Search Results (197)

Accurate electricity price forecasting (EPF) is essential for market participants to optimize trading strategies and for power systems to accommodate the increasing penetration of volatile renewable energy sources. However, electricity price series are characterized by strong nonlinearity, high volatility, and significant structural breaks, which pose substantial challenges to conventional forecasting models. Although numerous hybrid deep learning models have been proposed for EPF, most existing approaches either overlook structural breaks or treat them as outliers rather than as signals of regime shifts, often resulting in systematic forecasting degradation when market conditions change abruptly. To address this issue, this study proposes COCAL-TTL, a novel multi-stage structural break-aware forecasting framework that integrates regime-adaptive data partitioning with a functionally differentiated hybrid deep learning architecture. First, a joint detection scheme combining the Iterated Cumulative Sum of Squares (ICSS) algorithm and the Chow test is employed to partition Spanish electricity market data from 2014 to 2023 into distinct regimes. Within each regime, CEEMDAN is applied to extract multi-scale features, which are subsequently reconstructed into trend, periodic, and random components based on an independent sample t-test and Fast Fourier Transform (FFT). The CNN-SE Attention-LSTM (CAL) model, with hyperparameters optimized by the Osprey Optimization Algorithm (OOA), serves as the primary forecasting engine. In addition, a dedicated heterogeneous error correction module, namely TTL, is introduced, in which Temporal Convolutional Network, Transformer, and LSTM are designed to capture local transients, long-range dependencies, and transitional dynamics in the residual series, respectively. Empirical results demonstrate that compared with the Naive benchmark, COCAL-TTL achieves percentage MAPE improvements of 58.48% and 48.97% in low- and high-volatility regimes, respectively. These findings indicate that the proposed structural break-aware framework provides a robust data-driven solution for EPF under heterogeneous market conditions and offers technical support for stable electricity market operation in the context of renewable energy integration. Full article

Landslide susceptibility algorithms demonstrate high reliability in quantifying the likelihood of landslide occurrence. However, traditional methods are often limited by computationally intensive sampling strategies and models with limited adaptability. In this study, we propose an adaptive sampling strategy based on hotspot analysis to enhance the reliability of the generated samples. Additionally, we develop an improved meta-ensemble (IME) stacking-based heterogeneous framework for landslide susceptibility assessment by integrating a support vector machine (SVM), random forest (RF), and XGBoost. To further reduce factor complexity, a Monte Carlo-based frequency ratio analysis is employed. The Baihetan Reservoir area along the Jinsha River was selected as the study area. A total of 26 conditioning factors were considered, supplemented by 120 Sentinel-1A images to cover the study area. The proposed sampling strategy was then used to generate high-quality samples. Finally, to evaluate the performance of the proposed method, the proposed ensemble learning framework was applied to assess landslide susceptibility with eight models using five evaluation metrics. The experimental results demonstrated that: (1) the adaptive sampling strategy improved both the quantity and quality of the training samples; (2) the adoption of the Monte Carlo strategy increased the sample partitioning rate; and (3) despite the formally highest IME metrics, the inclusion of InSAR information did not lead to a statistically significant improvement in the forecast compared to the high-quality basic sampling strategy. Overall, the proposed methodology provides valuable support for regional geohazard susceptibility assessment in dynamic environments. Full article

Effectively integrating graph topology and node attributes, while assigning nodes with both semantic similarity and structural closeness to the same community, remains a key challenge in attributed graph community detection. To address this challenge, this study proposes TVHCL-CD, a two-view hierarchical contrastive learning-driven method for community detection. The proposed method constructs an attribute view and a modularity view from the node attribute matrix and the modularity matrix, respectively, to model attribute semantics and high-order community structure priors. Structure-aware two-view representations are then learned in parallel through dual-view graph attention encoders incorporating multi-order neighborhood priors. Furthermore, a structure-enhanced Graph Transformer fusion module is designed to achieve node-level adaptive fusion of the two-view representations by introducing a learnable adjacency bias into global self-attention and a view-aware gating mechanism into the feed-forward network. To align the optimization objective with community semantics, a hierarchical contrastive learning strategy is further developed. Specifically, view-level consistency contrastive learning constructs modularity-guided augmented views to improve representation robustness, while community-level semantic contrastive learning incorporates partial ground-truth labels to enhance intra-community compactness and inter-community separation. Finally, clustering is performed on the fused representations to obtain community partitions. Experimental results on eight real-world attributed graphs and the generated tree-like attributed graph Tree-2500 indicate that TVHCL-CD achieves competitive performance under the semi-supervised transductive setting, while ablation results support the contributions of its main components. Full article

This study proposes a novel intrusion detection and response architecture that formulates network traffic analysis as a sequential decision-making problem rather than a static classification task. The architecture integrates a Transformer Encoder for temporal feature extraction with a Dueling Double Deep Q-Network (DDQN) to enable autonomous and risk-aware security decisions. Network flows are modeled within a Markov Decision Process, where the agent learns an optimal policy over a hierarchical action space consisting of IGNORE, LOG, ESCALATE, and BLOCK actions. To evaluate generalization capability, a transfer learning-based cross-domain adaptation strategy was employed. The CICIDS2018 and CICIoT2023 datasets were re-partitioned using a stratified 70/15/15 train/validation/test split. The proposed model achieved high detection performance on these datasets with F1-scores of 99.48% and 99.13%, respectively. After transfer learning to the AWID3 dataset, the model preserved strong generalization capability with F1-scores of 96.76% and 96.61%, demonstrating its robustness across wired, IoT, and wireless network environments. A risk-aware reward function is designed to balance detection accuracy and operational cost, while Integrated Gradients-based explainability is incorporated to analyze decision behavior. Experimental results further show that the proposed Transformer–DDQN framework achieves more stable learning, lower optimization loss, and more consistent action policies compared to alternative reinforcement learning-based approaches. The model operates with high computational efficiency while maintaining real-time processing capability in high-throughput network environments. Full article

Particle swarm optimization (PSO) is a simple and efficient metaheuristic algorithm that has been widely applied to solving various practical problems. However, PSO has some inherent limitations, such as a tendency to get trapped in local optima and an imbalance between global exploration and local exploitation. To overcome these challenges, this paper proposes a novel algorithm called the multi-swarm particle swarm optimization algorithm with multi-learning strategy (MPLPSO). First, the entire swarm is randomly partitioned into multiple sub-swarms, each comprising three distinct types of particles, which enables the algorithm to explore multiple potential solutions simultaneously. Next, a pool elite learning strategy combined with a convergence learning mechanism is employed to effectively reduce the risk of premature convergence. Furthermore, an elimination-replacement mechanism is integrated with a hierarchical competition strategy to further enhance the solution accuracy. Extensive experiments conducted on the CEC 2017 and CEC 2022 benchmark test suites demonstrate that the proposed MPLPSO significantly outperforms the classical PSO and several state-of-the-art PSO variants. Additionally, MPLPSO is also applied to the traveling salesman problem, and the experimental results further validate the superior performance and robustness of the proposal. Full article

Processing-in-memory and in-memory computing (PIM/IMC) are increasingly explored to mitigate the von Neumann data-movement bottleneck that limits deep neural network (DNN) performance and energy efficiency. Progress, however, remains fragmented across device substrates, architectural prototypes, mapping and scheduling methods, compiler toolchains, and benchmarking practices, making results hard to compare and slowing deployment. This survey synthesizes developments from 2019–2025 along four coupled axes: (i) memory substrates and architectural design, (ii) mapping, partitioning, and scheduling, including learning- and graph-based strategies, (iii) compilers and end-to-end deployment flows, and (iv) benchmarking datasets, metrics, and reporting norms. Drawing on over twenty representative platforms spanning static random-access memory (SRAM) and dynamic random-access memory (DRAM), emerging non-volatile, capacitive, and photonic substrates, we clarify the trade-offs separating analog/charge-domain IMC from digital SRAM/DRAM-centric PIM, including reported peaks up to 600 TOPS/W and 1.5 TOPS/mm². We organize mapping frameworks into a unified reference taxonomy, identify recurrent evaluation pitfalls that undermine reproducibility, and highlight persistent gaps in training support, robustness under non-idealities, and coverage of large-scale GNN workloads. Finally, we outline a five-phase roadmap from benchmark standardization to industrial validation toward compiler-integrated, GNN-informed PIM/IMC systems validated on production-scale workloads. Full article

This study presents a structure-aware inference framework for tooth numbering and lesion localization in visible light oral images. Tooth numbering is often compromised by class imbalance and structural inconsistency caused by the uneven distribution of tooth types, motivating the integration of anatomical priors into the inference process. The framework first partitions the dental arch into quadrants using a deep learning-based detection module to establish spatial organization. Based on this, an Anchor-Teeth-Guided Inference (ATGI) strategy reconstructs globally consistent tooth numbering by leveraging dental arch continuity, bilateral symmetry, and confidence-guided anchor selection, thereby improving the recognition of underrepresented tooth classes. Visually suspicious lesion regions are independently detected and spatially associated with numbered teeth, enabling joint structural and lesion-aware analysis. Evaluated on a multi-source dataset, the method achieves a weighted F1-score of 0.813 for 32-class tooth numbering, outperforming end-to-end baselines while improving spatial consistency. Lesion localization yields F1-scores of 0.850 for caries-related regions and 0.789 for gingivitis-related regions. These results demonstrate that incorporating anatomical constraints enhances numbering robustness and improves rare-class recognition in visible light dental image analysis, showing potential for screening-oriented oral assessment and teledentistry applications. Full article

As the offshore wind industry scales toward 15 MW capacity, the reliability of Direct-Drive Permanent Magnet Synchronous Generators (DD-PMSGs) becomes critical. However, real-world run-to-failure data for these massive, multi-pole machines is virtually non-existent, creating a barrier for developing effective data-driven diagnostic systems. This study proposes a high-fidelity framework for detecting permanent magnet faults in the International Energy Agency (IEA) 15 MW Reference Wind Turbine. Using Finite Element Analysis (FEA), a dataset (magnetic flux and back electromotive-force (EMF)) capturing the electromagnetic signatures of healthy and faulty states of a PMSG under varying severities is generated. To improve the power of computer vision, 1D time-series signals were transformed into 2D images. Specifically, Gramian Angular Fields (GAFs) and Recurrence Plots (RPs) were applied to magnetic flux density signals, while Markov Transition Fields (MTFs) were applied to back-EMF signals. These representations were then fused into multi-channel Red-Green-Blue (RGB) images and processed via a ResNet-18 Deep Transfer Learning model using a strictly non-overlapping, leakage-free dataset partitioning strategy. The proposed framework achieved a classification accuracy of 99.45% on noise-free data. Furthermore, robustness testing under varying levels of Additive White Gaussian Noise (AWGN) (30 dB, 40 dB, and 50 dB Signal-to-Noise Ratio (SNR)) demonstrated sustained high performance, maintaining over 90% accuracy even under severe 30 dB noise conditions. Comparative analysis proved that this multi-channel fusion significantly outperforms single-channel encoding methods, which collapse under heavy noise, validating the scalability of the framework and applicability for next-generation condition monitoring in harsh offshore environments. Full article

Feature selection is essential for improving classification performance in high-dimensional biomedical data, yet conventional metaheuristic algorithms often suffer from premature convergence and loss of population diversity. To address these issues, this paper proposes a Feature Grouping and Improved Hybrid Breeding Optimization framework (FGIHBO). First, the original feature space is hierarchically partitioned using the Maximum Relevance Minimum Redundancy criterion and Symmetric Uncertainty analysis to alleviate the curse of dimensionality. Then, a Multi-Strategy Synergistic Improved Hybrid Breeding Optimization (MSIHBO) algorithm is developed by incorporating Grey Wolf Optimizer (GWO) guidance and a Shannon entropy-adaptive simulated annealing mechanism to balance exploration and exploitation. Experimental results on the CEC2022 benchmark suite demonstrate that MSIHBO provides robust optimization performance across diverse problem categories. Furthermore, evaluations on eleven high-dimensional biomedical datasets show that FGIHBO achieves average classification accuracies ranging from 92.77% to 97.66%. Compared with representative algorithms, including Multi-strategy Improved Grey Wolf Optimizer (MIGWO), Hybrid Whale Optimization Algorithm based on Gathering strategy (HWOAG), Dynamic Crow Search Algorithm (DCSA), GWO, Hybrid Breeding Optimization (HBO), Hybrid Breeding Optimization based on Lévy flight and Elite Opposition-Based Learning strategy (LEHBO), and MSIHBO, the proposed framework improves average classification accuracy by 1.47–27.46%, with the largest gain observed on dataset D10 relative to HWOAG. These results confirm the effectiveness, robustness, and scalability of the proposed framework for high-dimensional biomedical feature selection. Full article

Hyperspectral image (HSI) classification enables precise classification of land-cover types from rich spectral and spatial information. Recent methods combine convolutional neural network (CNN) and Mamba branches to exploit their complementary local and global modeling capabilities for HSI classification. However, most of these methods treat all spectral channels uniformly in feature fusion, failing to account for the discriminability differences across spectral bands. Moreover, most methods rely on a single classification head at the final layer, which may lead to vanishing gradients in shallow layers. To address these limitations, a spectral group-wise gated CNN–Mamba network with cross-stage mutual distillation, called SGGCMNet, is proposed. To address the first limitation, a CNN–Mamba spectral group-wise gating block (CMSB) is designed at the feature-fusion level. Specifically, the CMSB partitions channels into multiple sub-groups along the spectral dimension. Each sub-group learns its own fusion weights that balance local spectral–spatial cues produced by a CNN pathway with long-range context produced by a Mamba pathway. To address the second limitation, two loss-level optimization strategies are proposed jointly: A progressive deep supervision strategy with uncertainty-based dynamic weighting is proposed to attach classification heads at all network stages. A temperature-regulated cross-stage mutual-distillation mechanism is further designed to enable bidirectional knowledge transfer among classification heads at different stages. On three benchmark HSI datasets, SGGCMNet achieves state-of-the-art accuracy. Full article

Accurate prediction of loan defaults is essential for financial institutions seeking to minimize credit losses and maintain portfolio stability. In the vehicle financing segment of emerging markets, real-world datasets frequently exhibit extreme class imbalance ratios that far exceed those encountered in standard benchmark corpora, posing severe challenges for conventional machine learning pipelines. This study introduces a gradient boosting framework integrating Weight of Evidence (WoE) transformation, Bayesian hyperparameter optimization, and three complementary classifiers—Extreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM), and Categorical Boosting (CatBoost)—to predict vehicle loan default risk. The methodology is evaluated on a large-scale, fully anonymized Turkish vehicle loan dataset ($N=207,572$) with an extreme imbalance ratio of 1:1133 (183 defaults versus 207,389 non-defaults). A strict three-way data partition (60% training, 20% validation, 20% test) is adopted to ensure leakage-free model selection and unbiased performance estimation. A multi-stage experimental pipeline is developed encompassing: (i) statistical feature selection via Mann–Whitney U and chi-square tests with adaptive thresholding, (ii) a comparative analysis of seven resampling strategies including Synthetic Minority Oversampling Technique (SMOTE) variants, Adaptive Synthetic Sampling (ADASYN), and focal loss weighting, (iii) a greedy forward selection ensemble procedure for heterogeneous model fusion, and (iv) a systematic training-set size sensitivity analysis across eight majority undersampling ratios. Under the leakage-free evaluation protocol, the highest-AUC individual model (LightGBM with SMOTE-ENN) achieves an Area Under the Curve (AUC) Receiver Operating Characteristic (ROC) of 0.710 (95% bootstrap CI: 0.614–0.798), while CatBoost with cost-sensitive weighting exhibits superior operational metrics (KS $=0.389$, PR-AUC $=0.011$). The greedy ensemble procedure exhibits high selection instability with only 37 validation-set positives, providing a methodological finding on the minimum sample requirements for reliable ensemble construction under extreme scarcity. Ablation results confirm that WoE encoding contributes 3.1 percentage points to the overall AUC gain. Tree SHAP-based interpretability analysis identifies the financing-to-age ratio, WoE-encoded occupation group, and log financing amount as the primary predictive drivers, with cross-model stability confirmed via Spearman rank correlation. A decision support analysis provides precision–recall curves, a Brier score of 0.0082, reliability diagrams, and threshold-dependent performance at operationally plausible review rates. Fairness evaluation across gender and marital status subgroups demonstrates that threshold-dependent metrics such as Disparate Impact Ratio and Equalized Odds Gap are inherently compromised under extreme minority scarcity, whereas rank-based subgroup AUC analysis with bootstrap 95% confidence intervals preserves meaningful discriminative assessment. These findings provide an empirically validated framework for credit default prediction in highly imbalanced and data-scarce financial environments. Full article

Targeting non-apoptotic regulated cell death (RCD) modalities, such as ferroptosis and cuproptosis, offers a new avenue for overcoming resistance to conventional antitumor therapies, while deep learning provides a powerful tool for discovering bioactive molecules from multi-source data. This review delineates the core methodologies and application advances of deep learning in this domain, covering end-to-end molecular representations, multimodal fusion strategies, dataset partitioning criteria, and deep learning frameworks, thereby establishing a preliminary technical framework tailored to the study of non-apoptotic RCD mechanisms. Subsequently, the applications of deep learning in non-apoptotic RCD are discussed along three dimensions: direct applications, adjacent applications, and speculative outlooks. Future directions should focus on constructing high-quality annotated databases capable of distinguishing multiple cell death modalities and establishing standardized blind test benchmarks, developing explainable AI methods, designing mechanism-oriented few-shot learning algorithms, and building dynamic context-aware models. Advances along these directions may help propel the application of deep learning in drug discovery targeting non-apoptotic RCD mechanisms, from computational prediction toward experimental validation and translational research. Full article

This paper applies physics-informed neural networks (PINNs) to laterally excited liquid sloshing in a two-dimensional rectangular tank, where near-resonant forcing (${\omega }_e/{\omega }_1=0.9$) produces a multi-frequency beating response with a period of approximately $10T_1$. Linearized potential flow theory governs the problem; the network learns the velocity potential $\phi (x,z,t)$ while the free-surface elevation $\eta$ is injected analytically. Two training obstacles specific to forced sloshing are analyzed. First, a zero-solution trap arises because the trivial solution $\widehat{\phi }=0$ satisfies all equations except the free-surface conditions, whose residuals are roughly ${10}^{-4}$ times smaller than the Laplace residual; characteristic-scale normalization combined with loss weighting (${\lambda }_D={\lambda }_K=100$) breaks this trap. Second, spectral bias prevents standard MLPs from resolving the three co-existing frequencies (${\omega }_1$, ${\omega }_e$, $\mathsf{\Delta }\omega$); a Fourier time embedding that augments the input from 3 to 9 dimensions overcomes this limitation. Two additional techniques further reduce errors: a hard-wall boundary condition enforced exactly via a $\cos (\pi x/B)$ spatial embedding, which eliminates wall collocation points; and a gradient-enhanced Laplace regularizer (${\parallel \nabla \left({\nabla }^2\widehat{\phi }\right)\parallel }^2$) that constrains velocity smoothness through third-order automatic differentiation. An ablation study shows that these four techniques progressively reduce the horizontal velocity error from ${\epsilon }_u=12.46\%$ to $0.84\%$. Results are validated against a viscous finite-difference benchmark. Over one beating cycle the errors are ${\epsilon }_{\eta }=0.15\%$, ${\epsilon }_u=0.84\%$, and ${\epsilon }_w=1.65\%$. A frequency parameter study across ${\omega }_e/{\omega }_1$ = 0.5–1.1 gives ${\epsilon }_{\eta }<0.25\%$ and ${\epsilon }_u<2.3\%$ for all near-resonance cases. For long-time simulation, a time-domain decomposition strategy with transfer learning partitions the domain into one-beat windows; extending to five beating cycles ($50T_1$) yields ${\epsilon }_u=3.43\%$ and ${\epsilon }_{\eta }=0.30\%$ with no monotonic error accumulation across windows. The methodology is then extended to a three-dimensional rectangular tank ($B\times W\times H$) with bi-directional lateral excitation. The 3-D formulation introduces the y-dimension into the Laplace equation (${\nabla }^2\phi ={\phi }_{xx}+{\phi }_{yy}+{\phi }_{zz}=0$), adds transverse wall boundary conditions ($\partial \phi /\partial y=0$) enforced exactly via a $\cos (\pi y/W)$ embedding, and extends the Fourier time embedding from 9 to 16 dimensions to accommodate six physical frequencies. The bi-directional excitation excites both $(m,0)$ and $(0,n)$ modal families, producing a genuinely three-dimensional beating response. Experimental results verify that the proposed methods can be well generalized to three-dimensional scenarios. Within a single beating cycle, the relative errors reach ${\epsilon }_{\eta }=0.24\%$, ${\epsilon }_u=1.31\%$, ${\epsilon }_v=1.78\%$ and ${\epsilon }_w=2.32\%$, with a total training time of 2499 s. By applying time domain decomposition to carry out two-cycle three-dimensional simulations, the model can steadily maintain satisfactory prediction precision across segmented time intervals, achieving overall errors of ${\epsilon }_{\eta }=0.30\%$ and ${\epsilon }_u=1.32\%$. Full article

Post-disaster Search and Rescue (SAR) missions demand rapid coordination of Heterogeneous Unmanned Aerial Vehicle (UAV) fleets under stringent payload and flight range limitations. Traditional heuristic solvers struggle to solve the Large-Scale Heterogeneous Team Orienteering Problem (LSH-TOP) within operational time limits due to the coupled complexity of task allocation and route planning. A Hierarchical Deep Reinforcement Learning framework decomposes this high-dimensional combinatorial problem into tractable sub-problems. An upper-level policy, guided by Monte Carlo Tree Search (MCTS), partitions the global target set to balance fleet workload distribution, whereas a lower-level Transformer-based model constructs near-optimal trajectories for individual agents. A Curriculum-Integrated Alternating Cooperative Training (C-ACT) protocol resolves the convergence difficulties associated with sparse feasible solutions in constrained environments. This protocol incorporates a dynamic constraint annealing strategy and a virtual agent buffer to progressively shape the solution space from relaxed to strictly constrained formulations. Experiments conducted on real-world geographic data demonstrate the proposed approach consistently outperforms all baselines across scales of 80 to 300 targets, improving over the strongest competitor by 0.63–8.51% and over conventional heuristics by up to 53.27% in objective value. Results indicate a task completion rate of 27.5% at the 300-target scale (versus 25.1% for the strongest baseline MCTS + OR) and balanced workload distribution, validating framework adaptability to complex emergency response scenarios. Full article

Background: Preoperative diagnosis of medullary thyroid carcinoma (MTC) is clinically challenging due to sonographic overlap with other thyroid tumors. To address this, we aimed to develop a multi-vendor, multimodal radiomic framework for accurate MTC identification, comparing its diagnostic performance with that of experienced radiologists. Methods: This retrospective study included 467 pathologically confirmed thyroid nodules (94 MTCs, 373 non-MTCs) acquired across multiple ultrasound platforms. The dataset was randomly partitioned into training (80%) and internal testing (20%) sets. In total, 2250 radiomic features were extracted from grayscale and color Doppler images, followed by Z-score normalization to mitigate batch effects. A robust feature selection strategy (LASSO and recursive feature elimination) identified optimal signatures for developing machine learning classifiers (SVM, LR, RF). The optimal model was further validated on an independent, balanced cohort (n = 60; comprising 12 cases each of MTC, papillary carcinoma, follicular carcinoma, follicular adenoma, and nodular goiter) and compared with experienced radiologists across seven classification tasks. Results: The RF model achieved an AUC of 0.993 in distinguishing MTC from papillary carcinoma. The LR model showed an AUC of 0.991 for identifying MTC from all other nodules. In the independent validation cohort, the models maintained superior discriminatory ability, showing better diagnostic performance compared to the image interpretation by radiologists (AUC 0.993 vs. 0.488, p < 0.001). Conclusions: The proposed multi-vendor, multimodal radiomic system demonstrated good discriminative ability in the diagnosis and stratification of MTC. By integrating grayscale and Doppler ultrasound features while overcoming scanner variability, this model shows potential as a non-invasive adjunctive tool. Full article