Search Results (1,281)

Multi-site water quality sequences are jointly affected by seasonal periodicity, meteorological disturbances, and inter-site differences in the Jianghuai Watershed region. Conventional quality prediction models struggle to simultaneously achieve multi-scale feature extraction, spatial heterogeneity characterization, and prediction uncertainty expression. This study used daily-scale monitoring data on dissolved oxygen (DO), pH, and ammonia nitrogen (NH3N) from 32 monitoring stations within the region in 2025 and proposed the FT-TransONet (Fourier-enhanced Temporal Transformer Operator Network) multi-site probabilistic water quality prediction model. Within a Transformer framework, the model employed a FourierTime learnable frequency-domain filtering module, a GeoBias (Geographic Bias) attention bias mechanism, and a multi-residual ensemble strategy composed of a multilayer perceptron (MLP), a gated recurrent unit (GRU), and a temporal convolutional network (TCN) combined with a mass conservation constraint, thereby achieving both point and interval prediction of key water quality indicators. The results showed that FT-TransONet achieved the lowest Macro_RMSE among all compared methods on the multi-site water quality prediction task. At a prediction horizon of three days, its Macro_RMSE reached 0.2255, which was 21.89% lower than that of the long short-term memory network and 5.57% lower than that of the strongest baseline MC-Dropout. For the three individual indicators, the model attained coefficients of determination of 0.9135, 0.9338, and 0.8660 for dissolved oxygen, pH, and ammonia nitrogen, with corresponding root-mean-square errors of 0.5145, 0.1098, and 0.0523, confirming its potential to characterize the temporal variation in the main water quality indicators. Under multi-step prediction, the error grew gently, with the Macro_RMSE rising only from 0.2255 to 0.2384 as the horizon extended from three to seven days, and the ablation experiments, together with the probabilistic prediction results, further supported the effectiveness of the proposed structural design. Validated on 32 water quality monitoring stations in the Jianghuai Watershed, the method improved multi-site prediction accuracy while accounting for stability and uncertainty quantification, providing a preliminary reference for regional water quality early warning and management. Full article

Accurately predicting the carcinogenicity of compounds is of great significance for drug discovery, clinical drug safety, and chemical risk assessment. Traditional methods for assessing carcinogenicity rely on animal testing, which suffers from limitations such as time-consuming processes, high costs, significant interspecies differences, and low predictive throughput. In recent years, computational modeling-based prediction methods (such as Quantitative Structure–Activity Relationships, QSAR) have made some progress, but they still face challenges such as insufficient molecular feature information and poor model interpretability. To overcome these barriers, the multimodal deep learning framework LMF-CP (Late Multimodal Fusion of Carcinogenicity Prediction) is proposed to enhance the performance and interpretability of compound carcinogenicity prediction. First, to comprehensively characterize the structural and physicochemical properties of compounds, a multimodal representation system based on four molecular modalities is constructed, namely SMILES sequences, molecular fingerprints, molecular images, and molecular graph structures. Specifically, Text Convolutional Neural Network (TextCNN), Multi-Layer Perceptron (MLP), Visual Geometry Group Network (VGGNet), as well as Molecular Graph Attention Network (MGAT) are employed to process this information, respectively. Second, to integrate information from different molecular representations, a late-stage fusion strategy based on Lasso stacking is employed. On the test set, LMF-CP achieves an area under curve (AUC) of 0.828, an accuracy (ACC) of 0.782, an F1 score of 0.786, a sensitivity (SEN) of 0.786, and a specificity (SPE) of 0.779. In addition, this paper combines Shapley Additive Explanations (SHAP) analysis with Bemis–Murcko scaffold analysis to interpret the model results from two perspectives. Finally, a visual online platform for predicting the carcinogenicity of compounds is designed, providing a convenient tool for the rapid assessment of compound carcinogenicity and structural interpretation. Full article

Hyperspectral anomaly detection (HAD) identifies targets by spectral differences. However, large deep learning models overfit under the small-sample conditions typical of remote sensing, where anomalies are sparse and annotations costly. We propose a lightweight network on the GT-HAD transformer backbone for hyperspectral imagery. The design includes: (1) capacity reduction via multi-layer perceptron (MLP) ratio and embedding dimension optimization; (2) a Decoupled Projection Gating Fusion (DPGF) module enforcing spatial–spectral decoupling via dual-path projection and complementary regularization; (3) analysis of capacity–generalization tradeoffs across six airborne hyperspectral datasets. Experiments reveal an inverted-U relationship between capacity and generalization. The Mini configuration (121 K parameters) achieves a 53% parameter reduction. It yields AUC improvements on three datasets, maintains negligible performance loss (<0.1%) on two datasets, and exhibits a measurable decline on only one dataset, compared with the 255 K baseline. Notably, the baseline attains higher accuracy with only 25% training data on Pavia (+3.69% area under the receiver operating characteristic curve, AUC). This dataset-specific observation provides empirical evidence that capacity–data mismatch can induce severe overfitting in deep HAD models. Under a 25% training ratio, DPGF shows observable performance trends on spectrally homogeneous scenes, and zero initialization ensures identical performance to that of the baseline at the initial training stage, eliminating insertion risk during module deployment. However, the limited diversity of sensors and scene types constrains the generalizability of the observed trends. These results demonstrate that spatial–spectral decoupling with lightweight design suppresses overfitting in few-shot HAD, guiding compact models for onboard satellite and unmanned aerial vehicle (UAV) applications. Full article

This paper presents a novel method based on microwave reflection measurements and machine-learning techniques to estimate the water level in a multimode microwave applicator. Accurate water level monitoring is essential to maximize heating efficiency and protect the microwave source from excessive reflected power. To supplement conventional physical sensors, four regression models were evaluated: a one-dimensional convolutional neural network (CNN-1D), a multilayer perceptron (MLP), a support vector regressor (SVR), and a random forest (RF) model. These models estimate the water level based on the reflection coefficient $S_{11}$ measured under low-power conditions over the 2.2–2.8 GHz band using a waveguide-based measurement setup for embedded sensing. The models were trained and evaluated using either the magnitude, phase, or both of $S_{11}$ over the full-band or two selected reduced sub-bands. Results demonstrated that several models can accurately estimate water level from low-power microwave measurements, especially when magnitude is used as input data. The results demonstrate the potential of microwave measurements combined with machine-learning as a non-invasive supplementary sensing approach for water level monitoring. Full article

Connected vehicles and vehicle-to-everything (V2X) communication create new opportunities for estimating urban origin–destination (OD) demand from continuously collected mobility data. However, in realistic deployment conditions, only a fraction of vehicles may be connected, making OD reconstruction a highly underdetermined problem under low penetration rates. This paper proposes a supervised deep-learning framework that reconstructs full OD matrices from synthetic connected-vehicle data in a simulated Manhattan network from New York City. Vehicle movement information is aggregated into intra-zonal and adjacent-zone traffic counts using K-means traffic analysis zones. These partial connected-vehicle observations, represented by the zonal movement matrix, outgoing and incoming zonal-movement summaries, diagonal movement counts, and penetration-rate features, form a compact input to a conditional Multi-Layer Perceptron (MLP) that predicts the complete OD matrix of all vehicles. The training objective separates OD spatial shape from total traffic volume and adds losses on marginals, diagonal elements, and log-space reconstruction to embed basic flow-conservation properties. A single conditional MLP is trained across multiple connected-vehicle penetration-rate scenarios by appending the penetration rate $\rho$ and $\log \left(\rho \right)$ to the input representation. The model is evaluated over ten random connected-vehicle sampling seeds. Results show that the proposed estimator remains stable down to 20% penetration, with test sMAPE increasing only from $20.70\pm 0.00\%$ at full penetration to $21.71\pm 0.38\%$ at 20% penetration. Marginal and total-flow errors increase more gradually as penetration decreases, while clear degradation appears below approximately 2–1% penetration. Baseline and ablation comparisons further show that penetration-rate conditioning and the conservation-aware loss are essential for improving OD reconstruction and total-flow consistency. Within the evaluated simulated Manhattan scenarios, these findings suggest the potential of conditional neural estimators for OD reconstruction under limited connected-vehicle penetration. Validation across longer periods, additional demand regimes, and real-world data is required before the results can be generalized to broader urban traffic conditions. Full article

Training neural networks with gradient-based methods such as backpropagation is the dominant paradigm, but it depends on differentiable loss functions and is sensitive to initialization and local minima. Evolutionary algorithms offer a gradient-free alternative, yet the influence of their internal operators on training quality remains insufficiently characterized. This study presents a systematic comparison of backpropagation and ten variants of a genetic algorithm (GA) for training multi-layer perceptrons (MLPs), with particular focus on the role of crossover operators. The evaluation covers four MLP architectures and ten classification datasets from the UCI Machine Learning Repository, differing in sample size, dimensionality, and number of classes. Each configuration was assessed using stratified 4-fold cross-validation with 30 independent repetitions, and accuracy served as the primary performance metric, with macro-F1 reported to assess classifier behavior on class-imbalanced datasets. Backpropagation achieved higher mean accuracy than every GA variant on nine of the ten datasets, with the largest margins on high-dimensional problems. The genetic algorithm proved competitive on simpler, class-balanced datasets, where its better-performing variants matched the gradient-based baseline within one to two percentage points, and, on the Heart disease dataset, every GA variant reached a higher mean accuracy than backpropagation across all four architectures, though absolute performance remained modest on this five-class problem. Among crossover operators, BLX-$\alpha$ and BLX-$\alpha$-$\beta$ combined with tournament selection and a high crossover probability yielded the strongest configurations, while averaging crossover performed worst, as it restricts offspring to the midpoint of the parents and cannot explore beyond the range already present in the population. Tournament selection consistently led to higher mean accuracy than roulette-wheel selection, and shallow but moderately wide architectures, which encode fewer trainable parameters and thus a shorter chromosome, proved more amenable to evolutionary training than the two-layer alternative. These findings clarify when gradient-free training is competitive and which evolutionary operators drive its effectiveness. Full article

Ensuring the safety and reliability of photovoltaic (PV) installations requires accurate electrical arc fault detection. This work presents a computational arc fault detection framework that combines fixed-length windowing, Fast Fourier Transform (FFT)-based features, and supervised machine learning classifiers. Data were acquired using an Arc Fault Circuit Interrupter (AFCI) test bench developed based on IEC 63027. Current and voltage signals were partitioned into 200-sample windows, DC-offset corrected, and Hann-windowed signals. Each window generated 204 statistical and spectral attributes used to train and evaluate Random Forest (RF), K-Nearest Neighbors (KNN), Multilayer Perceptron (MLP), and Convolutional Neural Network (CNN) models. Hyperparameters were tuned by grid search with TimeSeriesSplit cross-validation, comparing min–max normalization and Z–Score standardization. On a 15% hold-out test set, CNN with Z–Score achieved F1 = 0.9982 and recall = 0.9975, followed by MLP (F1 = 0.9957) and RF (F1 = 0.9821). Amortized per-window inference latencies were ≈0.0035 ms for RF, ≈0.0016 ms for MLP with Z–Score, and ≈0.032 ms for CNN. These classifier-stage timings indicate computational compatibility with edge-oriented implementation but do not constitute an end-to-end IEC 63027 AFCI compliance assessment. The framework targets integration into PV inverters at Mackenzie Presbyterian University’s solar plant. Full article

This paper presents a novel approach to 3D semantic occupancy prediction that leverages the integration of 2D and 3D features. Traditional 3D voxel representations, while detailed, are computationally intensive. Our method addresses this challenge by encoding 3D voxel features into a Bird’s-Eye View (BEV) representation, then decoding them back into voxels using a multi-layer perceptron (MLP). This fusion approach reduces computational resources compared to voxel-only methods while maintaining state-of-the-art accuracy. By leveraging multi-scale features and deformable attention mechanisms, our network achieves a mean intersection-over-union (mIoU) of 41.26% on the Occ3D-nuScenes dataset, outperforming recent methods. By introducing a 3D backbone and multi-scale voxel-to-BEV-to-voxel feature transformation, the model achieves improved mIoU with moderate computational overhead. Our results demonstrate the effectiveness of integrating 2D and 3D information for accurate 3D semantic occupancy prediction, providing a potentially useful scene representation for downstream autonomous driving tasks. Full article

Precise estimation of lithium-ion battery State of Health (SOH) is highly demanded for reliable battery management systems, lifetime prediction, and safety assurance in electric vehicle and energy storage applications. Traditional data-driven approaches such as multilayer perceptron (MLP) often suffer from poor generalization and may produce non-physical degradation trends due to the absence of domain knowledge constraints. To address these limitations, this work proposes a monotonic Physics-Informed Residual MLP neural network framework for SOH estimation using the NASA battery dataset (B0005, B0006, B0007, and B0018). The proposed model incorporates a physics-based monotonic degradation constraint by penalizing positive gradients of SOH with respect to cycle index, thereby enforcing physically consistent capacity fade behavior. A loss function is employed to improve robustness and enhance late-cycle learning. Experimental results demonstrate that the proposed approach achieves an RMSE of 0.0287, MAE of 0.0181, and MAPE of 2.69%, indicating accurate and stable SOH prediction across multiple degradation patterns. The use of physics-informed constraints markedly enhances deterioration consistency and diminishes overfitting relative to solely data-driven models. The proposed structure offers a faithful solution for State of Health estimation in practical battery management systems. Full article

The main objective of this paper is to present a comparative analysis between the finite element method (FEM) and artificial neural networks (ANNs) for predicting sheet metal springback, while addressing the “black-box” nature of machine learning through explainable AI (XAI). To achieve this novelty, a multi-layer perceptron (MLP) architecture was implemented and evaluated against numerical simulations during the bending of a hat-shaped profile. The experimental framework utilized dual-phase HCT600X steel (0.8 mm thickness), supplemented by deep-drawing DC06 and high-strength RAK40/70 steels to ensure dataset diversity and robust generalization capability. A key contribution of this work is the integration of local SHAP (Shapley additive explanations) analysis to interpret the ANN outputs, allowing for a precise quantification and rank ordering of how individual material, design, and process parameters govern the resulting springback angle. The developed MLP model (comprising two hidden layers with five neurons each) achieved high predictive fidelity, yielding an overall correlation coefficient R = 0.99074 alongside robust error metrics (RMSE and MAE). Full article

This tutorial takes the reader on a historical and technical journey from the simple Perceptron (1958) to modern Convolutional Neural Networks (CNNs) that dominate spatial data processing (images and video). We start with the Perceptron’s linear classifier, then expose its inability to learn non-linear patterns (e.g., XOR), which motivates the Multi-Layer Perceptron (MLP) and the backpropagation algorithm. Next, we discuss the limitations of MLP when faced with structured data like images—parameter explosion, loss of spatial information, and lack of translation invariance—and use these limitations as a natural springboard to the core ideas of CNNs: local connectivity, weight sharing, and hierarchical feature learning. Throughout, we provide intuitive explanations, mathematical formulations, and step-by-step numerical examples (e.g., a complete forward and backward pass for a small network, and a manual 2D convolution). Clear graphical representations and examples help readers understand each concept. The tutorial concludes with a detailed walkthrough of influential CNN architectures (LeNet-5, AlexNet, VGG, GoogLeNet, ResNet, DenseNet, and EfficientNet) and also discusses more recent attention-based models (e.g., Vision Transformers and ConvNeXt), explaining why each was necessary and how it advanced the field. Aimed at students and practitioners with a basic knowledge of calculus and linear algebra, this tutorial connects foundational ideas to state-of-the-art deep learning, focusing on spatial data. It is designed for readers who want to understand why each architectural choice was made, not just what the final model looks like. Full article

This paper presents an efficient framework for surrogate modeling and rapid optimization of a dual-band circular patch antenna with a C-shaped slot (DB-CPAC) using multilayer perceptron (MLP) neural networks. Although highly accurate, traditional full-wave electromagnetic simulations are computationally expensive for geometric optimization due to complex slot-induced surface current perturbations. To address this limitation, a hybrid optimization framework based on Latin Hypercube Sampling (LHS) is proposed, combining the developed MLP model with a Method-of-Moments (MoM) simulator. The surrogate model uses an advanced modular architecture consisting of an ensemble of MLP neural networks for regressing center frequencies and classification MLP modules with a softmax output layer to estimate the probabilities of achieving bandwidth and gain targets. All networks are trained using the Levenberg–Marquardt algorithm with early stopping on data generated by a dedicated DB-CPAC_MoM_Sim software package. The proposed LHS-based optimizer employs the surrogate model for rapid global search and targeted local optimization before final MoM verification. Results show that this hybrid approach achieves an order-of-magnitude acceleration of the optimization process compared to conventional MoM methods while maintaining high accuracy. Full article

This paper presents a stacked neural network (StNN) framework for tool-holder health classification. The proposed architecture integrates a denoising autoencoder (DAE), used as a feature-reconstruction stage, with a multi-layer perceptron (MLP) classifier to distinguish between Healthy and Damaged tool-holder conditions. Vibration data were collected from Axial and Radial tool-holders tested under different rotational speeds and CNC machines. Six vibration descriptors—root mean square (RMS), average amplitude (AA), peak-to-peak (P2P), mean square frequency (MSF), gravity center frequency (GF), and mean spectrum amplitude (MSA)—were selected using training data only and combined with spindle speed and tool-holder type as classifier inputs. To avoid information leakage and pseudo-replication, model development and evaluation were performed using a single stratified group-wise hold-out split based on physical tool-holder units. The proposed StNN was compared with a direct MLP baseline, Random Forest, and SVM-RBF classifiers under clean/original and feature-level corrupted test conditions. On clean/original test data, the StNN achieved performance comparable to the direct MLP baseline, with balanced accuracy values of 0.8778 and 0.8694, respectively. Under feature-level corrupted test conditions, the StNN retained the highest balanced accuracy (0.8822), outperforming the direct MLP, Random Forest, and SVM-RBF baselines and showing essentially no degradation with respect to the clean/original condition. These results indicate that DAE-based feature reconstruction can preserve clean-data classification performance while improving robustness under the adopted feature-level corruption, supporting its use for feature-based tool-holder condition monitoring, with generalization assessed on unseen physical tool-holder units within a stratified group-wise hold-out split. Full article

Sanitary sewer collection systems are among the least observable urban infrastructure assets, with most utilities operating fewer than one sensor per several hundred pipes; placement drives operational value. We develop DPP-SP, a Descriptive–Predictive–Prescriptive Sensor-Placement framework that links machine-learning failure prediction with risk-weighted maximum-coverage placement and apply it to a 33,349-pipe sewer system. The geographic information system (GIS) topology is rebuilt, raising the largest connected component from 29.8% to 89.4% of nodes. Random Forest (RF), eXtreme Gradient Boosting (XGBoost), and a multilayer perceptron (MLP) are trained on combined 2020–2025 failure data; RF achieves the highest receiver-operating-characteristic area under the curve (ROC-AUC) of 0.7626 and supplies per-pipe risk weights, while repeated stratified cross-validation confirms this model ranking and permutation-importance and SHAP analyses identify pipe age and length as the dominant risk drivers. A budgeted maximum weighted coverage problem is solved over 680 candidate sites using greedy, genetic algorithm (GA) and tabu search (TS). At $K=48$, RF with greedy covers 32.26% of network risk against an 11.73% baseline, a 174.9% improvement; all three optimizers converge on the same solution. Extending to $K=400$ exposes a 56.58% coverage ceiling—set jointly by residual network fragmentation and the upstream detection range, and specific to the baseline candidate set and radius—and a six-radius sensitivity study (200–2500 m) identifies detection range as the most influential design parameter over the ranges tested. Risk coverage can be nearly tripled by redeploying the existing 48 stations without purchasing additional sensors. Full article

Olive cultivation is one of the most economically important agricultural activities in the Mediterranean region, yet its productivity is significantly threatened by olive peacock spot, caused by the fungus Cycloconium oleaginum. This disease reduces photosynthetic activity, induces premature defoliation, deteriorates fruit quality, and causes considerable yield losses. Although deep learning has significantly improved automated plant disease diagnosis, most existing approaches rely solely on leaf images and overlook environmental factors that influence disease development and progression. This study proposes an environmentally enriched multimodal deep-learning framework that integrates RGB images of olive leaves with heterogeneous environmental descriptors, including meteorological conditions, soil characteristics, rainfall-derived moisture indicators, vegetation indices, and environmental stress variables obtained from authoritative public data sources. Visual features are extracted using a fine-tuned ResNet50 convolutional neural network, while environmental descriptors are modeled using a multilayer perceptron (MLP). The extracted features are fused at the feature level to simultaneously perform seven-stage disease classification and continuous estimation of lesion coverage and leaf yellowing within a unified multi-task learning framework. Unlike synchronized field-sensor datasets, the proposed dataset combines publicly available olive leaf images with representative environmental observations, providing a reproducible proof-of-concept for multimodal disease diagnosis. Results demonstrate that incorporating environmental information substantially improves disease-stage recognition, achieving an accuracy of 97.77%, a macro F1-score of 0.9809, and a weighted F1-score of 0.9776. The proposed framework also achieved accurate severity estimation, with a mean MAE of 1.29%, RMSE values of 4.86% and 4.75%, and $R^2$ values of 0.957 and 0.969 for lesion coverage and leaf yellowing, respectively. These findings demonstrate the potential of multimodal deep learning to support precision agriculture and intelligent disease monitoring systems. Full article