Search Results (2,700)

We investigate the convergence of the Levenberg–Marquardt (LM) iterative regularization method for linear ill-posed inverse problems in Hilbert spaces under general source conditions characterized by admissible index functions. We introduce admissibility conditions tailored to the spectral filter structure of the LM iteration, extending the classical convergence theory beyond Hölder-type source conditions to a broader class of index functions, including logarithmic source conditions relevant to severely ill-posed inverse problems. Using the filter representation of the LM iteration, we derive convergence rate estimates under both a priori and an a posteriori parameter choice rules. The analysis provides a unified framework for convergence analysis that covers a broad class of admissible index functions. Numerical experiments involving a time-fractional backward heat problem and a Fredholm integral equation of the first kind are consistent with the theoretical convergence results and demonstrate close agreement between the predicted error estimates and the observed reconstruction errors under both Hölder-type and logarithmic source conditions. Full article

Railway settlement requires long-term monitoring with millimeter-level accuracy; however, conventional leveling provides limited spatial coverage and requires considerable field effort. This study presents a practical assessment of selected StaMPS parameter settings for ComSAR-based railway settlement monitoring using Sentinel-1 InSAR, with conventional StaMPS PS-InSAR used as a comparative workflow. A ComSAR-based compressed interferogram stack and a conventional PS-InSAR workflow were applied to a section of the Honam High-Speed Railway using 61 Sentinel-1A IW SLC scenes acquired between 28 December 2018 and 29 December 2020. Track Concrete Layer (TCL) leveling data acquired at three annual epochs in December 2018, 2019, and 2020 were projected onto the radar line of sight (LOS) using local incidence angles. The InSAR LOS displacement time series were then compared with the leveling-based linear reference trend using trend-referenced RMSE and a 5 m spatial matching criterion. Four StaMPS parameters were evaluated: spatial resampling grid size before unwrapping (unwrap_grid_size), Goldstein filter window size (unwrap_gold_n_win), temporal window for phase unwrapping (unwrap_time_win), and temporal low-pass filtering window (scn_time_win). A univariate parameter-effect assessment was conducted, in which each parameter was varied individually while the remaining parameters were held at their reference settings. Among the four parameters examined, unwrap_grid_size produced the clearest first-order RMSE response in the ComSAR-based workflow under the reference settings used in this study. Grid sizes of 10–20 m produced clear degradation in trend-referenced validation performance, with the RMSE increasing to approximately 6.7–7.4 mm, whereas 40 m was the smallest tested grid size that recovered practically stable RMSE-based validation performance, at approximately 5.58 mm. In contrast, unwrap_gold_n_win, unwrap_time_win, and scn_time_win produced only small or localized numerical RMSE variations, with no consistent deterioration trend across the tested settings. The conventional PS-InSAR workflow also showed relatively stable RMSE-based validation performance across the tested parameter ranges, with RMSE values of approximately 4.87–4.91 mm. These findings provide case-study-based practical guidance for selecting StaMPS parameter settings in ComSAR-based railway settlement monitoring. Full article

Low-Dose Computed Tomography (LDCT) reduces radiation exposure but introduces severe quantum noise and streak artifacts that degrade image quality. To address these challenges, we propose the Physics-Guided Sequential State Space Transformer ($P{S}^{3}T$), a projection-domain denoising framework that combines sequential state-space modeling with a photon-aware attention mechanism to capture long-range dependencies across projection angles with linear computational complexity. A differentiable Filtered Backprojection (FBP) layer further enforces reconstruction-domain consistency during training. The proposed framework was evaluated on the Mayo Clinic LDCT and Projection Dataset using patient-level dataset partitioning. Experimental results demonstrate that $P{S}^{3}T$ consistently outperforms state-of-the-art methods, including DRL, SADiff, and GEDFormer, across the abdomen, head, and chest datasets. On the abdomen dataset, $P{S}^{3}T$ reached a peak PSNR of 42.40 dB, an SSIM of 0.9020, and the lowest RMSE of 0.0076 across anatomical regions. Statistical analysis using 95% confidence intervals and paired Wilcoxon signed-rank tests confirmed that these improvements were significant ($p<0.05$). Furthermore, $P{S}^{3}T$ achieved the lowest reconstruction consistency loss (0.0128 at epoch 50), demonstrating stable convergence and the effectiveness of incorporating acquisition physics into projection-domain LDCT denoising. Full article

The increasing adoption of machine learning models in corporate valuation has substantially improved predictive accuracy but at the cost of interpretability, a critical limitation in regulated financial environments. This study investigates whether SHAP (Shapley Additive Explanations) can systematically enhance the transparency of XGBoost based valuation models and examines whether the resulting insights extend those of classical linear regression. An empirical analysis was conducted on a cross-sectional dataset of U.S. publicly listed firms (2018), including more than 200 financial indicators. After systematic preprocessing and a hybrid feature selection procedure combining XGBoost importance, mutual information, and correlation based filtering, both an OLS regression and an XGBoost model were trained and validated. XGBoost achieved substantially higher predictive performance (R² = 0.654 vs. 0.364), while SHAP values provided transparent global and local explanations of model decisions. Both models consistently identified earnings before tax and EV to sales as primary value drivers; however, SHAP additionally showed nonlinear effects, threshold behaviors, and context dependent interactions, particularly for EBITDA margin, share buybacks, and asset based indicators, that remain undetectable in linear models. These findings confirm that SHAP significantly enhances model transparency and generates economically meaningful insights beyond classical regression, supporting its application in auditable, regulatory compliant financial modeling. Full article

Reliable identification of genomic regions controlling complex agronomic traits across variable growing seasons remains a major challenge in soybean genetics and breeding. Here, a diverse panel of 252 soybean accessions was evaluated over six consecutive growing seasons (2018–2023) for flowering time, maturity, plant height, number of seeds per plant, seed yield per plant, and thousand-seed weight. Whole-genome resequencing and variant filtering yielded 2,019,772 high-quality SNPs, and association signals were evaluated using Inclusive Integrative Input Multiple-locus Random-SNP-effect Mixed Linear Model (IIIVmrMLM), Bayesian-information and Linkage-disequilibrium Iteratively Nested Keyway (BLINK), and Multi-Locus Mixed Model (MLMM) together with linkage disequilibrium (LD)-based locus consolidation. Cross-model prioritization retained 21 high-confidence loci supported by all three GWAS models and distributed across 10 chromosomes. Among the identified loci, 18 overlapped or co-localized with previously reported SoyBase genes and QTLs, whereas three (q.VER2.13-1, q.YP.01-1, and q.TSW.15-1) showed no positional overlap with known genes and QTLs and were therefore considered presumably novel. These three loci were associated with flowering time, yield per plant, and thousand-seed weight, accounting for 1.60%, 2.13%, and 5.53% of phenotypic variation, respectively. Ten loci co-localized with genomic regions containing established soybean regulators, including E2, E3, GmDt2, and POWR1, support the biological plausibility of the association results. Integration of genomic position, functional annotation, and tissue-expression evidence prioritized 112 candidate genes across 18 loci. These findings provide a focused set of genomic loci and candidate genes for independent validation and further investigation of the genetic basis of soybean adaptation and yield formation under variable continental growing conditions. Full article

Pneumatic artificial muscles, including McKibben-type actuators, offer high power-to-weight ratio, compliance, and inherent safety, but their nonlinear pressure–contraction behavior, hysteresis, saturation, and load-dependent dynamics make accurate position control challenging. This study develops a practical data-driven workflow that derives a branchwise feedforward compensator and an LQI or PID controller from one open-loop characterization experiment. Quasi-static characterization first identifies a conservative control-ready voltage window. A bounded random excitation within this window is replayed with $4\mathrm{s}$ holds to expose terminal and transient behavior. The same experiment supplies branchwise discrete plant models and a feedforward lookup. Two open-loop-derived transient layers, voltage creep compensation and dynamic pressure referencing, are applied to the raw lookup before simulation. Four controller variants are compared on a common simulated closed-loop benchmark built from the identified plant: a feedforward-only baseline, a branchwise proportional–integral–derivative (PID) baseline, a base linear quadratic integral (LQI) controller with displacement and pressure feedback, and a velocity-state LQI extension with a filtered velocity estimate. A multi-metric optimization score balances tracking RMS, settled oscillation, command activity, saturation, and gain magnitude. The score selects the base LQI within the LQI family. The selected gains and transient layers are deployed in a real-time implementation with manually reduced position gains. The controllers are then evaluated on a common reference stream against the physical actuator. Although simulation metrics cannot be transferred directly to the real system, the combined-metric ranking of the controllers remains unchanged. Full article

Industrial water treatment plants are increasingly dependent on cyber–physical systems (CPS) and automated control processes for their operational safety and efficiency. However, the embedding of digital control networks exposes these critical infrastructures to sophisticated cyber–physical attacks, including malicious tampering with chemical dosing units and physical actuators. This paper proposes a robust, AI-enhanced anomaly detection framework designed to identify multi-stage malicious activities in water treatment systems using real-world industrial datasets. The proposed system is developed and validated on the Secure Water Treatment (SWaT) dataset, which contains multivariate sensor and actuator time-series data collected from a fully operational physical testbed under both normal operations and targeted cyber–physical attacks. First, high-frequency sensor noise is filtered, and cross-channel measurement reliability is maximized using a Kalman filter-based sensor fusion module. Subsequently, the fused-state vector is analyzed using an unsupervised Isolation Forest algorithm optimized for high-dimensional boundary isolation. To eliminate false negatives caused by stealthy, low-amplitude data injections that bypass purely statistical models, a deterministic, rule-based verification layer derived from physical process control logic is integrated. By integrating a discrete linear Kalman filter with an unsupervised Isolation Forest and deterministic physical rules, the framework effectively suppresses high-frequency sensor noise, achieving a 67.8% reduction in root mean square error (RMSE), while maintaining high detection accuracy across complex industrial attack scenarios. Experimental results demonstrate that the proposed hybrid framework yields superior detection capability, achieving a Precision of ≈95%, a Recall of ≈93%, a scenario-level F1-score of 94.1 % (alongside a sample-level F1-score of 21.5 %) and an edge inference latency of 0.6 ms, effectively demonstrating its suitability for deployment within simulated real-time industrial edge computing environments. The findings further confirm that combining statistical machine learning, state-space sensor fusion, and invariant physical process logic provides a resilient defense paradigm for securing critical industrial infrastructure against modern cyber–physical threats. Full article

Global Navigation Satellite System (GNSS)-based measurements are widely used for sports monitoring and outdoor activity analysis; however, consumer-grade smartphones often produce degraded trajectories, inaccurate elevation profiles, and unreliable pace estimates. This study proposes a multi-stage framework for reconstructing low-fidelity GNSS running trajectories using an averaged high-fidelity wearable GNSS reference proxy. The framework combines activity-window selection, trajectory filtering and route-consistent projection, reference-based elevation correction, and pace reconstruction using two complementary approaches: a Linear Acceleration Influence Model and a Physics-Based Model. The methodology was validated using three high-fidelity and three low-fidelity recordings collected on a shared 5.726 km route. Within the common activity window, raw low-fidelity observations had a pooled nearest-route RMSE of 30.72 m, whereas retained route-consistent assignments had a residual RMSE of 14.84 m. Aggregate pace agreement improved from 2.40 to 1.74 min/km RMSE and from 32.84% to 25.46% MAPE. Raw smartphone elevation had a pooled RMSE of 190.27 m relative to the adopted reference profile, supporting reference-based elevation substitution. A constant-velocity Kalman RTS baseline reduced positional RMSE from 30.72 to 29.05 m (5.4%), whereas the complete route-association procedure eliminated severe backtracking that remained after distance thresholding alone. The proposed framework provides a transparent and reproducible solution for reconstructing sparse consumer-grade GNSS activities while preserving explicit uncertainty. Full article

This study conceptualizes RE100 not merely as a voluntary corporate campaign, but as an evolving socio-technical system that intersects with South Korea’s path-dependent energy regime. Applying BERTopic to a comprehensive corpus of 21,901 raw news articles (reduced to 13,507 analytically relevant articles after noise filtering; January 2019–October 2025), we trace the systemic evolution and multi-layered governance of RE100. Grounded in social constructionism and institutional systems theory, the research reveals how the global sustainability norm becomes structurally embedded in South Korea’s industrial and economic systems through a partially sequential evolutionary process of externalization, objectivation, and internalization. BERTopic identified 39 analytically meaningful topics, subsequently organized into four interpretive clusters: (1) local government-centered construction, (2) spatial reconfiguration and industrial-policy formation, (3) corporate-led market institutionalization, and (4) ESG-driven corporate governance. Collectively, these clusters demonstrate that RE100 has evolved from a peripheral international initiative into a multilayered governance norm embedded across local governance, industrial infrastructure, corporate strategy, and ESG systems. The findings hold theoretical and policy implications for energy-transition governance and climate communication. This progression is recursive rather than strictly linear: corporate-level internalization (Cluster 4) is conceptually linked to, and may interact with, regional and industrial-policy externalization (Clusters 1–2), consistent with a non-linear reading of institutionalization in which corporate site-selection criteria shape local-government infrastructure competition. Cluster 1’s prominence reflects local governments’ discursive visibility rather than substantive policy authority. Notably, RE100’s “normalization” denotes routinized corporate compliance practices, not political consensus on its legitimacy relative to alternatives such as carbon-free 100% or nuclear power. Full article

Air pollution poses a major environmental and public health challenge in Metropolitan Lima, Peru, where complex topography, coastal meteorological conditions, and intense urbanization generate highly dynamic patterns of pollutant concentrations. Accurate multi-horizon forecasting is therefore essential for supporting environmental monitoring and early-warning systems. This study proposes a hybrid deep learning framework, called VMD-Attention-BiLSTM, for forecasting hourly PM_2.5, PM₁₀, and NO₂ concentrations using a decade of hourly air quality observations (2015–2024) collected by the National Meteorology and Hydrology Service of Peru (SENAMHI). The proposed methodology integrates a strictly causal preprocessing pipeline—including forward-only imputation, spatial-corroborated percentile-95 outlier detection, and pollutant-calibrated Variational Mode Decomposition (VMD)—with a Bidirectional Long Short-Term Memory (BiLSTM) network enhanced by a Bahdanau-style attention mechanism. All transformations are fitted exclusively on the training partitions of a five-fold TimeSeriesSplit cross-validation to prevent information leakage. A systematic benchmark of 1008 imputation experiments was conducted to justify the choice of causal linear interpolation over Kalman Filter alternatives. Model performance was evaluated at 24-, 48-, and 72-h forecasting horizons. The optimized framework achieved competitive predictive performance across the evaluated horizons, yielding best RMSE values of 9.04, 19.93, and 9.41 µg/m³ for PM_2.5, PM₁₀, and NO₂ at 24 h, degrading to 9.79, 24.56, and 11.02 µg/m³ at 72 h. All metrics are reported on the original concentration scale. VMD sensitivity analysis revealed that the optimal mode count is pollutant-dependent ($K=12$ for particulate matter; $K=4$ for NO₂). Furthermore, the ablation study showed that the complete VMD-Attention-BiLSTM configuration provided competitive and frequently improved performance relative to the baseline and partial configurations, with the magnitude of the improvement varying according to pollutant and forecasting horizon. The obtained results indicate that integrating signal decomposition with attention-based bidirectional learning can provide complementary benefits for forecasting under highly non-stationary urban conditions, particularly at shorter forecasting horizons. The proposed framework provides a reproducible and scalable solution for intelligent air-quality forecasting and serves as a valuable decision-support tool for environmental monitoring and public health protection in complex metropolitan environments. Full article

Background: Modern lifestyles characterized by artificial light exposure, screen use, and altered sleep routines may be related to eating behaviors and dietary patterns relevant to the gut microbial environment. Although these factors have been investigated individually, limited evidence has examined self-reported light-exposure behaviors, sleep, eating behavior, and gut microbiota-related dietary patterns simultaneously within a single analytical framework. Objective: The aim of this study was to examine associations between five self-reported light-exposure behaviors and the global PSQI score, eating-behavior domains, and FFQ-GM-derived gut microbiota-related dietary pattern scores among adults in the recruited sample. Methods: A cross-sectional study was conducted among 518 adults aged 18–64 years recruited in Amman, Jordan, between April and May 2026 using convenience sampling. Participants completed the LEBA, the Arabic PSQI, the Arabic TFEQ-R18, and an Arabic-adapted gut microbiota-focused food frequency questionnaire (FFQ-GM). Associations were examined using Spearman’s rho and multivariable linear regression; multiplicity across the 25 LEBA factor–outcome coefficient tests was addressed using the Benjamini–Hochberg false discovery rate. Results: In the analyzed sample, 87.3% of participants had a global PSQI score > 5. Several LEBA factor–outcome coefficients had nominal p values < 0.05. Blue-light-filter use was positively associated with the global PSQI score (β = 0.29, p = 0.020), whereas phone use in bed (β = −0.43, p = 0.002) and light exposure before bedtime (β = −0.51, p = 0.009) were negatively associated with the global PSQI score. After false-discovery-rate adjustment across the 25 individual coefficient tests, most individual associations did not remain below the 0.05 threshold; therefore, these coefficient-level findings are interpreted as exploratory. Conclusions: Self-reported light-exposure behaviors were associated with questionnaire-derived sleep scores, eating-behavior domains, and FFQ-GM-derived gut microbiota-related dietary pattern scores. The cross-sectional design does not establish temporal direction, causality, or biological mechanism. Full article

In this paper, we present an algorithmic modeling approach based on Disjunctive Programming using Boolean logic “OR” to solve the Optimal Phasor Measurement Unit Placement (OPP). We propose a framework for modeling the optimal PMU placement subject to disjunctive constraints. A convex objective function is minimized subject to a bilinear equality constraint with a piecewise linear structure. The polynomial constraint constitutes a union of linear segments conceptually analyzed in the two-dimensional continuous space, separating the infeasible from the feasible region. This work investigates the trajectory of iterates from infeasible initial points to stationary solutions and analyzes the convergence behavior using Interior-Point Method (IPM) and Sequential Quadratic Programming (SQP). Our algorithmic model addresses the progress of infeasible and feasible iterates, step computation using line-search and trust-region mechanisms. Combined with second-order correction (SOC) and filter methods, these mechanisms enable the algorithm to maintain a unit primal step, even when starting from an infeasible initial point. Network observability constraints are transformed from a Conjunctive Normal Form (CNF) into a Disjunctive Normal Form (DNF) via Balas’s theory. This geometry transformation reformulates the feasible set into a union of convex affine pieces, effectively eliminating constraint curvature issues. This affine reformulation ensures that gradient-based algorithms maintain a smooth optimization trajectory along a convex local manifold. This trajectory enables the algorithm to preserve the full Newton step, maintaining a superlinear convergence rate. Its underlying piecewise linear convexity inherently enables the gradient-based algorithm to avoid the Maratos effect. Numerical results on IEEE power systems validate the optimization problem. Our framework uses the IEEE-14 bus system to address the high-degree non-convexities. Monte Carlo simulations further enhance the argument that piecewise convexity enables IPM and SQP to converge to binary local minima. Depending on multiple-run initialization, these methods reach the same objective function value. These local minima can be characterized as non-strict optimum points that are structurally symmetric but exhibit unequal basins of attraction. Hence, this geometry-driven formulation enables gradient-based methods to reliably identify binary-valued optimal solutions for the OPP. Full article

Clock steering, the core time-frequency technology for high-precision single-satellite time reference generation in global navigation satellite systems, can effectively combine the excellent short-term stability of Oven-Controlled Crystal Oscillators (OCXOs)—whose top performance indicators have already surpassed many space-borne atomic clocks in recent years—with the superior long-term stability of atomic clocks, to obtain time signals with optimal full-range stability. This paper proposes a novel clock steering scheme that integrates an adaptive variational Bayesian Kalman filter and a Proportional-Integral-Derivative (PID) automatic controller: the filter constructs a separable variational approximation for the joint posterior distribution of clock states and measurement noise parameters, to achieve real-time adaptive estimation of noise at each timestamp, while the PID controller performs closed-loop fine adjustment on the output frequency. Comparative simulations with the classic Linear Quadratic Gaussian (LQG) control scheme verify that the proposed method can generate steered time signals with better stability performance in both short-term and long-term dimensions. This work further investigates the influence of measurement noise at different intensity levels on clock steering performance and conducts corresponding mechanism analysis supported by quantitative data. The proposed scheme and conclusions can provide a reliable reference for selecting appropriate clock steering strategies under different noise conditions. Full article

PICO is an ultra-lightweight substitution–permutation network block cipher designed for resource-constrained devices such as Internet of Things terminals and edge agents. For fixed endpoints, summing the characteristic probabilities over an enumerated finite weight window gives a verifiable lower bound on the differential probability. We use a PICO-specific workflow that combines mixed-integer linear programming bounds on active substitution boxes, exact-weight Boolean satisfiability search, optional Matsui pruning, and fixed-endpoint enumeration. For the selected endpoints, enumeration over $W=63,\dots ,76$ and $W=66,\dots ,79$ gives finite-window lower bounds of $2^{-59.95}$ and $2^{-61.95}$ for 21 and 22 rounds, respectively. We prepend two rounds and append three rounds to the 21-round differential distinguisher. The resulting 26-round analysis is an analytical equivalent-round-key filtering-and-ranking procedure for a 108-bit tuple. The verified 21-round finite-window probability input is a factor of $2^{0.8032}\approx 1.745$ larger than the previously reported input, increasing the expected right-tuple support at fixed S under the analytical accounting. For the illustrative choice $S=2^{42}$, the analytical resources are $D=2^{62}$ chosen plaintexts, a normalized substitution-box filtering workload of $T=2^{100.11}$ equivalent 26-round encryptions, and $M=2^{62}$ stored plaintext–ciphertext records. This setting is not tied to a demonstrated success probability and does not establish an equal-success complexity advantage over prior work. Full article

Exposure to atmospheric particulate matter (PM) is a major public-health concern, in part because PM can perturb cellular redox homeostasis. This study evaluates an in vitro A549/DCFH-DA approach using repeated fluorescence readouts to assess PM2.5-induced oxidative activity. Untreated and assay-specific controls were combined with zymosan and NIST Standard Reference Material^® 2584 suspended in PBS, and fluorescence was monitored at multiple readout times over a 15 min–6 h window. Method performance was characterized using the coefficient of variation (CV) and signal-to-noise ratio (SNR). A dedicated three-concentration SRM 2584 series (0.02, 0.05 and 0.10 mg mL⁻¹) further showed readout-dependent concentration behaviour: at 60 min the untreated-control-corrected mean response increased across the tested concentrations and followed an approximate descriptive linear trend (R² = 0.90), whereas earlier readouts were non-monotonic. Substrate-related effects were examined using paired PTFE and quartz filters. Among the eight matched PTFE–quartz pairs included in the regression analysis, zero-intercept fits showed slopes close to unity for both mass- and air-volume-normalized responses (0.90 and 0.99, respectively; R² ≈ 0.99), demonstrating strong proportional agreement within this comparison set; the limited number of pairs does not support universal substrate interchangeability. Application to chemically characterized field PM2.5 samples from an urban-background site and a high-altitude site showed that DCFH-DA fluorescence did not track PM mass alone and is interpreted in terms of exploratory associations with particle composition, rather than causal effects of individual constituents. Taken together, these findings support the use of the method-performance-characterized workflow for assessing oxidative responses to field-collected PM2.5 across multiple readout times and for investigating their associations with particle chemical characteristics. Full article