Search Results (730)

To address the difficulty of accurately capturing the dynamic patterns of carbon emission time series—characterized by nonlinearity, non-stationarity, and complex fluctuations—this paper proposes a carbon emission prediction model based on an elite-guided collaborative social spider learning optimization algorithm (EGC-SSLO) integrated with a Long short-term memory (LSTM) network. First, considering the limitations of the standard stochastic social learning optimization (SSLO) algorithm in complex high-dimensional optimization problems, such as insufficient elite information guidance, weak local exploitation in the later stages, and a tendency to become trapped in local optima, three complementary improvement strategies are introduced. The adaptive elite mean-guided search strategy enhances the search directionality by incorporating the cooperative information of the best individual and the elite mean. The worst-individual hybrid Cauchy–Lévy search mechanism achieves a dynamic balance between early-stage global exploration and late-stage local exploitation through long-range Lévy flights and fine-grained Cauchy perturbations. The quadratic directional exploitation strategy further refines the search trajectory of candidate solutions, thereby improving convergence accuracy. These three strategies significantly enhance the optimization performance without increasing the time complexity order of the algorithm. Experimental results on the CEC2017 (30-dimensional), CEC2020 (20-dimensional), and CEC2022 (20-dimensional) benchmark suites demonstrate that EGC-SSLO consistently outperforms classical algorithms such as PSO, GWO, and HHO, as well as their improved variants, in terms of convergence accuracy, convergence speed, and robustness. Furthermore, the Wilcoxon rank-sum test and Friedman test confirm that the observed improvements are statistically significant. Finally, an EGC-SSLO-LSTM carbon emission prediction model is constructed and applied to daily carbon emission data in China from 2019 to 2025 for empirical analysis. The experimental findings show that the EGC-SSLO-LSTM model markedly outperforms both the standard LSTM and SSLO-LSTM approaches across key evaluation metrics, including mean absolute error (MAE), root mean square error (RMSE), and the coefficient of determination (R²). In particular, the MAE is decreased by 39.9% and 4.64% compared with the two benchmark models, respectively, which highlights the strong effectiveness and practical potential of the proposed method in real-world carbon emission forecasting applications. Full article

This paper proposes a Multi-Strategy Improved Connected Banking System Optimizer, named MICBSO, for numerical optimization and three-dimensional UAV path planning. MICBSO enhances the original CBSO through three coordinated strategies. First, a chaos–opposition learning initialization strategy is introduced to improve initial population quality and search coverage. Second, a Gaussian perturbation-based multi-elite guidance mechanism is designed to reduce dependence on a single best solution and strengthen the balance between exploration and exploitation. Third, a hybrid boundary control strategy combining reflective correction and random reinitialization is developed to improve solution feasibility and maintain population diversity. The proposed algorithm is evaluated on the CEC2017 benchmark suite and compared with 11 representative algorithms. Experimental results show that MICBSO achieves competitive convergence accuracy, stability, and robustness across different dimensional settings. In addition, MICBSO is applied to three-dimensional UAV path planning in four complex terrain scenarios. The results demonstrate that MICBSO can generate feasible and safe flight paths with lower comprehensive cost. Overall, the proposed method provides an effective optimization framework for both benchmark optimization and constrained UAV path planning tasks. Full article

When addressing high-dimensional complex optimization problems, the vanilla Dung Beetle Optimizer (DBO) suffers from slow convergence, frequent stagnation in local optima, and progressive degradation of population diversity. To overcome the above inherent defects, this paper proposes a multi-strategy hybrid improved DBO variant named the SWDBO, which incorporates three targeted enhancement modules. First, an adaptive population proportion strategy is developed to dynamically adjust the population sizes of rolling beetles, brood beetles, small beetles and thief beetles throughout iterations. More individuals are allocated for extensive global exploration at the early evolutionary stage, while more search agents are reserved for delicate local exploitation in later iterations, which maintains stable population diversity over the entire optimization process. Second, the bubble-net encircling and spiral predation mechanisms of the Whale Optimization Algorithm (WOA) are embedded into the position update formula of rolling beetles. This integration strengthens fine local search performance and accelerates the overall convergence rate. Third, a modified seagull optimization operator combined with Lévy random perturbation is introduced into the position updating rule of thief beetles. This improved jump mechanism optimizes individual movement trajectories and enables the algorithm to effectively escape local optimal traps. Numerical experiments are implemented on the 100-dimensional benchmark functions of CEC2017 and CEC2020. Moreover, the proposed SWDBO is validated on three classical constrained engineering optimization tasks, including three-bar truss design, ten-bar truss design and cantilever beam sizing optimization. Wilcoxon rank-sum tests statistically verify significant performance disparities between the SWDBO and competing optimizers. For the three structural engineering cases, the design solutions obtained by the SWDBO produce lighter structural mass while satisfying all constraint requirements. Overall experimental evidence proves that the proposed multi-strategy improvement framework can efficiently tackle high-dimensional numerical optimization and constrained engineering design problems, and the SWDBO exhibits prominent performance in balancing global exploration and local exploitation. Full article

Unmanned aerial vehicle semantic communications are increasingly required in low-altitude sensing, intelligent inspection, and emergency response, where raw image transmission is difficult to sustain under limited onboard resources and time-varying air-to-ground links. Meanwhile, the simultaneous transmission of visual semantic features and object-centre location metadata under third-party eavesdropping creates a dual-privacy vulnerability: an attacker can exploit both to reconstruct sensitive content. In this paper, we propose a differential privacy-based collaborative protection framework that inserts dedicated perturbations into visual semantic and location descriptors before transmission. For visual data, we design a region-aware differential privacy mechanism that applies stronger noise to sensitive semantic regions while preserving utility for non-critical areas. For location data, a scenario-adaptive strategy is developed, comprising randomized differential privacy for discrete grid-based location information (coarse spatial awareness) and Laplace-based differential privacy for continuous coordinates (fine-grained protection). To balance privacy and utility, we formulate a joint optimization problem. It maximizes legitimate-side semantic task performance by coordinating the visual privacy budget, location privacy budget, and transmit power. A BCD-based algorithm is developed to solve this non-convex problem. Attacker-side recoverability is verified empirically at the optimized operating point. Simulation results demonstrate stable convergence within a small number of iterations. Compared with uniform differential privacy, the proposed framework achieves a superior task-level privacy–utility trade-off and provides selective sensitive-region protection, with the two mechanisms yielding comparable whole-image attack suppression. Full article

Lightweight, privacy-aware and adversarial robust intrusion detection is required for the proliferation of Internet of Things (IoT) devices. In the Industrial Internet of Things (IIoT), centralized detectors can be compromised by adversarial perturbations via gradient-based attacks, making them susceptible to raw traffic. We suggest Federated Learning-Adaptive Gated Recurrent Unit (FL-AdGRU), a Federated approach that combines a lightweight Gated Recurrent Unit (GRU) classifier with alternating adversarial fine-tuning on each client using FGSM and PGD, without any communication overhead. A two-stage resampling scheme (UCAS-SMOTE) reduces the class-imbalance ratio from 4081:1 to ≈4:1, followed by 61 features being reduced to 40 by a mutual-information selector (MI-SelectK). Under this scenario, FL-AdGRU achieves 99.9% accuracy and 0.999 weighted F1 (+6.5 p.p. over the federated DNN baseline), with no loss of accuracy when facing clean attacks, and boosts Fast Gradient Sign Method FGSM/Projected Gradient Descent (PGD) robustness by +19.3/+19.0 p.p. at the same level of $ϵ$ = 0.1, thus effectively balancing the accuracy–robustness trade-off. It is robust (97.8%/84.2% on UNSW-NB15) and generalizes well to UNSW-NB15, while decaying slowly in skeptical scenarios (≈99.9% weighted F1 for moderate skew, 93.9%/86.7% for severe). Assuring data-locality privacy through exchange of only model weights; defenses against inference attack are left for future work. FL-AdGRU, with a total communication of 43.8 MB (≈50× less than centralized training), is deployable on bandwidth-constrained IIoT networks. Full article

Federated learning (FL) is a promising paradigm for privacy-preserving IoT intrusion detection, but its effectiveness can be substantially degraded by the combination of heterogeneous non-IID client distributions and severe multi-class imbalance. Under such conditions, conventional size-based aggregation may overemphasize large yet highly skewed clients, limiting the representation of minority attack classes in the global model. To address this issue, we propose Mean/Std, a lightweight distribution-aware aggregation strategy that combines a client-size proxy with two complementary statistics of local label distributions, namely the standard deviation and the dominance gap of class proportions, while preserving a communication footprint comparable to FedAvg. Experiments on the N-BaIoT benchmark, comprising seven heterogeneous IoT clients and eleven traffic classes, are conducted under a privacy-oriented update-perturbation setting inspired by secure aggregation workflows. The results show that Mean/Std consistently provides the strongest imbalance-aware performance among the evaluated FL baselines, achieving a Macro-F1 score of 0.8418 and a Balanced Accuracy of 0.8722 while improving the representation of minority attack classes. Additional experiments across five independent random seeds and a comprehensive hyperparameter sensitivity analysis further confirm the robustness and stability of the proposed aggregation mechanism. Overall, the results demonstrate that lightweight distribution-aware aggregation offers an effective, robust, and practically deployable solution for mitigating aggregation bias under simultaneous non-IID heterogeneity and severe multi-class imbalance in FL-based IoT botnet detection. Full article

This study develops an interval-valued q-spherical fuzzy rough set TOPSIS framework (IVq-SFRS-TOPSIS) for multi-criteria group decision-making when expert judgments contain interval uncertainty, neutrality, and granular indiscernibility. The revised framework clarifies the relationship between interval-valued q-spherical and interval-valued T-spherical fuzzy models, defines admissible approximation operators over compatible domains, and introduces a radial projection step that guarantees closure under the IVq-SFN constraint whenever component-wise extrema would otherwise violate it. The proposed framework provides a mathematically balanced representation of interval-valued q-spherical fuzzy information, reflecting the concept of symmetry and supporting reliable group decision-making under uncertainty. The TOPSIS procedure is then formulated through expert aggregation, benefit–cost normalization, entropy-based criteria weighting, ideal-solution distance calculation, and closeness-coefficient ranking. The method is illustrated through a sustainable smart city development case using four AI-based alternatives and six criteria. Rather than claiming unconditional superiority, the revised comparative and sensitivity analyses examine how the ranking changes under alternative fuzzy decision models, different q values, perturbations to criteria weights, and perturbations to the decision matrix. The results indicate that the proposed framework provides an interpretable rough-boundary representation and a reproducible ranking mechanism for complex MCDM problems under interval-valued q-spherical uncertainty. Full article

At 1 km resolution, NDVI projections for heterogeneous landscapes can appear spatially coherent in aggregate while concealing substantial class-level prediction weaknesses, a limitation that has received limited systematic attention in the NDVI projection literature. This study applies a four-component assessment workflow to Northeast China (NEC) for 2040 under SSP1-2.6, SSP2-4.5, and SSP5-8.5, integrating multi-stage model selection, land-cover-stratified validation, quantile-regression-based uncertainty characterization, and validation-priority ranking. Among three candidate tree-based models evaluated using spatial block cross-validation, temporal holdout validation, long-jump extrapolation, and climatic perturbation tests, LightGBM showed the most balanced and consistent performance, with spatial CV R² = 0.654 ± 0.123, temporal holdout R² = 0.710, and long-jump R² = 0.671, and was therefore selected for the 2040 projection. Projected regional mean NDVI increased modestly from 0.393 in 2020 to 0.414–0.417 across scenarios, with limited divergence among SSP pathways at this near-term horizon. Class-stratified validation of the 2020 holdout prediction revealed that global model performance masked strong class-level heterogeneity, with R² values ranging from 0.576 for Construction land to −0.886 for Unused land. Water bodies and Unused land exhibited negative R² values, indicating weak class-level predictive support relative to a simple class-mean benchmark. Residual decomposition showed that Water bodies combined high random error with elevated systematic deviation, whereas Unused land was mainly characterized by systematic bias, suggesting different needs for class-specific model improvement. The Uncertainty Risk Index (URI), derived from 95% prediction intervals, was highest in Construction land and lowest in Cropland across all scenarios. Integrating historical residuals with future URI-identified Water bodies, Unused land, and Construction land as the highest-priority classes for future targeted validation. These priorities arise from both limited class representation and intrinsic NDVI-related complexity, including low vegetation signal, mixed-pixel effects, and heterogeneous land-surface composition. These results demonstrate that land-cover-stratified error decomposition and uncertainty-informed priority ranking reveal class-specific projection limitations that aggregate accuracy metrics can conceal. Full article

Complex-valued knowledge graph embedding (KGE) models like ComplEx effectively capture asymmetric relations but are fundamentally constrained by integer-order transformations. This restriction limits their ability to model multi-scale interactions, hierarchical correlations, and non-local semantic dependencies inherent in heterogeneous graphs. To address these limitations, this paper introduces FracComplEx, a novel fractional-order extension that embeds fractional calculus into the complex latent space. By leveraging fractional operators, the framework introduces non-local dynamics and memory-aware mechanisms to continuously generalize standard linear transformations. The core architecture employs a fractional-order parameter $\alpha$ as a controllable scaling mechanism that balances local relational details with global topology, optimizing representation smoothness and flexibility. We provide rigorous theoretical findings demonstrating that fractional transformations enhance the embedding’s expressive capacity, spectral characteristics, and perturbation robustness beyond conventional integer-order benchmarks. Extensive experiments on FB15k-237, WN18RR, and CoDEx-M establish the empirical superiority of FracComplEx, yielding significant improvements in Mean Reciprocal Rank (MRR) and Hits@K metrics over classical baselines, particularly under severe structural data sparsity. Full article

Point cloud denoising is a fundamental preprocessing task in 3D vision and geometry processing, where the main challenge is to suppress corruption while preserving sharp features, thin structures, and local surface fidelity. Classical geometric filters are computationally efficient and interpretable, but they commonly rely on fixed local supports or globally selected parameters, which limits their effectiveness under spatially heterogeneous corruption. More adaptive non-local and learning-based approaches can improve robustness, yet they often introduce higher computational complexity, stronger modeling assumptions, or substantial training-data dependency. In this work, we propose Noise-Adaptive Bilateral Normal Projection (NABNP), a learning-free point cloud denoising framework that introduces explicit patch-wise adaptation to local corruption conditions. NABNP estimates a robust dimensionless local noise level from point-to-plane residuals and uses this quantity to adapt the angular bandwidth of bilateral normal refinement, the balance between weighted local averaging and normal projection, and the magnitude of the positional update. This design enables conservative smoothing in locally reliable neighborhoods while applying stronger geometric correction in more severely corrupted regions. We evaluate NABNP on standard benchmark models under 13 stress scenarios covering additive Gaussian noise, outlier contamination, sparsity, and rotation perturbation, resulting in 650 trials per method. The experimental results show that NABNP provides strong aggregate behavior among the evaluated learning-free baselines, particularly under low-to-medium corruption, sparsity, and rotation perturbations, while its advantage becomes less pronounced under the most severe noise and outlier settings. The method remains training-free and interpretable, with a moderate computational cost associated with repeated neighborhood analysis and covariance-based local updates. Full article

Aging is the dominant risk factor for most chronic diseases, yet the mechanisms driving this relationship remain poorly integrated across biological scales. Existing frameworks have catalogued key hallmarks of aging but do not explain how these processes converge to produce organism-level decline and multimorbidity. A systems-level framework is introduced in which aging is conceptualized as progressive destabilization of interacting regulatory networks. Mitochondrial quality control, nutrient-sensing pathways, and chronic inflammatory signaling form a putative high-centrality network core: mitochondria coordinate redox balance, bioenergetics, and transcriptional adaptation, while NAD⁺-dependent signaling and NLRP3 inflammasome activation propagate perturbations across regulatory layers. This architecture provides a mechanistic basis for the convergence of neurodegenerative, cardiovascular, metabolic, and oncological phenotypes as emergent consequences of shared network instability. Reframing the hallmarks as coupled network nodes shifts the explanatory focus from isolated mechanisms to system-level resilience and non-linear dynamics. This narrative and conceptual review integrates evidence across mitochondrial biology, metabolic signaling, and inflammatory pathways to develop these arguments, with explicit acknowledgment that the proposed framework is hypothesis-generating rather than formally validated. Interventions targeting high-centrality nodes, including mTOR modulation, NAD⁺ restoration, mitophagy activation, and anti-inflammatory strategies, may exert system-wide effects by reconfiguring network dynamics rather than correcting individual pathways. This perspective suggests that biomarker-stratified, network-calibrated interventions may offer a broader systems-level therapeutic rationale than single-pathway approaches. Full article

Deep neural networks have substantially improved the performance of hyperspectral image classification, yet they remain vulnerable to adversarial attacks. Existing attack methods usually manipulate pixel spectra directly, ignoring the physical mixing mechanism of remote sensing imaging and potentially generating adversarial samples with limited physical consistency and interpretability. Moreover, balancing attack effectiveness and perturbation imperceptibility remains a challenging multi-objective optimization problem. To address these issues, this paper proposes an evolutionary multi-task multi-objective adversarial attack framework based on inter-task knowledge transfer. Instead of perturbing raw pixel spectra, the proposed method introduces perturbations into abundance maps obtained through spectral unmixing, thereby improving the physical plausibility of the generated adversarial samples. The generation of class-specific universal perturbations is formulated as a collaborative multi-task optimization problem. To solve this problem, we develop a Self-Adaptive Multi-Objective Multi-Factorial Evolutionary Algorithm for Adversarial Attacks (SAMO-MFEA-AA). By modeling the attack generation processes for different land-cover classes as distinct yet correlated optimization tasks, SAMO-MFEA-AA dynamically captures synergistic relationships among tasks. An asymmetric adaptive cooperation matrix is employed to regulate the intensity of knowledge transfer, allowing beneficial perturbation patterns to be shared across related classes while reducing the risk of negative transfer. Extensive experiments on the Indian Pines and Salinas datasets demonstrate that the proposed framework achieves competitive hypervolume performance and favorable solution diversity compared with existing multi-objective optimization algorithms. In adversarial attack scenarios, the proposed method achieves effective attack success rates against representative classification networks while maintaining the physical plausibility of abundance-space perturbations. Full article

Urban innovation districts (UIDs) provide an important space for innovation-led development. However, their scale and the heterogeneity within the districts are not easy to capture with administrative units, especially in urban centers of great denseness. Therefore, we develop a parcel-based geospatial framework to locate and typify the UIDs in Shanghai using multi-source geospatial data. Morphology parcels are used to measure an innovation asset index, and a hierarchical density-based spatial clustering algorithm (HDBSCAN) is applied to identify functionally coherent innovation agglomerations beyond administrative boundaries. The procedure identifies 18 UIDs; an exploratory stability check further indicates that high-value core parcels are largely retained under moderate density-connectivity perturbations, whereas peripheral boundary parcels are more sensitive. Patent-density and policy-alignment evidence are used as supporting external evidence for the delineated innovation geography and as one part of a convergent evidence check. To describe spatial heterogeneity, we collate 31 conceptual indicators, operationalized as 38 PCA variables after dummy coding, to represent locational features, urban form, function and environmental quality. Principal component analysis reduces these variables to six components explaining 81.59% of the total variance, and hierarchical clustering classifies the districts into five types. A k = 2–8 cluster-quality assessment supports the five-type solution as a balanced and interpretable classification rather than a uniquely optimal statistical partition. The results reveal mismatches between administrative boundaries and functional innovation spaces, as well as systematic built-environment differences across district types. Full article

Power data terminals deployed in smart-grid environments generate massive amounts of operational data, yet the sensitive nature of these data and the existence of cross-region silos make centralized model training difficult in practice. Federated learning offers a natural alternative by enabling collaborative model optimization without transferring raw data, but its direct use in power terminal scenarios is still limited by four coupled challenges: update leakage, malicious or abnormal client behavior, constrained terminal-side resources, and severe Non-IID data heterogeneity. To address these issues, we develop SFL-PDT, a hierarchical federated learning framework tailored to power data terminals. The proposed method is built on a server–edge–terminal architecture. Within this architecture, edge nodes aggregate terminal updates from relatively homogeneous regional groups and perform local robustness screening, while the central server aggregates edge-level updates across heterogeneous regions and coordinates the privacy budget schedule for protected updates. It combines adaptive privacy-aware update perturbation, robust suppression of suspicious regional updates, terminal-oriented update compression, and similarity-guided aggregation for heterogeneous data distributions. Experiments on two representative power-system tasks, load forecasting and fault diagnosis, demonstrate that SFL-PDT achieves a superior overall balance among privacy protection, robustness, efficiency, and predictive performance. Compared with the evaluated baselines, the proposed method more effectively reduces reconstruction-related leakage under different privacy budgets, lowers leakage similarity under gradient inversion attacks, and maintains robust performance when malicious clients participate. It also converges faster and more stably under heterogeneous data partitions. In addition, SFL-PDT achieves the best overall predictive results, reaching an MAE of 0.021 for load forecasting and an accuracy of 88.2% for fault diagnosis, while reducing average terminal-side local training time from 4.3 s to 2.9 s and per-round upload volume from 4.2 MB to 1.5 MB relative to FedAvg. These results suggest that SFL-PDT is a practical solution for secure, efficient, and heterogeneity-aware collaborative learning in power data terminal environments. Full article

Phosphene-based visual neuroprostheses require encoding schemes that preserve task-relevant information while remaining feasible under safety-constrained stimulation. This study proposes a stress-based evaluation framework that reframes phosphene encoder assessment as a tri-objective operating-envelope problem rather than a single-metric comparison. Four representative encoders—rate, sparse, temporal, and optim—were evaluated under structured perturbations using two simulated prosthetic-vision(SPV) benchmarks: EMNIST Letters for symbolic recognition and a COCO-derived balanced subset for a reduced four-class COCO-derived image-level classification task. The final experimental configuration used 20,000/3000/3000 train/validation/test samples for EMNIST and 4000/1000/1000 for the COCO-derived benchmark. The framework combined structured stress sweeps, residual proxy safety-burden analysis, topology-based representation metrics, and stress-integrated utility summaries within a simulation-based evaluation setting. The results showed that clean-conditioning accuracy alone was insufficient to predict encoder behavior under stress. EMNIST exhibited broad operator-dependent degradation, whereas the COCO-derived benchmark showed a lower but more compressed performance regime. Residual proxy safety burden was only loosely aligned with performance, with moderate dissociation between performance and residual proxy safety burden in EMNIST and weaker alignment between these two axes in the COCO-derived benchmark. In the point-estimate utility summaries, the sparse encoder tended to yield comparatively favorable tri-objective utility values within the present single-run simulation-based framework, simplified SPV percept-synthesis operator, and fixed benchmark-specific decoder setting, primarily because it maintained an almost negligible residual proxy safety burden while preserving competitive performance and topology-based representation metrics. Topology-based analysis further indicated that topology-based representation metrics largely tracked task degradation in EMNIST, whereas topology-based representation metrics showed larger relative variation than decoder accuracy within the evaluated simulation setting under degraded COCO-derived conditions. Taken together, these findings provide an exploratory, benchmark-specific assessment suggesting that phosphene encoder evaluation may benefit from a multi-axis operating-envelope-oriented analysis that jointly considers stressed functional performance, residual proxy safety burden, and topology-based representation metrics within the present simplified SPV and fixed-decoder evaluation setting. These results should therefore be interpreted as simulation-level, configuration-dependent observations under a simplified SPV percept-synthesis operator, with safety-related quantities treated as residual proxy safety-burden summaries rather than as direct physiological, electrochemical, clinical, or implant-specific safety measurements. Full article