Search Results (486)

Rapid surrogate models are attractive for iterative computational fluid dynamics (CFD) design loops, though defining their operating envelope remains crucial. This study proposes RenaNet, a Reynolds-aware convolutional gated recurrent unit (ConvGRU) surrogate, for predicting two-dimensional laminar and transitional flows past cylinder and square obstacles. Using two initial flow snapshots and a Reynolds-number map, the model predicts spatiotemporal flow states up to 2000 time steps into the future, with Lattice Boltzmann Method (LBM) simulations serving as ground truth. Trained on Reynolds numbers of $1\le \mathrm{Re}\le 500$ (cylinder) and $1\le \mathrm{Re}\le 250$ (square), RenaNet achieves a minimum validation mean squared error (MSE) of $1.47\times {10}^{-5}$. A Reynolds-number ablation shows that removing the conditioning channel increases the validation MSE to $1.17\times {10}^{-3}$, while a ConvLSTM baseline gives $9.94\times {10}^{-4}$ with 24% more parameters. RenaNet also uses a direct long-horizon prediction interface for distant target frames. Auxiliary physics diagnostics confirm that predictions trained via MSE maintain acceptable continuity residuals across fitting, interpolation, and extrapolation cases. The average inference time for a 1000-step prediction horizon is approximately 1.25 s, delivering a 500-fold speedup over the reference LBM solver. Interpolation errors range from ${10}^{-4}$ to ${10}^{-2}$ depending on Reynolds number and geometry, while extrapolation beyond the training regime increases errors to the order of ${10}^{-2}$. These results establish RenaNet as a robust, parameter-efficient surrogate for laminar and transitional flows, with a clearly characterized operational boundary that informs future extensions into turbulent regimes. Full article

Precise weather forecasting plays a critical role in sectors such as agriculture, transport, energy management, and climate change adaptation, and machine learning and deep learning algorithms have been widely used for data-driven time series forecasting problems. In this work, we explore the application of machine learning and deep learning models for multi-weather variable forecasting in a dataset recorded over a period of ten years (2015–2025) for five weather stations in Palestine. The dataset comprises measurements for temperature, relative humidity, wind speed, precipitation, atmospheric pressure, and sunshine hours. To avoid the issue of temporal leakage, a chronological training, validation, and test set splitting approach was used in the evaluation experiments. The models used in this study include ARIMA, SARIMA, Random Forest, XGBoost, CNN, LSTM, GRU, ConvLSTM, CNN-GRU, and CNN-LSTM with station embeddings. Our experimental results indicate that the XGBoost model achieved the highest performance in predicting temperature and relative humidity (R² = 0.953 and R² = 0.670, respectively), while deep learning methods exhibited high accuracy across several weather features. The CNN-LSTM model was successfully able to learn temporal–spatial patterns via station embeddings, while recurrent neural networks performed impressively in forecasting sunshine hours and atmospheric pressure. Full article

Objectives: Peripheral nerve ultrasound is used to evaluate nerve enlargement in immune-mediated neuropathies; however, assessment can be challenging because the distribution and severity of nerve enlargement vary among patients and are often accompanied by indistinct nerve boundaries and heterogeneous echogenicity. Although deep learning-based segmentation has been reported, most studies have focused on limited regions or single anatomical sites, primarily in compressive neuropathies. This study aimed to evaluate the performance of temporal deep learning-based segmentation for assessing diffuse or focal nerve enlargement in immune-mediated neuropathies using continuous ultrasound scans. Methods: Twenty-five healthy participants and five patients with immune-mediated neuropathy and nerve enlargement were included. Continuous ultrasound scanning from the wrist to below the elbow was performed. A static DeepLabV3+ model and temporal models incorporating convolutional long short-term memory (ConvLSTM) or Temporal Mamba were constructed and compared. Results: In healthy participants, segmentation performance was comparable across models. In contrast, in patients with nerve enlargement, temporal models demonstrated higher Dice coefficients and reduced frame-to-frame variability. The ConvLSTM-based model showed the highest performance, with mean Dice coefficients ranging from 0.87 to 0.92. Conclusions: Temporal deep learning showed potential for nerve segmentation in selected cases with nerve enlargement associated with immune-mediated neuropathies. Temporal models achieved improved segmentation performance and reduced frame-to-frame variability in these preliminary cases. This approach may facilitate more consistent quantitative ultrasound evaluation and warrants further validation in larger cohorts. Full article

Urban acoustic environments and performing-arts spaces strongly influence speech communication quality, acoustic comfort, and public wellbeing, particularly in noise-exposed shared environments such as transport hubs, campuses, healthcare spaces, public service facilities, music-education settings, and rehearsal or performance-related spaces. To address speech–noise separation in low signal-to-noise ratio and acoustically complex scenarios, this study proposes a lightweight two-stage deep learning framework termed LCGRU–Wave-SkipConvNet. In the preprocessing stage, a Lightweight Convolutional Gated Recurrent Unit (LCGRU) model is employed to achieve preliminary separation of target speech and background noise by capturing both spatial and temporal acoustic features. In the post-processing stage, a Wave-SkipConvNet model is introduced to further suppress residual noise and enhance speech quality. Experimental results demonstrate that the proposed framework achieves superior performance under different signal-to-noise ratios, sound-source angles, and target angle errors. For example, in the preprocessing stage, the LCGRU model achieved a perceptual evaluation of speech quality (PESQ) score of 2.64 at source angles between 0° and 30°, outperforming the convolutional neural network-long short-term memory (CNN-LSTM) model by 1.17. In the post-processing stage, the Wave-SkipConvNet model achieved higher short-time objective intelligibility (STOI) and segmental signal-to-noise ratio (segSNR) values than the comparison models under different SNR conditions. The proposed framework provides an effective and deployment-oriented AI solution for improving speech accessibility and acoustic comfort in urban acoustic environments and performing-arts spaces. Beyond speech enhancement, it offers practical potential for supporting healthier, more inclusive, and more equitable acoustic environments in noise-sensitive public and educational spaces. It should be noted that this study focuses on the objective acoustic environment and signal-level speech enhancement, rather than subjective soundscape perception, musical perception, or human perceptual evaluation. Full article

Accurate vessel traffic flow prediction provides an important data basis for intelligent shipping management, including maritime traffic monitoring, navigational risk awareness, waterway organization, and emission-related assessment. Although recent studies have advanced spatiotemporal, graph-based, and hybrid forecasting methods, improving the predictive ability of a conventional ConvLSTM backbone without introducing substantially more complex model structures remains underexplored in grid-based waterway scenarios. This study proposes a lightweight CA-ConvLSTM framework for grid-level vessel inflow and outflow prediction. AIS-derived flow data and MERRA-2 meteorological variables are rasterized onto a common spatial grid and fused at an early stage. A residual dilated convolution module with dilation rates of 1, 2, and 4 is used to extract multi-scale spatial dependencies, and a channel attention mechanism is applied before ConvLSTM-based temporal prediction to adaptively reweight the fused flow-meteorological feature channels. Experiments using AIS and MERRA-2 data from the northern Bohai Strait waterway show that the proposed framework improves baseline ConvLSTM performance. Compared with ConvLSTM, CA-ConvLSTM reduces MSE and MAE by 24.93% and 12.55% for outflow prediction, and by 24.80% and 12.82% for inflow prediction. These results suggest that spatially aligned meteorological fusion, multi-scale spatial feature extraction, and channel-wise feature weighting can effectively enhance ConvLSTM-based grid-level vessel traffic flow prediction without relying on complex model fusion or heavy graph-based architectures. Full article

Soluble solids content (SSC) is one of the most important indicators of sweetness, ripeness, and market quality in sweet cherries. However, conventional SSC determination is destructive, labor-intensive, and unsuitable for rapid or large-scale quality assessment. Therefore, there is a need for fast, non-destructive, and data-driven sensing approaches that can estimate internal fruit quality without damaging the sample. This study aimed to develop a non-destructive approach for SSC prediction in sweet cherries by combining open-ended coaxial probe dielectric spectroscopy with deep learning models. An open-ended coaxial probe measurement system was designed and developed to determine the dielectric properties of sweet cherries and was coupled with an Agilent E4991A impedance analyzer operating over a frequency range of 5–3005 MHz. A total of 10,080 dielectric measurements and 2100 reference SSC measurements were collected over 26 experimental days. The dielectric constant (ε′), loss factor (ε″), and loss tangent (tan δ) were extracted and used to construct separate ε′, ε″, tan δ, and integrated combined datasets. Six deep learning architectures, namely convolutional neural network (CNN), long short-term memory (LSTM), bidirectional long short-term memory (BiLSTM), gated recurrent unit (GRU), CNN-LSTM, and convolutional long short-term memory (ConvLSTM), were trained and optimized using Bayesian optimization and early stopping. CNN achieved the best performance on the tan δ dataset (test R² = 0.9099, RMSE = 0.8354 °Brix, MAE = 0.6599 °Brix), whereas GRU yielded the highest accuracy on the integrated combined dataset (test R² = 0.8622, RMSE = 1.0331 °Brix, MAE = 0.7958 °Brix). ConvLSTM provided the most consistent performance across all four datasets (test R² = 0.8081–0.8651), demonstrating strong predictive capability and practical computational efficiency. These findings confirm the potential of reduced-range dielectric spectroscopy combined with deep learning for rapid, non-destructive SSC assessment in sweet cherries. Full article

High-frequency FY-4B aerosol optical depth (AOD) observations provide useful spatiotemporal constraints for dust nowcasting, but their application over bright deserts is limited by retrieval gaps and high-AOD uncertainty. This study develops a physics-informed spatiotemporal learning framework for 15–60 min FY-4B AOD nowcasting over the Taklimakan Desert. Historical FY-4B AOD, valid masks, ERA5 dynamic fields, model-level diagnostics, and surface constraints are organized on a unified 48 × 64 grid. An LSTM–TCN–Transformer temporal backbone is combined with spatial-context encoding, mask-aware observation encoding, and structured source–transport prediction heads to represent both temporal evolution and spatial plume structures. A physics encoder represents boundary-layer mixing, vertical wind shear, source-region emission, upwind transport, and deposition loss. Mask-aware encoding and structured prediction heads are used to handle missing retrievals, source and transport increments, high-AOD tails, and low-confidence regions. Results show that FY-4B AOD constrains the main dust-belt position and spatial extent within 1 h, with skill decreasing from 15 to 60 min. High-coverage samples show more stable spatial structures, whereas low-coverage and extreme high-AOD cases have larger peak underestimation and boundary errors. The proposed framework improves high-AOD event detection and spatial-structure preservation compared with persistence, advective persistence, ConvLSTM, and ST-UNet baselines. An additional case-based comparison with MODIS MAIAC AOD and MERRA-2 dust optical depth shows partial spatial colocation between predicted high-value footprints and independent aerosol-enhancement references; however, the reported skill scores should still be interpreted mainly as spatiotemporal consistency with the FY-4B AOD product field rather than direct validation of true atmospheric dust loading. Full article

Drought is a slowly evolving, multi-driver hydro-meteorological hazard whose accurate early prediction is a cornerstone of climate-smart agriculture and water-resource planning. Existing data-driven drought forecasting frameworks suffer from three persistent limitations: (i) most models concatenate heterogeneous climate variables into a single flat feature vector, implicitly assuming a single dominant driver such as precipitation, even though atmospheric moisture demand, radiation and wind-mediated evapotranspiration co-determine drought onset; (ii) wavelet preprocessing is typically applied to the full series, introducing future-information leakage that violates the operational causality requirement of forecasting; and (iii) most architectures predict a single horizon and provide no causal attribution explaining when, where and which climatic variables initiated the event. This study proposes WaveDroughtNet, a multi-modal, multi-horizon deep-learning framework that addresses these limitations through five integrated components: (a) a strictly causal Daubechies-4 wavelet decomposition computed in a rolling fashion; (b) six modality-specific encoders with stochastic modality dropout (p = 0.15); (c) cross-modal multi-head attention with four heads; (d) a four-layer temporal convolutional network (TCN) backbone with dilation factors yielding a 240-step receptive field; and (e) a post hoc DroughtOriginTracer that combines temporal attention, modal-attribution and inter-district propagation scans. The Standardised Precipitation Evapotranspiration Index (SPEI), used as the supervisory target, is computed following the canonical Vicente-Serrano formulation. water balance $D=P-\mathrm{P}\mathrm{E}\mathrm{T}$ (Hargreaves PET) at a 4-week (≈1-month) timescale, fitted with a three-parameter log-logistic distribution via L-moments, validated by Kolmogorov–Smirnov goodness-of-fit testing ($\alpha =0.05$) per district, and standardised through the inverse-normal cumulative distribution function. Trained on 18,304 weekly district records from NASA POWER reanalysis (2014–2025) covering all 32 districts of Tamil Nadu, India, WaveDroughtNet uses only 256,869 parameters and produces, in a single forward pass, four forecasts (1 week, 1 month, 3 months, 1 year). On the held-out 2024 test partition ($N=1728$), the model attains weighted $F1=0.9221$ and $R^2=0.8512$ at the 1-week horizon, and weighted $F1=0.8498$ and $R^2=0.6812$ at the 1-year horizon. Diebold–Mariano tests confirm that WaveDroughtNet significantly outperforms naive persistence, seasonal naive, LSTM, ConvLSTM and a vanilla Transformer at the 3-month and 1-year horizons (p < 0.001). The DroughtOriginTracer successfully back-projects 15 Coimbatore events to causal origins 29–41 weeks prior to onset. We explicitly acknowledge three limitations that constrain operational deployment in its current form—zero severe events in the 2024 test partition ($F{1}_{severe}$ = 0.000), static inter-district modelling, and absence of vegetation-index supervision—and propose concrete mitigation pathways in the Discussion. 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

Total Electron Content (TEC) is a key parameter for characterizing the state of the ionosphere, and its spatiotemporal variations can significantly affect satellite navigation, radio communication, and space weather monitoring. To address the pronounced diurnal periodicity in global TEC map forecasting and the commonly neglected continuity at longitudinal boundaries, this study proposes an encoder–decoder ConvLSTM model that integrates periodic-matched residual prediction with longitude-circular boundary-aware convolution, namely the Longitude-Circular Periodic-Residual ED-ConvLSTM (LC-PR-EDConvLSTM). In the proposed model, the TEC map at the same temporal phase on the previous day is used as a periodic background field, enabling the network to focus on learning the residual variation in future TEC relative to this background. Meanwhile, longitude-circular padding is introduced into the convolution operations to preserve the spatial continuity of global TEC maps across the −180° and 180° meridians. Experiments were conducted using CODE global ionospheric map products from 2009 to 2019, with 12 TEC maps from the previous day used as inputs to predict 12 TEC maps for the following day. The results show that LC-PR-EDConvLSTM achieves RMSE values of 3.68 TECU and 1.37 TECU on the 2015 high-solar-activity test set and the 2019 low-solar-activity test set, respectively, outperforming the C1pg, ED-ConvGRU, and ED-ConvLSTM benchmark models. Ablation experiments further verify the effectiveness of the periodic-matched residual prediction strategy and the longitude-circular boundary-aware convolution. Analyses of typical space weather events and latitudinal regions demonstrate that the proposed model provides stable forecasting performance under complex space weather conditions and across most latitude regions. Full article

In the geographic environment, mesoscale ocean eddies and similar phenomena exhibit continuous and gradual changes. However, due to limitations in remote sensing observation technology, the obtained observational data are discrete, which contradicts the continuously evolving characteristics of these phenomena. Although spatiotemporal interpolation is a key tool for bridging this gap, existing single-model methods fail to fully consider continuous process features, making it difficult to obtain consistent high-quality datasets. To solve this problem, this paper combines deep learning and geostatistics to propose an Evolutionary Process-embedded Marine Spatiotemporal Interpolation Model (EPMSIM). EPMSIM first applies Seasonal and Trend decomposition using Loess (STL) to decompose marine time-series fields into trend, seasonal, and evolutionary components. Then, a Convolutional Bidirectional Long Short-Term Memory (ConvBiLSTM) model is adopted to interpolate the trend and seasonal components. Meanwhile, a Process-based Spatiotemporal Dynamic Tracking Interpolation Method (PSDTIM) is designed to interpolate the evolutionary component. Finally, these components are combined through additive coupling to produce the final interpolation result. In case studies of mesoscale eddy interpolation using SST and SLA data, EPMSIM outperforms traditional geostatistical and deep learning baselines in RMSE, MAE, and SSIM. Experimental results confirm that the model achieves significant interpolation effects in marine environmental element fields with evolutionary characteristics, validating its effectiveness in capturing continuous evolution features of marine phenomena and its feasibility for generating high-temporal-resolution spatiotemporal datasets. This study provides a methodological reference for data interpolation of evolutionary process phenomena in marine information science, and this method can be extended to other similar marine environmental variables, serving research on marine ecological environments and dynamic processes. Full article

Flash-flood and floodplain inundation simulations are highly sensitive to the spatiotemporal variability of convective rainfall, particularly during the initial runoff generation stage. However, coarse-resolution numerical weather prediction (NWP) forcing tends to smooth localized rainfall extremes, limiting its ability to accurately represent hydrological responses in low-relief floodplains. In this study, we couple a diffusion-enhanced radar nowcasting model, Diff_ConvLSTM, with a spatial resolution of 1 km and a temporal resolution of 6 min, to assess the hydrological value of high-resolution rainfall forcing over the middle Yangtze River floodplain. We introduce a monotone piecewise cubic Hermite interpolation scheme to ensure a stable transition from discrete high-frequency rainfall inputs to continuous hydrodynamic integration. Evaluation using a radar dataset from 2023 to 2024 shows that Diff_ConvLSTM better preserves intense convective echoes and rainband structures compared to the baseline ConvLSTM, increasing the Probability of Detection at the 40 dBZ threshold by 65.8%. A forcing-replacement experiment for the flood event on 30 June 2023 demonstrates that AI-based nowcasting rainfall forcing reduces peak-discharge underestimation, improves volumetric consistency, and produces inundation patterns that are closer to the observation-driven reference than those generated by low-resolution forecast forcing, although positive biases in inundation area and water depth persist. An additional event in 2024 confirms that the improvements are primarily reflected in discharge magnitude and flood volume representation, while enhancements in peak timing remain limited. Overall, the results illustrate both the added value and the remaining limitations of AI-enhanced nowcasting for hydrologically informed flood forecasting. Full article

Fish disease and species identification is critical for intelligent aquaculture, directly influencing productivity, sustainability, and economic viability. However, existing approaches largely treat species identification and pathological classification as independent tasks, limiting their ability to capture interdependent features under complex real-world conditions such as occlusion, low contrast, dynamic backgrounds, and high inter-class similarity. Moreover, challenges including class imbalance, cross-species variability, and fine-grained feature discrimination remain insufficiently addressed. To overcome these limitations, this paper proposes a hybrid ConvNeXt–BiLSTM–multi-head self-attention (MHSA) framework for joint fish species and disease classification, where a ConvNeXt-Small backbone extracts hierarchical spatial features that are transformed into a structured sequence and processed by a bidirectional LSTM to capture contextual dependencies, followed by an MHSA module for adaptive feature refinement. An auxiliary species classification branch is incorporated to provide multi-task regularization without additional inference costs. The training pipeline integrates CLAHE-based image enhancement, square-root inverse-frequency focal loss, targeted minority oversampling, and a two-stage progressive learning strategy with differential-rate cosine annealing, complemented by five-view test-time augmentation. For practical deployment, a YOLOv8s detector is employed for fish localization prior to classification. The experimental results demonstrate that the proposed model achieves superior performance, attaining overall top-1 classification accuracy of 94.33%, precision of 97.1%, recall of 90.9%, 96.1% mAP50, and an F1-score of 0.9264, while achieving a macro AUC of 0.994 and maintaining high computational efficiency (213.3 FPS), demonstrating a robust and efficient solution for real-time fish disease screening. Full article

Getting soil moisture right matters for fighting drought and stopping land from turning into desert. Aiming at the problems of insufficient spatiotemporal modeling and redundant attention mechanisms in global soil moisture prediction, we built a new deep learning model called CTA-GraphConvLSTM to better capture how soil moisture changes across both space and time, and provide technical support for drought early warning, precision agriculture and water resource management. It combines graph convolutional networks to map geographic relationships and uses a 3D-SENet attention mechanism to pull out key temporal patterns. Using the LandBench dataset, we compared the proposed model with LSTM, GraphLSTM, and ConvLSTM across multiple lead times and drought levels. Performance was evaluated using root mean square error (RMSE) and R². The CTA-GraphConvLSTM achieved the highest predictive accuracy (R² = 0.555 for 1-day lead), outperforming ConvLSTM (R² = 0.444), LSTM (R² = 0.430), and GraphLSTM (R² = 0.088). This value reveals that the model can hardly explain the variance in the data and presents extremely poor prediction performance, performing just slightly better than a simple mean predictor. The comparison results fully verify that the proposed model has higher prediction accuracy. These results demonstrate the effectiveness of graph-scale spatiotemporal modeling for soil moisture prediction. Our research has direct practical applications: it can support precision agriculture by optimizing irrigation schedules, enhance water resource management through improved reservoir operation, and strengthen drought early warning systems, thereby contributing to sustainable land use and food security. Full article

Motor imagery (MI)-based brain–computer interfaces (BCIs) have led to great interest as a result of their potential use in neurorehabilitation, assistive robotics, and human–computer interaction. However, decoding electroencephalographic (EEG) signals with high accuracy continues to be a difficult task due to the weak signal-to-noise ratio, differences among subjects, and the complicated temporal–spectral–spatial neural dynamics. Deep learning methods recently developed, such as convolutional neural networks, recurrent architectures, graph neural networks, and adversarial transfer learning, have enhanced MI decoding performance, yet many models are still concentrating on a single representation domain or they need costly adaptation phases in terms of computation. To tackle these shortcomings, we present NeuroCrossNet, a unified tri-modal deep learning model that is able to learn the temporal, spectral, and spatial EEG features jointly for robust and calibration-free MI decoding. The suggested network combines a Temporal HyperMixer Block for capturing long-range temporal dependencies, a wavelet transformer for learning localized time–frequency representation, and a Graph Attention Network for EEG topology-aware spatial reasoning. Additionally, a Dynamic Residual Attention Gate (DRAG) has been developed to adaptively merge heterogeneous feature streams, and a compact subject-aware normalization (SAN) method enhances cross-subject generalization without the use of labeled target-domain calibration data. Our proposed model was tested following the rigorous leave-one-subject-out (LOSO) approach on BCI Competition IV-2a and High-Gamma datasets. NeuroCrossNet reached a classification accuracy of 91.30%, surpassing several strong benchmark methods, including CNN-LSTM, EEGNet, DeepConvNet, spectral CNN, and graph-based EEG decoding frameworks. Furthermore, a large number of ablation studies reveal that the integration of temporally, spectrally, and spatially complementary representations considerably boosts robustness and inter-subject consistency. Full article