Search Results (662)

Autoencoders (AEs) have been widely used for industrial process fault detection owing to their ability to learn nonlinear representations from normal operating data. However, conventional AE methods rely heavily on reconstruction errors and may miss weak faults due to overgeneralization. In addition, insufficient modeling of temporal evolution and operating condition variations may reduce their sensitivity to dynamic faults. To address these issues, this study proposes a memory-enhanced and prediction-assisted conditional variational autoencoder named MI-CVAE for unsupervised fault detection. In the proposed framework, statistical features extracted from sliding windows are used as condition information to describe variable operating states. A memory module stores representative normal prototypes to constrain reconstruction and reduce overgeneralization to faulty samples. Meanwhile, an Informer branch captures temporal dependencies and provides complementary prediction residuals. Reconstruction and prediction residuals are fused to construct squared prediction error and squared Mahalanobis distance statistics, with control limits determined by kernel density estimation. The proposed method is validated on the Benchmark Simulation Model No. 1 wastewater treatment benchmark and a real papermaking process dataset. The results show that MI-CVAE outperforms the evaluated comparison methods, particularly in detecting weak and dynamic faults, while maintaining a low false alarm rate. Full article

Accurate fault diagnosis of power transformers using Dissolved Gas Analysis (DGA) depends on effective feature extraction to reduce redundancy and improve classification performance. This study compares linear and nonlinear feature extraction methods viz. Principal Component Analysis (PCA) and bottleneck Autoencoders (AE) to determine whether nonlinear representations provide diagnostic advantages for transformer fault classification. A dataset of 595 IEC 60599-labeled DGA samples covering six fault classes (PD, D1, D2, T1, T2, T3) was used. A 15-dimensional feature space was constructed from gas concentrations, total hydrocarbon content, and IEC-aligned gas ratios. PCA and AE were applied for dimensionality reduction across latent dimensions (k = 1–15), followed by an identical Artificial Neural Network (ANN) classifier. Performance was evaluated using test accuracy, cross-validation stability, and per-class F1-scores. The PCA+ANN model achieved a maximum accuracy of 68.9% at k = 11, outperforming AE+ANN, which achieved 66.4% at k = 4. PCA also demonstrated greater cross-validation stability (62 ± 3.5%) compared to AE (62 ± 6.6%). However, AE improved F1-scores for discharge faults (D1 and D2) by enhancing nonlinear separation of overlapping samples. PCA provides superior overall accuracy and stability for transformer fault diagnosis, while AE offers targeted advantages in distinguishing discharge-related faults. These findings establish a consistent benchmark for future studies and highlight the complementary roles of linear and nonlinear feature extraction in DGA-based diagnostic systems. Full article

To address the challenges of small foreign matter size, severe occlusion, and complex backgrounds in Pu-erh ripe tea processing, this study drew inspiration from primate visual mechanisms and proposed an improved YOLOv13-based network, AE-YOLOv13-S. To mitigate loss of fine details, the weakening of discriminative features, and the frequent occurrence of missed and false detections, the Adaptive Sparse Self-Attention Network was introduced to optimize the backbone of the network, inspired by the sequential cognitive pattern of primates involving target search, local verification, selective integration, and final decision making. To address insufficient long-range semantic associations and the submergence of fine-grained differences in background noise, Emulating Self-Attention with Convolution was employed to optimize part of the Conv modules of the network, drawing on the hierarchical information processing mechanisms of primates from peripheral perception to central fine analysis. In response to the limitations of bounding boxes, such as approximate target enclosure, the large amount of geometric supervision noise, the obvious localization deviation, and delayed model convergence, a Scale-based Dynamic Loss, inspired by primate visual perception mechanisms, was introduced to optimize the network’s loss function. The results showed that, during training, compared with the baseline, AE-YOLOv13-S achieved lower training loss values: Box Loss declined by 6.76%, Cls Loss by 6.52%, and DFL Loss by 8.65%. On the validation dataset, the model demonstrated reductions of 6.58%, 16.39%, and 8.33% for these respective metrics. After the overall improvements, AE-YOLOv13-S achieved increases of 1.43, 4.85, and 2.69 percentage points in precision, recall, and mAP@50, respectively, with only a 0.3 G increase in FLOPs. The improved model can classify and detect foreign matter in Pu-erh ripe tea efficiently and accurately, providing not only a new technical pathway for foreign matter detection in tea processing but also a practically meaningful technical solution for intelligent quality control and food safety assurance in the tea processing chain. Full article

Industrial control systems (ICS) are increasingly targeted by cyberattacks that manipulate physical processes while evading data-driven detectors trained on raw time-series data. This paper extracts 34–41 control-theoretic features, including tracking error, valve mismatch, sensor liveness, and their temporal derivatives, from Proportional–Integral–Derivative (PID) control loops and evaluates them using an Isolation Forest combined with a maximum z-score. On HAI 21.03, Stage 1 achieves a PA-F1 score of 0.8945, detecting 48 out of 50 attacks. On HAI 23.05, Stage 1 attains a PA-F1 score of 0.9210, surpassing seven deep-learning baselines by at least 23 PA-F1 points; the closest baseline, a learned Graph Neural Network (GNN), achieves 0.6890. Re-implementations of ConvBiLSTM-AE (PA-F1 = 0.6689) and TranAD (PA-F1 = 0.6838) on the same evaluation split confirm this performance gap. A controlled USAD experiment, with PA-F1 = 0.7343 for physics features versus 0.6687 for raw Supervisory Control and Data Acquisition (SCADA), demonstrates that the extracted features provide the detection signal independently of the model architecture. Adding a bidirectional Gated Recurrent Unit (GRU) refinement stage improves PA-F1 by 8.1 percentage points on HAI 21.03, but the same stage reduces it by 6.8 percentage points on HAI 23.05, where attacks manifest as brief perturbations; four alternative Stage 2 designs reproduce this degradation. We therefore characterize temporal refinement as beneficial only for sustained-deviation attacks and identify Stage 1 as the primary deployable detector. This study is the first to apply physics-informed features, report both PA-F1 and eTaPR on HAI 23.05, and perform per-window error diagnosis on this dataset. Results show that 10 of 15 detected windows are covered by fewer than 10% of their timesteps, revealing a structural tension between PA-F1 and eTaPR. Full article

Predictive modeling of thermal aging in solid propellants is challenging because HTPB-based propellants are highly filled particle-reinforced polymer systems with sparse experimental data, nonlinear parameter coupling, and partially unclear aging mechanisms. This study proposes a Physics-Informed Manifold Neural Operator (PIMANO) framework for multi-parameter prediction of polymer aging in HTPB solid propellants. Accelerated thermal aging, stress relaxation, and swelling experiments were used to obtain aging temperature, aging time, crosslinking density, and viscoelastic Prony-series parameters. A continuous aging-state field was first reconstructed over the temperature–time domain by radial basis function interpolation. Crosslinking density was then introduced as a physically interpretable bridge-state variable linking aging conditions with viscoelastic responses. Among three candidate kinetic models, the modified Arrhenius–Avrami model gave the best fitting performance for crosslinking-density evolution, with R² = 0.988 and MRE = 0.0199. By combining local multi-scale neighborhood features, manifold latent representations, and DeepONet-based operator learning, PIMANO established a unified mapping from aging conditions to multi-parameter viscoelastic responses while incorporating bridge-state consistency, parameter non-negativity, and evolution-direction constraints. Under the RBF-augmented validation setting, PIMANO-ae achieved RMSE = 0.7847, MAE = 0.3366, R² = 0.9995, and MRE = 0.0027. Compared with the traditional model, RMSE, MAE, and MRE were reduced by 94.93%, 96.47%, and 96.85%, respectively. Temperature leave-one-out validation further yielded average R² values of 0.9469–0.9647 and MRE values of 4.98–6.21% at unseen aging temperatures. These results demonstrate that PIMANO provides an accurate, stable, and physically interpretable framework for multi-parameter aging prediction and life-assessment modeling of polymer-based energetic materials. Full article

Five new Daphniphyllum alkaloids, daphmacropomines A–E (1–5), were isolated from Daphniphyllum macropodum Miq. and structurally characterized. Compound 1 is an unprecedented Daphniphyllum alkaloid that features a unique 6/9/7/5/5 ring with a rare N-nitroso group. Compound 2 is the second Daphniphyllum alkaloid featuring a highly rearranged 6/6/6/7/5/6-fused skeleton. Compounds 3 and 4 are new yuzurimine-type alkaloids, and their structures were determined by extensive techniques including HRESIMS, NMR, ECD, and single-crystal X-ray. A plausible biosynthetic pathway for 1–4 is proposed. Furthermore, compounds 3 and 4 can induce lysosomal biogenesis and promote autophagic flux. Full article

Reliable detection of lubrication conditions in sliding bearings is crucial for condition monitoring and predictive maintenance. Despite advances in tribological research, there remains a need for accurate diagnostics that indicate worsening of lubricity in mixed and boundary lubrication states. In this study, a dynamic test procedure is utilised to classify lubrication conditions with the help of a boosted tree classification algorithm. A radial journal bearing test rig is built and equipped with a high-frequency acoustic emission (AE) sensor on which experiments consisting of repeated dynamic speed and load alterations are conducted. AE signal features are extracted, compared and used to train an Extreme Gradient Boosting (XGBoost) classification model. The model achieves high accuracy (97.57%) in distinguishing adequate vs. starved lubrication conditions in mixed friction. Misclassifications are mainly observed at the lowest load or speed conditions, where residual lubrication effects make the classes less separable. The model’s generalisability is evaluated by applying it to tests with differing viscosity classes and alternative bearing materials without retraining, with the classifier retaining good performance. The model is also used to detect anomalies in a grease-lubricated system, where it successfully detects poor lubrication conditions. While it is known prior to this publication that AE is a good tool to detect anomalous behaviour in hydrodynamic journal bearings, the findings presented highlight the potential for the transferability of anomaly detection models trained in a laboratory setting and applied to different real-world applications to reduce life-cycle maintenance costs and increase uptime in industrial applications. Full article

Introduction: Schizophrenia is characterised by distributed abnormalities in electrophysiological dynamics and large-scale brain networks, yet unimodal EEG or MRI alone cannot fully explain how fast neural computations relate to spatially organised circuit dysfunction. Multimodal EEG–MRI approaches offer a bridge across temporal and anatomical scales by explicitly modelling cross-modal coupling. Methods: Following PRISMA 2020 guidance, we conducted a systematic, mechanistic review of human studies (adults ≥ 18 years) comparing schizophrenia-spectrum groups with healthy controls using EEG combined with at least one MRI modality (fMRI, structural MRI, and/or diffusion MRI) and explicit EEG–MRI integration (e.g., EEG-informed fMRI, joint ICA, mCCA/MCCA, coupled matrix–tensor factorisation, DCM-based fusion). Searches were performed in PubMed/MEDLINE, Embase, Web of Science, Scopus, PsycINFO, IEEE Xplore, ResearchGate, and Google Scholar for January 2000–December 2025, supplemented by citation tracking. Risk of bias was assessed with ROBINS-I, and due to heterogeneity, results were synthesised narratively by integration of families. Results: From 148 records, 23 studies met the inclusion criteria. Studies used mainly simultaneous EEG–fMRI at 3T and spanned resting-state designs and task paradigms dominated by auditory processing (oddball, MMN/N100–P200, ASSR/aeGBR), with additional work in affective context, working memory, semantic processing (N400), sensory gating, and pharmacologic challenge. Across tasks, the most reproducible multimodal signature was disrupted coupling between electrophysiological markers and the recruitment of large-scale networks, rather than isolated changes in EEG or fMRI metrics. Target detection/oddball paradigms converged on reduced late ERP responses (especially P300, sometimes N2) alongside reduced expression or loss of coupling to salience/ventral attention and control circuitry (including ACC/anterior insula/TPJ). Resting-state studies most consistently indicated altered “coupling rules” (frequency specificity, timing/lag structure, and directionality), including abnormalities detectable even when unimodal summaries were weak. Extended multimodal studies (adding sMRI/DTI and/or classification) suggested that combining modalities can improve discrimination, though performance was sensitive to sample size, demographic imbalance, and feature-selection/validation choices. Conclusions: Multimodal EEG–MRI studies support schizophrenia as a disorder involving persistent structural and circuit-level abnormalities whose functional expression varies dynamically across cognitive states and task demands. Future progress will depend on harmonised acquisition/artefact-control practices for simultaneous EEG–fMRI, larger and more diverse samples (including early/CHR and longitudinal designs), and cross-site replication of mechanistically interpretable coupling biomarkers. Full article

Evaluating intrusion detection methods at the level of individual MITRE Adversarial Tactics, Techniques, and Common Knowledge (ATT&CK) for Industrial Control System techniques requires Operational Technology traffic in which each attack sequence carries its MITRE technique identifier as ground truth. Publicly available Industrial Control System datasets either provide coarse attack-versus-benign labels (SWaT, WADI, CIC-APT-IIoT) or require ex-post technique reconstruction from CALDERA operation logs, and therefore do not support per-technique benchmarking. We describe one primary contribution and two supporting contributions, demonstrated on one Modbus/Raspberry-Pi programmable logic controller/CALDERA/convolutional bidirectional Long Short-Term Memory autoencoder (CNN-BiLSTM-AE) use case. The primary contribution is an in-orchestrator labelling methodology for per-technique-labelled Industrial Control System attack capture. Its single load-bearing property is that the campaign orchestrator owns the label primitive and writes each per-sequence technique identifier into the capture artefact at injection time, eliminating ex-post log-to-packet alignment. The first supporting contribution is a protocol-aware detection pipeline. Its load-bearing architectural choice is a priority-ordered protocol router that dispatches each labelled flow to a per-protocol detector plug-in (protocol-aware features here, with generic-flow features admissible as an alternative plug-in policy on the same router). The second supporting contribution is a suite of four reproducible CALDERA chains (three Information-Technology-to-Operational-Technology kill chains plus one enterprise-side control) that exercise the labelling methodology end-to-end and the detection pipeline along complementary detection paths. All three contributions are platform-independent: any ATT&CK-aligned emulator and any fieldbus protocol can host the labelling methodology, and any detector trained on an admissible feature space can plug into the router. The dataset contains 40,000 benign and 9997 attack Modbus sequences spanning four ATT&CK techniques (T0802 Automated Collection, T0831 Manipulation of Control, T0836 Modify Parameter, T0846 Remote System Discovery). On this dataset, the CNN-BiLSTM-AE reaches a 100% true-positive rate (TPR) at the 98th-percentile benign threshold across all four techniques and a 99.7% overall TPR at the tighter 99.5th-percentile threshold, with per-technique TPR between 96.1% (T0836 Modify Parameter) and 100% (T0802 Automated Collection, T0846 Remote System Discovery). Across the four CALDERA chains, the Modbus autoencoder produces 234 protocol-layer detections and the Security Information and Event Management (SIEM) rule set produces 30 alerts, with per-chain tactic coverage between 0.714 and 0.786 and CALDERA-ability success rates between 0.800 and 0.857. Full article

Accurate classification of transformer partial discharge (PD) severity is essential for insulation diagnostics yet remains challenging due to nonlinear feature relationships and class imbalance. This study evaluates whether feature extraction improves PD severity classification and compares the effectiveness of linear and nonlinear extraction methods. A dataset of 294 samples was categorized into four IEC-aligned severity classes. Two raw measurements (discharge magnitude and applied voltage) were expanded into a 15-dimensional feature space. Principal Component Analysis (PCA) and a bottleneck Autoencoder (AE) were used for linear and nonlinear feature extraction, respectively. Extracted features were classified using an identical Multilayer Perceptron (MLP). Both feature extraction methods improved classification performance over raw and full-feature baselines (96.6%). PCA+ANN achieved 100.0% accuracy (k = 9), while AE+ANN achieved 98.3% (k = 8). The AE misclassified one minority “Normal” sample due to poor latent boundary representation. Reconstruction analysis showed the highest error for the Normal class, reflecting imbalance-driven optimization bias. Feature extraction enhances PD severity classification, with linear PCA outperforming nonlinear AE in this near-linearly separable dataset. PCA’s deterministic projection preserves minority class boundaries more effectively, whereas AE performance is limited by class imbalance. These findings suggest that nonlinear methods provide advantages only in more complex feature spaces. Full article

Evaluation of fiber–matrix interfacial properties is essential for understanding composite performance and exploring the feasibility of real-time diagnostic approaches. In this study, the interfacial behavior between glass fiber and epoxy resin was examined using acoustic emission (AE) features obtained during microdroplet pull-out tests. Four AE features (amplitude, energy, rise time, and Fast Fourier transform peak frequency) were used as input variables to Random Forest models for both regression and classification tasks, targeting interfacial shear strength estimation and failure mode identification (interfacial debonding vs. fiber fracture). In regression analysis, energy and amplitude showed stronger associations with interfacial shear strength, although overall regression performance remained limited. In classification analysis, amplitude alone provided the most stable discrimination between fiber fracture and interfacial debonding, while combining multiple features offered only a marginal additional benefit due to feature redundancy. These results suggest that intensity-related AE parameters are closely associated with interfacial debonding behavior and failure modes. Overall, this exploratory study indicates that AE-based machine learning can serve as a supplementary tool for indirect and trend-level assessment of fiber–matrix interfacial behavior, with potential relevance to real-time monitoring applications. Full article

Reliable estimation of battery remaining useful life (RUL) becomes difficult when the capacity trajectory contains regenerative rebounds, short-term oscillations, and long-range temporal dependence. To address this problem, an adaptive decomposition and hybrid deep-learning framework is proposed. First, the phototropic growth algorithm (PGA) is used to tune variational mode decomposition (VMD), allowing the capacity series to be separated into low-frequency trend information and high-frequency fluctuation information so that the influence of regeneration and noise is weakened. Next, a component-level predictor combining a temporal convolutional network (TCN), an attention mechanism (AM), and a Transformer is constructed. In this architecture, TCN learns multi-scale local features, AM enhances salient degradation cues, and the Transformer captures global long-horizon dependencies. To deduce the future capacity degradation path and the associated RUL, these estimated elements are synthesized. Results on the NASA, CALCE, and BIT datasets verify the effectiveness of the proposed framework. On NASA dataset, the average root mean square error (RMSE), mean absolute error (MAE), and absolute error (AE) reach 0.0123 Ah, 0.0073 Ah, and 0.5 cycles, respectively, improving on the strongest baseline by 11.9%, 19.7%, and 50.0%. On CALCE dataset, the corresponding values are 0.00695 Ah, 0.00499 Ah, and 1.75 cycles, and all R² values are higher than 0.9989, indicating strong accuracy and robustness in the presence of complex regeneration behavior. Supplementary BIT validation on three higher-capacity cells further achieves average RMSE, MAE, and AE of 0.01201 Ah, 0.00771 Ah, and 1.0 cycle, respectively. Full article

Background: Image-guided thermal ablation has been used for the treatment of primary lung carcinoma but its use in the treatment of multiple lung carcinoma and effects on survival have not been well established. Objective: This study compares the long-term survival metrics for stage 1 single primary lung cancer and multiple primary lung cancer (MPLC) in patients treated with image-guided thermal ablation (IGTA). Methods: A retrospective institutional review included 37 NSCLC patients (mean age 71.6 ± 8.8 years) with ≥5 years follow-up. In total, 119 IGTA procedures were performed. Among patients with a single tumor (n = 14, 37.8%), each underwent a single ablation session. In contrast, patients with MPLC (n = 23, 62.2%) underwent 88 ablation sessions to treat 105 tumors. Data included demographics, tumor features, procedural details, safety, adverse events, and outcomes. Primary endpoints were 5-year overall survival (OS), progression-free survival (PFS), and cancer-specific survival (CSS). Results: All ablations were completed successfully. Severe AEs occurred in 5.8% (7/119) of the ablations and were limited to pneumothorax requiring chest tube placement with hospitalization. At the time of ablation, individual nodules were staged at T1A = 46 (38.7%), T1B = 54 (45.4%), T1C = 16 (13.5%) and T2A = 3 (2.5%). Local recurrence was observed in 4/119 (3.3%) ablated tumors, all at stage T1B, and all were retreated with ablation. The 5-year OS was better for patients with MPLC at 85.6% compared to patients with a single tumor at 35.7% (HR = 0.14, p = 0.003, 95% CI: 0.037, 0.51). The 5-year OS for tumors based on T classification for T1A, T1B, TIC and T2A was 71.4%, 66.8%,66.7% and 0%. The 5-year PFS was 77.4% for patients with MPLC compared to 35.7% for patients with single primary lung cancer (HR = 0.25, p = 0.014, 95% CI: 0.084, 0.76). The 5-year CSS was 95.2% for patients with MPLC compared to 83.1% for patients with single primary lung cancer (HR = 0.21, p = 0.16, 95% CI: 0.018, 2.33). Conclusions: IGTA is an effective and safe treatment for patients with stage 1 single primary lung cancer and MPLC with limited local recurrence. Tumor size up to 3 cm did not have significant impact on survival. Overall survival was improved in patients with MPLC compared to those with single NSCLC. Clinical Impact: IGTA can be safely performed in patients with single primary lung cancer and MPLC, with limited local recurrence rate. Highlights: Key Findings: IGTA effectively treats patients with stage 1 single primary lung cancer and MPLC, with 3.3% recurrence, which can be retreated with ablation. The five-year OS was higher in patients with MPLC (85.6%) versus those with single lung cancer (35.7%, p = 0.003). OS by T classification: 71.4% for T1A, 66.8% for T1B, 66.7% for TIC, and 0% for T2A. Importance: IGTA effectively treats patients with single primary lung cancer and MPLC with low recurrence. Tumor size < 3 cm showed no impact on overall survival. Full article

Genotype imputation (GI) plays a critical role in predicting missing genetic information for genomic studies and breeding applications. Although recent reference-free deep learning approaches have demonstrated promising performance, they often fail to exploit local genomic information, which limits further improvements in prediction accuracy and stability. In this study, we developed ALGI, a novel method based on a sparse convolutional denoising autoencoder, which uniquely integrates local genomic window information with group-specific feature learning. Unlike conventional convolutional or autoencoder-based approaches, ALGI first applies K-means clustering to group samples according to local genomic windows, then learns hidden genotype configurations specific to each group, capturing fine-scale local patterns and complex haplotype structures. Systematic evaluation was conducted across yeast, human, and pig MHC regions under multiple scenarios, including different window sizes, missing rates, sample sizes, and numbers of variants. Results show that ALGI demonstrates consistent improvements over conventional methods (Beagle) and state-of-the-art deep learning approaches (AE, SCDA) under the evaluated settings, with enhanced accuracy, stability, and robustness. In addition, ALGI is user-friendly and publicly available. While evaluated on highly polymorphic MHC regions, its strong performance suggests applicability to less complex regions, though broader genome-wide validation is needed. This approach provides a powerful tool for genomic selection and advancing complex trait genetics in livestock and other species. Full article

This research introduces a quantum machine learning (QML) approach for predicting heart disease (HD). The method combines preprocessing of data with quantum feature map (QFM) and quantum classification techniques. In the method, clinical data of HD are preprocessed, and then features are optimized using principal component analysis (PCA). After that, the resulting features are encoded into quantum states with five different QFM methods, namely angle encoding (AE), amplitude encoding (AmE), basis encoding (BE), Pauli encoding (PE), and ZZ feature map (ZZFM). Finally, four quantum classifiers, such as quantum support vector machine (QSVM), quantum k-nearest neighbor (QKNN), quantum random forest (QRF), and variational quantum circuit (VQC), are evaluated to predict the HD from the encoded states. Experimental results show that QSVM with AE achieved the best performance, with an overall accuracy of 90.26%, specificity of 83.42%, sensitivity of 92.16%, precision of 88.89%, F1-score of 89.68%, and kappa value of 0.7608. These results are superior to those from classical state-of-the-art methods. This research finding suggests QML methods can capture complex nonlinear relationships in clinical data effectively and thus improve diagnostic reliability. Full article