Search Results (47)

Future industrial human activity recognition (HAR) may require discrimination among many operational elements, including movements with partially overlapping kinematics. This controlled pilot study characterized recognition across 22 upper-body movement classes with varying structural similarity and examined classifier, signal scaling, and temporal window length. Sixteen adults performed the movements with an XSENS motion-capture system; eight upper-limb inertial measurement units provided 80 synchronous time-series channels. Thirteen participants were used for model development, and three whose observed execution patterns differed comparatively from the remainder were deliberately reserved as a challenge-oriented holdout. Support vector classifier (SVC), random forest (RF), Gaussian naive Bayes (NB), and long short-term memory (LSTM) models were evaluated with min–max or maximum-absolute scaling and 62-, 93-, or 124-frame windows. RF with maximum-absolute scaling and a 124-frame window achieved the best aggregate holdout performance (accuracy = 0.950; F1 = 0.939). Importantly, this performance was obtained on three entirely unseen participants who were deliberately reserved because their observed execution patterns and fluency differed from those of the model-development participants, providing a controlled, challenge-oriented test of transfer across inter-individual execution variability. Nevertheless, strong aggregate performance did not translate into uniform class-level reliability, and prominent errors remained concentrated in specific movement pairs. These findings provide empirical evidence for both the participant-independent transfer capability and the class-specific limitations of motion-only recognition, supporting its role as a methodological precursor to future AI-assisted work study and human–robot collaboration rather than as evidence of end-to-end industrial HAR. Full article

With the escalating frequency of sophisticated zero-day attacks, overcoming the critical limitations of signature-based Intrusion Detection Systems (IDSs) has become paramount. This study proposes a hybrid multi-layered intrusion detection framework combining traditional machine learning, Deep Neural Architectures (DenseNN), and ensemble methods to evaluate zero-day resilience within cloud-level backend connectivity interfacing EV and V2X management ecosystems. Using the comprehensive CSE-CIC-IDS2018 benchmark as a surrogate environment, a code-executed Leave-One-Attack-Out (LOAO) cross-validation protocol across 13 distinct attack families was implemented to assess unseen-attack-family generalization within the benchmark to unseen threats. Furthermore, Explainable Artificial Intelligence (XAI) auditing, utilizing SHapley Additive exPlanations (SHAP) and Integrated Gradients, was integrated to inspect decision boundaries and resolve feature-attribution failure modes. Critically, the audit identified an artifact-driven data leakage caused by the Timestamp and identifier features, demonstrating that models learned temporal schedules rather than behavioral network signatures. Re-executing all experiments post-leakage removal quantified performance drops across all classifiers (e.g., Gaussian NB dropping by up to 20.88 percentage points in accuracy (at the 60% training ratio; 18.30 points at the 80% ratio)). Under standard binary classification metrics, tree ensembles (Random Forest and Extra Trees) achieved high in-distribution detection (F1 > 0.95) with rapid inference latency (≈0.05–−0.07 ms/sample). However, the rigorous LOAO evaluation revealed a substantial generalization penalty on truly unseen zero-day families (e.g., SQL Injection and Infiltration), where simpler linear models demonstrated broader generalization robustness (mean LOAO F1 = 0.397) compared with complex tree-ensemble models. By rectifying dataset leakage and benchmarking deployment trade-offs (training runtime, throughput, and memory footprint), this study delivers actionable, transparent guidelines for deployment-oriented IDS evaluation in dynamic network infrastructures. Full article

Deep neural networks have achieved substantial success in image, text, and signal analysis, but their advantage is less consistent for heterogeneous tabular data, where tree-based ensemble methods often remain strong baselines. This study proposes CANON (Cross-Attention Neuro-symbolic Ontology Network), a neuro-symbolic architecture that integrates a hierarchical a priori concept ontology with a modular mixture-of-experts mechanism. CANON is designed to combine data-driven representation learning with explicit domain structure and to reduce the influence of irrelevant or weakly informative features. The architecture was evaluated on two clinical tabular cohorts—848 patients with ANCA-associated vasculitis described by 142 features, and 200 patients assessed for coronary artery stenosis described by 593 features—using stratified 8-fold cross-validation, and was compared with tree-based ensembles, dedicated tabular deep learning models, and classical neural architectures. CANON achieved the highest mean AUC on both cohorts (0.9426 and 0.9484). Two findings are reported. First, CANON significantly outperformed in AUC all four non-tree baselines included in the paired statistical analysis: all eight paired comparisons against GaussianNB, FT-Transformer, TabNet and LSTM were significant on both cohorts and favored CANON in 8 of 8 folds, with FT-Transformer and TabNet tuned separately for each cohort under an identical budget of 20 Optuna trials each. Second, CANON performed comparably to the tree-based ensembles, significantly outperforming Random Forest on the coronary artery stenosis cohort ($\Delta =+0.124$, $p=0.004$), while the remaining comparisons against Random Forest and XGBoost did not reach significance. An ablation study comprising 40 cross-validation folds per variant shows that the semantic organization of features into concepts and the non-linear feature tokenizer contribute measurably to the harder, higher-dimensional cohort, whereas the directional fusion mechanism does not improve predictive performance and is retained for its interpretability role. These findings indicate that ontology-guided neural architectures can provide a competitive and interpretable basis for clinical decision support and may also be useful in other regulated domains in which predictive models must remain consistent with domain-specific knowledge. Full article

Network intrusion detection systems (NIDSs) deployed in dynamic environments face concept drift from evolving attacks and traffic patterns, causing model reliability to degrade over time. Continual learning (CL) offers an adaptive solution, yet many methods misalign drift detection, memory updating, and optimization: drift is often judged with low-dimensional statistics, while adaptation occurs in representation space, limiting consistency under buffer constraints. To address concept drift in non-stationary network traffic and catastrophic forgetting during online intrusion detection updates, we propose a continual-learning framework built upon SSF that combines feature-space Gaussian-kernel Maximum Mean Discrepancy (MMD) drift detection with gradient-matching coresets for memory admission. The proposed framework retains strategic forgetting and steady-state distillation while replacing low-dimensional drift tests with feature-space MMD and mask-based selection with gradient-matching coresets, thereby improving incremental updates under a limited memory budget. On NSL-KDD and UNSW-NB15 under a unified multi-seed streaming protocol, the proposed method improves detection performance and knowledge retention. Experimental results demonstrate that gradient-matching coreset selection is the primary contributor to the observed performance improvements, while the effectiveness of MMD-based drift scheduling and strategic forgetting depends on the underlying data distribution and drift-trigger threshold. The proposed framework employs batch-level MMD scheduling to coordinate memory admission and online optimization, providing a practical path toward robust continual intrusion detection. Full article

Securing IoT networks presents fundamental challenges rooted in hardware constraints: firmware is often non-upgradeable and every security boundary is fixed at manufacture. Machine learning-based intrusion detection offers a scalable response, yet nearly all published systems assume clean training data and clean inference conditions. Production IoT environments satisfy neither assumption. Sensors degrade, packets drop, and adversaries deliberately corrupt telemetry streams to evade detection. The framework described here is built around that reality. The proposed framework is distinguished from prior work by four design decisions. First, three encoding branches, a residual DNN, a 1D-CNN, and a BiLSTM, are run in parallel and are fused by concatenation, each capturing structural patterns in tabular traffic data that the others miss. Second, a dual-view consistency loss trains the model under simultaneous feature masking and Gaussian noise, penalizing prediction divergence between two independently corrupted views of the same sample. Third, we introduce entropy-weighted attention: rather than fixed learned weights, per-feature importance is adjusted dynamically from information entropy measured across training batches, giving higher-entropy features stronger influence because they carry more discriminative variation. Fourth, branch-dropout regularization randomly silences entire branches during training, forcing each to develop independently useful representations instead of co-adapting. Class imbalance is handled through severity-aware loss weighting which scales contributions by the operational cost of missing each attack category, not purely by inverse frequency. On UNSW-NB15, the full model achieves 99.99% accuracy, 100% precision, 99.97% recall, and a false-negative rate of 2.65 × 10⁻⁴—the lowest across all compared architectures. Full article

Machine learning helps intrusion detection systems learn new assaults quickly. These systems train on a dataset with several threats and may identify odd behavior. This research detects intrusion using Random Forest, KNN, and Gaussian Naive Bayes. We run the model on a comprehensive dataset. Dynamics Feature Selector (DFS) improves performance. This technique eliminates unnecessary inputs and improves predictions using statistical analysis and feature significance. DFS effectiveness is tested using the NSL-KDD dataset. The recommended hybrid approach, Gaussian NB, Random Forest, and KNN are compared in meta-learning. Getting excellent accuracy with fewer characteristics is the aim. In order to demonstrate how the model may function in actual cybersecurity scenarios, the final test makes use of common performance metrics such as accuracy, precision, recall, and F1-score. The proposed method outperforms previously reported results with around 96.09% accuracy, 93.21% precision, 92.53% recall, 92.79% F1-score, and 93.65% average performance. Full article

Relaxor ferroelectric single crystals, such as Pb(In₁/₂Nb₂/₃)O₃–Pb(Mg₁/₂Nb₂/₃)O₃–PbTiO₃, possess extraordinary electro-optic (EO) coefficients, offering immense potential for next-generation integrated modulators. However, the application of PIN-PMN-PT in fiber-optic gyroscopes (FOGs) is hindered by the challenge of fabricating high-quality optical waveguides with strict mode selectivity, as conventional diffusion typically excites multi-mode propagation. Here, the fabrication of high-quality, mode-selective waveguides is achieved in rhombohedral PIN-PMN-PT via a nickel in-diffusion technique. The resulting graded-index structures exhibit a Gaussian profile with a maximum refractive index change (∆n) of 1.53% while preserving the single crystal structure. Under specific processing conditions, we achieve precise mode selectivity, enabling exclusive transverse electric (TE) mode transmission. This mode selectivity fulfills the requirements for single-mode Y-branch geometries, establishing a robust platform for ultra-compact, low driving voltage modulators and advancing the miniaturization of inertial navigation and integrated photonic systems. Full article

This study provides a detailed comparative analysis of a three-hybrid intrusion detection method aimed at strengthening network security through precise and adaptive threat identification. The proposed framework integrates an Autoencoder-Gaussian Mixture Model (AE-GMM) with two supervised learning techniques, XGBoost and Logistic Regression, combining deep feature extraction with interpretability and stable generalization. Although the downstream classifiers are trained in a supervised manner, the hybrid intrusion detection nature of the framework is preserved through unsupervised representation learning and probabilistic modeling in the AE-GMM stage. Two benchmark datasets were used for evaluation: NSL-KDD, representing traditional network behavior, and UNSW-NB15, reflecting modern and diverse traffic patterns. A consistent preprocessing pipeline was applied, including normalization, feature selection, and dimensionality reduction, to ensure fair comparison and efficient training. The experimental findings show that hybridizing deep learning with gradient-boosted and linear classifiers markedly enhances detection performance and resilience. The AE–GMM-XGBoost model achieved superior outcomes, reaching an F1-score above 0.94 ± 0.0021 and an AUC greater than 0.97 on both datasets, demonstrating high accuracy in distinguishing legitimate and malicious traffic. AE-GMM-Logistic Regression also achieved strong and balanced performance, recording an F1-score exceeding 0.91 ± 0.0020 with stable generalization across test conditions. Conversely, the standalone AE-GMM effectively captured deep latent patterns but exhibited lower recall, indicating limited sensitivity to subtle or emerging attacks. These results collectively confirm that integrating autoencoder-based representation learning with advanced supervised models significantly improves intrusion detection in complex network settings. The proposed framework therefore provides a solid and extensible basis for future research in explainable and federated intrusion detection, supporting the development of adaptive and proactive cybersecurity defenses. Full article

This study aims to investigate the predictive value of pre-treatment multi-phasic contrast-enhanced computed tomography (CECT) radiomic features for treatment resistance in patients with rat sarcoma virus (RAS)-mutated colorectal liver metastases (CRLMs) receiving bevacizumab-based chemotherapy. Seventy-three samples with RAS-mutated CRLMs receiving bevacizumab-combined chemotherapy regimens were evaluated. Radiomic features were extracted from arterial phase (AP), portal venous phase (PVP), AP-PVP subtraction image, and Delta phase (DeltaP, calculated as AP-to-PVP ratio) images. Three groups of radiomics features were extracted for each phase, including peritumor, core tumor, and whole-tumor regions. For each of the four phases, a two-sided independent Mann–Whitney U test with the Bonferroni correction and K-means clustering was applied to the remnant features for each phase. Subsequently, the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm was then applied for further feature selection. Six machine learning algorithms were then used for model development and validated on the independent testing cohort. Results showed peritumoral radiomic features and features derived from Laplacian of Gaussian (LoG) filtered images were dominant in all the compared machine learning algorithms; NB models yielded the best-performing prediction (Avg. training AUC: 0.731, Avg. testing AUC: 0.717) when combining all features from different phases of CECT images. This study demonstrates that peritumoral radiomic features and LoG-filtered pre-treatment multi-phasic CECT images were more predictive of treatment response to bevacizumab-based chemotherapy in RAS-mutated CRLMs compared to core tumor features. Full article

Local Interpretable Model-agnostic Explanations (LIME) is a widely used technique for interpreting individual predictions of complex “black-box” models by fitting a simple surrogate model to synthetic perturbations of the input. However, its standard perturbation strategy of sampling features independently from a Gaussian distribution often generates unrealistic samples and neglects inter-feature dependencies. This can lead to low local fidelity (poor approximation of the model’s behavior) and unstable explanations across different runs. This paper presents CoLIME, which is a copula-based perturbation generation framework for LIME, designed to capture the underlying data distribution and inter-feature dependencies more accurately. The framework employs bivariate (2D) copula models to jointly sample correlated features while fitting suitable marginal distributions for individual features. Furthermore, perturbation localization strategies were implemented, restricting perturbations to a defined local radius and maintaining specific property values to ensure that the synthesized samples remain representative of the actual local environment. The proposed approach was evaluated on a network intrusion detection dataset, comparing the fidelity and stability of LIME under Gaussian versus copula-based perturbations, using Ridge regression as the surrogate explainer. Empirically, for the most dependent feature pairs, CoLIME increases mean surrogate fidelity by 21.84–50.31% on the merged CIC-IDS2017/2018 dataset and by 29.28–60.24% on the UNSW-NB15 dataset. Stability is similarly improved, with mean Jaccard similarity gains of 3.78–5.45% and 1.95–2.12%, respectively. These improvements demonstrate that dependency-preserving perturbations provide a significantly more reliable foundation for explaining complex network intrusion detection models. Full article

Parkinson’s disease (PD) impairs balance and gait through neuromotor dysfunction, yet conventional assessments often overlook subtle postural deficits during dynamic tasks. This study evaluated the diagnostic utility of center-of-pressure (CoP) features captured by pressure-sensing insoles during the Timed Up and Go (TUG) test. Using 39 PD and 38 control participants from the recently released open-access WearGait-PD dataset, the authors extracted 144 CoP features spanning positional, dynamic, frequency, and stochastic domains, including per-foot averages and asymmetry indices. Two scenarios were analyzed: the complete TUG and its 3 m walking segment. Model development followed a fixed protocol with a single participant-level 80/20 split; sequential forward selection with five-fold cross-validation optimized the number of features within the training set. Five classifiers were evaluated: SVM-RBF, logistic regression (LR), random forest (RF), k-nearest neighbors (k-NN), and Gaussian naïve Bayes (NB). LR performed best on the held-out test set (accuracy = 0.875, precision = 1.000, recall = 0.750, F1 = 0.857, ROC-AUC = 0.921) using a 23-feature subset. RF and SVM-RBF each achieved 0.812 accuracy. In contrast, applying the identical pipeline to the 3 m walking segment yielded lower performance (best model: k-NN, accuracy = 0.688, F1 = 0.615, ROC–AUC = 0.734), indicating that the multi-phase TUG task captures PD-related balance deficits more effectively than straight walking. All four feature families contributed to classification performance. Dynamic and frequency-domain descriptors, often appearing in both average and asymmetry form, were most consistently selected. These features provided robust magnitude indicators and offered complementary insights into reduced control complexity in PD. Full article

As the aging population grows, ensuring effective and sustainable health management for elderly individuals has become a critical challenge. This study explores the integration of smart healthcare technologies and ESG (Environmental, Social, and Governance) principles to enhance elderly health management through data-driven strategies. Using the MIMIC-III database, this study evaluates five machine learning models (Adaboost, Bagging, Catboost, GaussianNB, and SVC) through ten-fold cross-validation to predict 3-day and 30-day mortality rates among elderly ICU patients. The Bagging model achieved the best performance with an AUROC of 0.80, demonstrating the potential of smart healthcare in mortality prediction. These technologies enhance predictive accuracy, enabling the timely identification of high-risk patients and effective intervention. Through the application of smart data integration methods, this study demonstrates how combining clinical indicators with socioeconomic factors can improve healthcare equity and efficiency. Furthermore, by aligning smart healthcare development with ESG concepts, we emphasize the importance of sustainability, social responsibility, and governance transparency in future healthcare systems. The findings offer valuable contributions toward building an interoperable and ethical health ecosystem, supporting early risk identification, improved care outcomes, and the promotion of healthy living for the elderly population. Full article

The application fields of high-temperature titanium alloys are mainly concentrated in the aerospace, defense and military industries, such as the high-temperature parts of rocket and aircraft engines, missile cases, tail rudders, etc., which can greatly reduce the weight of aircraft while resisting high temperatures. However, traditional high-temperature titanium alloys containing multiple types of elements (more than six) have a complex impact on the solidification, deformation, and phase transformation processes of the alloys, which greatly increases the difficulty of casting and deformation manufacturing of aerospace and military components. Therefore, developing low-component high-temperature titanium alloys suitable for hot processing and forming is urgent. This study used data augmentation (Gaussian noise) to expedite the development of a novel quinary high-temperature titanium alloy. Utilizing data augmentation, the generalization abilities of four machine learning models (XGBoost, RF, AdaBoost, Lasso) were effectively improved, with the XGBoost model demonstrating superior prediction accuracy (with an R² value of 0.94, an RMSE of 53.31, and an MAE of 42.93 in the test set). Based on this model, a novel Ti-7.2Al-1.8Mo-2.0Nb-0.4Si (wt.%) alloy was designed and experimentally validated. The UTS of the alloy at 600 °C was 629 MPa, closely aligning with the value (649 MPa) predicted by the model, with an error of 3.2%. Compared to as-cast Ti1100 and Ti6242S alloy (both containing six elements), the novel quinary alloy has considerable high-temperature (600 °C) mechanical properties and fewer components. The microstructure analysis revealed that the designed alloy was an α+β type alloy, featuring a typical Widmanstätten structure. The fracture form of the alloy was a mixture of brittle and ductile fracture at both room and high temperatures. Full article

Walking is a fundamental human activity, and analyzing its complexities is essential for understanding gait abnormalities and musculoskeletal disorders. This article delves into the classification of gait phases using advanced machine learning techniques, specifically focusing on dividing these phases into five distinct subphases. The study utilizes data from 100 individuals obtained from an open-access platform and employs two distinct training methodologies. The first approach adopts stratified random sampling, where 80% of the data from each subphase are allocated for training and 20% for testing. The second approach involves participant-based splitting, training on data from 80% of the individuals and testing on the remaining 20%. Preprocessing methods such as Min–Max Scaling (MMS), Standard Scaling (SS), and Principal Component Analysis (PCA) were applied to the dataset to ensure optimal performance of the machine learning models. Several algorithms were implemented, including k-Nearest Neighbors (k-NNs), Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), Naive Bayes (Gaussian, Bernoulli, and Multinomial) (NB), Linear Discriminant Analysis (LDA), and Quadratic Discriminant Analysis (QDA). The models were rigorously evaluated using performance metrics like cross-validation score, Mean Squared Error (MSE), Root Mean Squared Error (RMSE), accuracy, and $R^2$ score, offering a comprehensive assessment of their effectiveness in classifying gait phases. In the five subphases analysis, RF again performed strongly with a 94.95% accuracy, an RMSE of 0.4461, and an $R^2$ score of 90.09%, demonstrating robust performance across all scaling methods. Full article

Unreinforced masonry buildings are highly vulnerable to earthquake damage due to their limited ability to withstand lateral loads, compared to other structures. Therefore, a detailed assessment of the seismic response and resultant damage associated with such buildings becomes necessary. The present study employs machine learning models to effectively predict the seismic response and classify the damage level for a benchmark unreinforced masonry building. In this regard, eight regression-based models, namely, Linear Regression (LR), Stepwise Linear Regression (SLR), Ridge Regression (RR), Support Vector Machine (SVM), Gaussian Process Regression (GPR), Decision Tree (DT), Random Forest (RF), and Neural Networks (NN), were used to predict the building’s responses. Additionally, eight classification-based models, namely, Naïve Bayes (NB), Discriminant Analysis (DA), K-Nearest Neighbours (KNN), Adaptive Boosting (AB), DT, RF, SVM, and NN, were explored for the purpose of categorizing the damage states of the building. The material properties of the masonry and the earthquake intensity were considered as the input parameters. The results from the regression models indicate that the GPR model efficiently predicts the seismic response with larger coefficients of determination and smaller root mean square error values than other models. Among the classification-based models, the RF, AB, and NN models effectively classify the damage states with accuracy levels of 92.9%, 91.1%, and 92.6%, respectively. In conclusion, the overall performance of the non-parametric models, such as GPR, NN, and RF, was found to be better than that of the parametric models. Full article