.30% and an average F1-score of 78.63% in cross-subject evaluations, significantly outperforming state-of-the-art baselines. Moreover, an exploratory interpretability analysis suggests that the extracted topological sub-networks exhibit spatial patterns consistent with specific brain network reorganizations, which encompass the transition from global distributed monitoring to temporal multimodal integration and parietal–occipital parallel processing during workload regulation. The framework provides a robust and analytically transparent pBCI solution for adaptive automation in modern aviation. Full article

This study establishes a GCN–NSGA-III-based framework for determining gas-leak sensor locations. Candidate layouts are optimized simultaneously with respect to installation expenditure, spatial coverage, leak-identification performance, and time to alarm. In the proposed framework, the GCN extracts spatial correlations and leakage-risk features among candidate monitoring locations, whereas NSGA-III optimizes the network weights and thresholds to support the selection of improved sensor placement schemes. By combining the image-based spatial feature extraction capability of CNNs, the graph-structured feature learning capability of GCNs, and the multi-objective optimization strength of NSGA-III, the model achieves significant improvements in detection accuracy and risk assessment efficiency. The model was validated using a hybrid gas-leak dataset comprising 758 training samples, including 189 real-world monitoring samples and 569 simulated samples, and 229 test samples, including 73 real-world monitoring samples and 156 simulated samples. Its engineering applicability was further evaluated using a simulated chlorine leakage scenario at a chemical plant in Changzhou, China, with a leakage rate of 2 kg/s and an ambient easterly wind speed of 1 m/s. Experimental results confirm its reliability and practical applicability in real-world engineering contexts. Compared with the original pre-optimization sensor layout under the same leakage and environmental conditions, the proposed optimization strategy reduces deployment costs by 19%, increases monitoring coverage by 8.2%, improves detection accuracy by 14.1%, and shortens alarm response time by approximately 15%. Full article

Circular RNAs (circRNAs) participate in gene regulation and disease progression, but experimental mapping of circRNA-disease associations (CDAs) remains costly and incomplete. We developed NARVGA, a hybrid prediction framework that combines non-negative matrix factorisation (NMF) with an adversarially regularised variational graph autoencoder (ARVGA). Functional, semantic and Gaussian interaction-profile kernel similarities are integrated; K-means-derived co-membership graphs reduce diffuse similarity connections; ARVGA learns nonlinear topological embeddings; and NMF captures complementary low-rank association patterns. An Extra Trees classifier then scores candidate circRNA-disease pairs. In the original transductive stratified five-fold benchmark on CircR2Disease, NARVGA achieved an area under the receiver operating characteristic curve (AUC) of 0.9868, an area under the precision–recall curve (AUPR) of 0.9887, 94.32% accuracy and a 94.44% F1-score. Ablation analysis identified cluster sparsification as the largest individual contributor, with additional gains from adversarial regularisation and low-rank augmentation. Without task-specific retuning, the model obtained AUCs of 0.9503 on LncRNADisease and 0.9700 on HMDDv4. Performance was stable across GCN depths and cluster settings, and 19 of the 20 highest-ranked hepatocellular carcinoma candidates had supporting published evidence. NARVGA provides a within-network prioritisation tool for RNA-disease association studies, although prospective validation, hard-negative assessment and independent-cohort testing remain necessary. Full article

Graph Neural Networks achieve strong closed-set accuracy in network intrusion detection but cannot flag zero-day attacks, because the closed-world assumption forces every input into a known class. OpenMax adds an Extreme Value Theory reject option, yet its Euclidean distance ignores the class-conditional covariance that encodes attack-specific structure, which produces unreliable tail models and rejection thresholds. We propose MaGOS-IDS, which extracts topology-aware embeddings with an edge-aware GCN that fuses flow-level edge features directly into message passing, whitens each class with a regularized Mahalanobis distance so the reject decision respects per-class variance and correlation, and calibrates a per-class EVT tail on these distances to set an attack-pattern-aware rejection boundary without a hand-tuned cutoff. We provide a theoretical justification via a peaks-over-threshold argument: whitening removes the per-class covariance dependence of the distance tail, so a single extreme-value tail model calibrates consistently across classes. On three benchmarks (NF-BoT-IoT, CIC-IDS-2017, UNSW-NB15) under withheld zero-day families, MaGOS-IDS raises open-set AU-PR over the Euclidean OpenMax baseline (0.932 vs. 0.848 on UNSW-NB15) while adding negligible inference cost. Full article

Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system characterized by inflammatory demyelination, neurodegeneration, and progressive disability. Clinically isolated syndrome (CIS) often represents the first overt presentation of MS. We performed an exploratory cross-sectional transcriptomic investigation of peripheral blood mononuclear cells (PBMCs) from healthy controls (HC, n = 40), CIS patients (n = 49), and Relapsing-Remitting MS (RRMS, n = 53) patients using the ArrayExpress dataset E-MTAB-11415. Differential expression analysis was performed using limma, adjusting for age and sex. Functional enrichment and network analyses were conducted using clusterProfiler, STRING, and Cytoscape. Although Principal Component Analysis (PCA) showed partial overlap among groups, pathway-level analyses revealed coherent alterations in translation, ribosome biology, mitochondrial protein synthesis, and stress-response regulation. In CIS compared with HC, cytoplasmic and mitochondrial ribosomal genes were predominantly downregulated, suggesting reduced translational capacity in PBMCs. This was accompanied by altered expression of translation-initiation and transcriptional regulators, whereas genes involved in stress-adaptive translational control, including GCN1, YARS1, QARS1, and SIL1, were selectively upregulated. Conversely, RRMS compared with CIS showed upregulation of ribosome-related and biosynthetic programs. Overall, these findings suggest that CIS may be characterized by peripheral translational restraint and adaptive stress response activation, whereas RRMS progression is associated with biosynthetic reactivation. Full article

Accessibility-oriented sign language recognition plays an important role in intelligent human–computer interaction, but cross-subject recognition remains challenging because test signers are unseen during training. Variations in appearance, motion style, and modality quality among different signers introduce significant distribution differences between training and testing samples, limiting model generalization. Although multimodal approaches commonly assume that RGB, skeleton, and depth modalities provide complementary information, their effectiveness under cross-subject recognition remains insufficiently explored. This study investigates modality effectiveness in SLR500 word-level sign language recognition under a cross-subject protocol. A unified evaluation framework is established to compare RGB-only, skeleton-only, RGB–skeleton fusion, and RGB–skeleton–depth fusion under consistent data splits, preprocessing procedures, training settings, and evaluation metrics. Sample-level prediction transitions are further analyzed to examine how fusion changes recognition outcomes relative to the skeleton-based reference model. The results show that the skeleton modality provides more reliable recognition performance under the evaluated cross-subject setting, while the tested fusion strategies do not consistently improve over the skeleton-only configuration. Based on the modality-effectiveness analysis, a skeleton-based reference model is established using a Spatial–Temporal Graph Convolutional Network (ST-GCN)-based spatio-temporal encoder, learned confidence-weighted temporal pooling, and a lightweight classifier. Multi-seed experiments further verify the stability of the model, which achieves 86.37% Top-1 accuracy and 96.86% Top-5 accuracy on the SLR500 cross-subject test set. These results emphasize the importance of evaluating modality reliability before designing complex fusion strategies for cross-subject sign language recognition. Full article

Power cable tunnels are typical enclosed cable-routing spaces in which cable overheating, insulation aging, and partial discharge may gradually develop into fire hazards. During the early stage of cable fires, abnormal sensor responses are often weak, localized, and continuously evolving, which increases the difficulty of early warning based on a single sensor or a single temporal feature. Motivated by cable fire early warning in power cable tunnels, this study uses cable-fire records from a publicly available indoor EN 54 fire-test-room dataset with distributed multi-sensor nodes to evaluate the proposed model under controlled laboratory conditions. In monitoring scenarios with fixed sensor nodes, temporal-only modeling methods often struggle to simultaneously characterize short-term variations, temporal evolution, and spatial differences in node responses. To address this limitation, this study proposes a multi-sensor spatiotemporal feature fusion model that integrates a gated recurrent unit (GRU), a Modern Temporal Convolutional Network (ModernTCN), and an enhanced graph convolutional network (GCN+). The proposed model adopts the ModernTCN as the temporal modeling backbone. A GRU module is introduced at the front end to encode local fluctuations and short-term continuous changes between consecutive time steps, while GCN+ is embedded at the intermediate feature stage of the backbone to model spatial correlations and cross-node coordinated responses among fixed sensor nodes. Experimental results show that the proposed model achieves strong classification performance in the cable fire early warning discrimination task, with a test accuracy of 0.9884 and a false negative rate (FNR) reduced to 0.0150. The comparative experimental results indicate that the proposed model achieves better overall performance than typical temporal baseline models. The ablation study further shows that, under the experimental settings of this study, the introduction of GRU and GCN+ leads to overall improvements in the main evaluation metrics, suggesting that both modules provide a certain enhancement to the cable fire early warning discrimination performance of the ModernTCN backbone. Full article

Advances in artificial intelligence have made hand gesture recognition an important human–computer interaction modality. Graph convolutional networks (GCNs) are widely used for skeleton-based hand gesture recognition, yet their performance can be limited by weak semantic topology modeling, underused feature channels, and shallow spatio-temporal fusion. We propose a Multi-scale Attention Graph Convolutional Network (MA-GCN) that combines three components within one skeleton framework: a hybrid topology that augments physiological connections with semantic priors; a Gaussian Multi-Scale Channel Attention (GMCA) module for coordinate denoising and adaptive channel weighting; and a Local-Global Fusion Module (LGFM) that combines local convolutional features with channel-wise global attention. Ablation studies quantify the independent and joint contributions of these components. MA-GCN obtains Top-1 accuracies of 97.50%/95.95% on SHREC’17 Track and 94.29%/92.86% on DHG14/28 for the 14-/28-class settings. In a SHREC’17 Track-to-FPHA pre-train-then-fine-tune evaluation, it reaches 94.09% Top-1 accuracy, providing preliminary evidence that the proposed framework maintains effectiveness under cross-dataset transfer. Full article

To address the challenges of integrating multi-source heterogeneous data, fragmented fault knowledge, and the limited capability of traditional rule engines in recognizing edge cases for ultra-high voltage (UHV) bushing fault diagnosis, this paper proposes a fault identification and decision-making method based on knowledge graph (KG) rule reasoning and inductive graph convolutional network (Inductive GCN). First, a triple-matching strategy is employed to perform entity extraction and relation mining from fault cases, constructing a fault knowledge graph that transforms unstructured fault case texts into a structured knowledge graph. Second, a rule engine based on a multi-source feature rule set is designed, utilizing the entropy weight method and the RETE algorithm to achieve interpretable symbolic reasoning. On this basis, a double-layer inductive graph convolutional network is introduced to learn implicit fault patterns by aggregating topological information from neighboring nodes, and a confidence-driven dynamic weighted fusion strategy is adopted to achieve complementary advantages between the two models. Finally, a large language model is introduced to generate operation and maintenance decision recommendations. Experimental results demonstrate that the proposed method achieves an identification accuracy of 98.1% on a test set of 159 samples, which is 10.7 percentage points higher than that of a single rule engine and 6.9 percentage points higher than that of a single inductive graph convolution network. The standard deviation of accuracy across different test batches is only 0.0029. These results demonstrate the effectiveness and stability of the proposed method, providing a practical technical solution for UHV bushing fault identification. Full article

The Internet of Medical Things (IoMT) is a promising technology for enabling smart and efficient healthcare systems through continuous physiological monitoring, early anomaly detection, and proactive patient management. However, IoMT sensors are vulnerable to False Data Injection Attacks (FDIAs), in which adversaries can manipulate sensor measurements to compromise diagnostic accuracy, mislead clinical decision-making, and threaten patient safety. Existing detection approaches often rely on population-level statistical models that may not fully capture individual physiological variations or residual-based thresholds designed for relatively simple attack scenarios, limiting their ability to exploit the spatio-temporal dependencies of multi-sensor physiological streams and detect stealthy or adversarial FDIAs. This paper proposes a patient-specific FDIA detection framework based on a Graph Convolutional Network–Gated Recurrent Unit (GCN–GRU) digital twin that learns an individual patient’s normal physiological behaviour from clean baseline telemetry. The trained digital twin is integrated into a Kalman filter as the state prediction model, and the resulting standardised innovation residuals are used as detection features. To characterise stealthy attack behaviours, four complementary window-based feature groups are extracted from the innovation sequence: innovation statistics, sensor correlation drift, temporal smoothness, and uncertainty mismatch. A CNN-1D classifier is then trained to learn discriminative temporal attack patterns from these features for accurate detection. A structured attack taxonomy comprising five stealthy and adversarial FDIA scenarios is developed, where attacks are injected as smooth gradual or abrupt coordinated modifications to sensor measurements while remaining within plausible physiological ranges. Experiments conducted on the WUSTL-EHMS-2020 benchmark dataset demonstrate that the proposed framework achieves an F1-score of 94.3%, outperforming Isolation Forest and PCA Reconstruction by 34 percentage points. Furthermore, the proposed framework reduces the false alarm rate to 3.6%, compared with 35.1% and 9.2% achieved by Isolation Forest and PCA Reconstruction, respectively. These results demonstrate the effectiveness of the proposed framework for reliable detection of stealthy FDIAs in IoMT-based healthcare systems. Full article

Dynamic link prediction, which aims to infer future edges from historical network structures, is a fundamental task in dynamic network analysis. Traditional models fail to capture high-order information, while existing methods neglect the distinct temporal evolution patterns between low-order and high-order structures, thereby limiting prediction accuracy. To address these issues, we propose DC-GHCN, a dynamic link prediction model based on a dual-channel architecture that integrates Graph Convolutional Network (GCN) and Hypergraph Convolutional Network (HGCN). Firstly, we extract closed motifs from dynamic network snapshots to construct an initial hypergraph, then refine it via nested motif pruning and node weight compensation strategies. Secondly, we design a dual-channel architecture: the GCN channel learns low-order structural features, while the HGCN channel learns high-order structural features. Furthermore, two independent Gated Recurrent Units (GRUs) separately model the temporal evolution of the two channels. Finally, the model employs a gating mechanism to adaptively fuse the dual-channel node representations for link prediction. Experiments on five real-world dynamic network datasets demonstrate that DC-GHCN outperforms baseline models, validating the effectiveness of the proposed model in dynamic link prediction. Full article