Search Results (40)

Accrual accounting reform can be formally adopted before the internal audit routines needed to assure it are operationally embedded. This exploratory diagnostic study examines perceived implementation risk among 95 internal audit professionals working in Saudi governmental entities. A content audit of an existing questionnaire showed that its original accrual-basis accounting application (ABAA), senior management support (SMS), and internal audit effectiveness (IAE) blocks combined conceptually adjacent benefit, support, resource, and outcome statements. The analysis therefore uses theory-guided, content-separated diagnostic domains and an item-level timeliness model rather than treating the original blocks as discriminant latent constructs or testing mediation. Within-respondent contrasts were assessed using paired t-tests and Wilcoxon signed-rank checks. Because timely completion is a five-category ordinal outcome, ordinal regression is the scale-respecting approach. The full five-predictor ordered-logit model was retained as a theory-complete diagnostic because E4 explicitly includes staffing, but its global proportional-odds restriction was rejected due to staffing. Accordingly, substantive ordinal inference and model-implied probabilities are based on the assumption-compatible reduced model excluding staffing, while the full model is reported transparently as a diagnostic sensitivity specification. Formal reform orientation exceeded applied reform capacity by 0.437 points (p < 0.001; dz = 0.441); visible governance and monitoring exceeded operational audit infrastructure by 0.674 points (p < 0.001; dz = 0.830); and perceived staffing sufficiency exceeded timely task completion by 0.800 points (p < 0.001; dz = 0.615). In the assumption-compatible reduced ordered-logit model, annual planning (b = 0.581, p = 0.016) and the accrual-aligned internal audit guide (b = 1.570, p < 0.001) were positively associated with timeliness; the corresponding full-model estimates were nearly identical. The OLS-HC3 robustness model accounted for 61.2% of respondent-level variation, although same-source response tendencies may contribute to this fit. The observed pattern is compatible with uneven layered institutionalization and a risk of decoupling between visible or formal elements and operational routines; it does not establish temporal sequencing, sector-wide prevalence, intentional symbolic compliance, or causal effects. Conclusions are bound by the cross-sectional, same-source design and the post hoc diagnostic use of an existing instrument that requires prospective validation. Full article

Second-line chemotherapy is widely used in metastatic urothelial carcinoma after progression on first-line platinum-based therapy, but its independent contribution to survival, as opposed to selection of healthier patients, remains unclear. In this multicenter retrospective cohort of 142 patients treated with first-line platinum-based chemotherapy across seven Turkish centers, overall survival (OS) from first-line progression was compared between patients who received second-line chemotherapy (n = 80) and those who did not (n = 62), using multivariable Cox regression, inverse probability of treatment weighting (IPTW), propensity-score matching, landmark analysis, and a time-dependent Cox model. Median OS was 7.4 versus 4.7 months (log-rank p = 0.064). Second-line chemotherapy was independently associated with improved OS on multivariable analysis (adjusted hazard ratio [aHR] 0.620; 95% confidence interval [CI] 0.423–0.907; p = 0.014); Eastern Cooperative Oncology Group (ECOG) performance status ≥ 2 (aHR 3.881; p = 0.001) and lower albumin (aHR 0.671; p = 0.018) were also independent predictors. The association remained significant after IPTW (HR 0.648; p = 0.025) and after a time-dependent Cox model (HR 0.632; p = 0.019), and was unchanged in ECOG-restricted and Bellmunt-adjusted analyses (p = 0.008, p = 0.029); it narrowly missed significance after propensity-score matching (HR 0.645; p = 0.051) and did not reach significance in the 3-month landmark analysis (HR 0.743; p = 0.180). Power was limited (~49%). In a time-dependent Cox model—the analysis least susceptible to immortal-time bias, as it retains the full cohort and classifies pre-treatment person-time as unexposed—second-line chemotherapy remained independently associated with improved OS (HR 0.632; p = 0.019), closely consistent with the primary multivariable estimate. The conventional Cox, IPTW, and propensity-score-matched analyses, which treat second-line receipt as a baseline exposure, were directionally concordant but share a common time-related bias and are therefore not independent confirmations. ECOG performance status was a consistent predictor throughout. Full article

Planning renewable energy bases with chronological source-grid-storage simulation makes exhaustive capacity screening impractical. This study develops a phase-adaptive constrained active-sampling framework whose core mechanism is a simulator-verified phase switch: a probability-of-feasibility-weighted lower confidence bound (PoF-LCB) directs the search until the first verified feasible plan is found, after which constrained expected improvement (CEI) directs economic refinement, supplemented by bounded optimal-neighborhood and constraint-boundary ranking refinements. Gaussian-process surrogates decide only the evaluation order; the reported objective and all four engineering constraints—photovoltaic curtailment, loss-of-load energy, capacity credit, and flexibility scarcity—are verified exclusively by the original 8760 h simulator. In a paired 2 × 2 factorial experiment over 30 common initial designs on a 125-candidate pool, the phase switch raised exact-optimum recovery from 24/30 to 30/30 (exact McNemar p = 0.03125), and the full rule maintained 30/30 under two unseen profile seeds where CEI achieved 24/30 and 22/30. The full rule further recovered the exact optimum of a 1224-candidate pool in 30/30 runs within 20 evaluations and of a six-variable 729-point grid within 75 evaluations, using roughly 2–10% of the exhaustive simulation budget. Constraint-slack and guard-band reporting, candidate-domain audits, and repeated wall-clock measurements turn the recommendation into auditable planning decisions, with all evidence drawn from a reproducible synthetic benchmark. Full article

The recently proposed Animated Oat Optimization (AOO) algorithm exhibits competitive search behavior, but its fixed branching rules and limited use of inter-individual information may cause diversity loss and premature stagnation. This study proposes an Improved Animated Oat Optimization algorithm (IAOO) that integrates the original AOO operator, a hybrid DE/rand/1–DE/best/1 operator with a decreasing scale factor, and an elite neighborhood-directed local search within a feedback-driven framework. Strategy probabilities are updated according to normalized successful fitness gains, enabling search effort to adapt to the current optimization state. IAOO was evaluated through 30 independent runs on the CEC2017 (dim = 30/100), CEC2020, and CEC2022 suites and achieved Friedman mean ranks of 1.50, 1.37, 2.70, and 1.83, respectively, achieving competitive Friedman mean ranks among the compared algorithms and demonstrating statistically supported advantages on most benchmark suites. In three-dimensional UAV reference-path planning, IAOO reduced the mean path cost from 406.26 for AOO to 298.94, corresponding to a 26.4% reduction, while the standard deviation decreased from 67.01 to 40.73. Its runtime increased only from 26.99 s to 27.13 s. These results indicate that feedback-based operator cooperation improves solution quality and robustness with limited computational overhead. The current UAV model produces geometrically feasible and kinematically constrained reference paths; full six-degree-of-freedom tracking validation remains future work. Full article

Online platforms need scalable risk-screening methods for rapidly spreading rumors, misleading content, and large-scale user responses. Platform-side content governance often needs to rank suspicious events when only partial propagation evidence is visible, so that limited review resources can be assigned first to high-risk events. This article focuses on that setting. We propose a Structure-Aware Bidirectional Graph Convolutional Network (SA-BiGCN), which combines a BiGCN-style propagation graph encoder with structural statistics computed from the currently visible propagation tree to estimate event-level rumor probability. We construct Main Weibo V2 from 4664 raw Weibo source posts and their propagation structures, encode node text with a unified vocabulary, and use fixed event-level splits. The model evaluates source-post veracity at Top 10, Top 30, Top 50, Top 100, and full observation windows according to a timestamp-audited propagation order, using only the currently visible propagation-tree subgraph. The experimental results show that SA-BiGCN reaches 92.72% Avg F1 and 91.28% Worst F1 on Main Weibo V2 while using one shared checkpoint for all observation windows. Feature-set, leave-one-feature-group, sampling-strategy, topology-corruption, deployment-cost, and transfer analyses indicate that structural statistics computed from the currently visible propagation tree provide auxiliary evidence mainly in early-window and weakest-window settings. We further extend SA-BiGCN with graph-attention, temporal-quality, response-style, window-aware, and RootRoBERTa fusion variants. These additional cues are retained as diagnostic trends and do not replace the main model. Overall, SA-BiGCN is best understood as a lightweight single-checkpoint propagation-tree scorer. Full article

The design optimization of electromechanical linear actuators (EMLAs) is challenged by coupled tribological, thermal, and dynamic effects. Such interactions can violate the additivity assumptions of compact orthogonal designs. Factor ranking may become unreliable when residual variance includes non-additive contributions rather than stochastic noise alone. This study presents an Extended Orthogonal Experimental Matrix (E-OEM) framework for robust factor screening under experimental-budget constraints. It preserves the reduced effort of an orthogonal design while using a latent allocation term to retain structured non-additive variation that would otherwise be treated as residual error. It combines compact experimentation, fitted probability distributions, Monte Carlo simulation, and weighted multi-criteria ranking under push force, robustness, and cost targets. The framework is demonstrated on 90 automotive EMLAs considering lubricant type, drive frequency, and operating temperature. E-OEM was benchmarked against the full-factorial experiment (27 combinations, 810 push-force measurements). Under the selected weighting strategy, both E-OEM and the benchmark identified the same nominal optimum (590 Hz, +25 °C, Addinol lubricant), with a measured push force of 77.32 ± 2.05 N (1.01% deviation from the prediction). The results show that E-OEM provides an efficient screening and decision-support approach for actuator design when full-factorial testing is constrained by time and cost. Full article

In the Argoverse 2 experiments reported here, the simplest Constant Turn Rate and Acceleration (CTRA) decoder was stable but missed many interaction-driven turns and merges, while residual decoders without enough control improved early displacement but increased final-horizon error. This paper therefore studies a compact decoder in which each of the six futures is represented as a CTRA anchor plus an autoregressive position residual. The residual gated recurrent unit (GRU) is initialized from fused target-history, top-k neighbor, and lane-polyline context, and its contribution is scaled by a mode-specific gate and learned exponential decay. On the 10k/2k sanity ablations, CTRA-only decoding reached ${\mathrm{minFDE}}_6=7.189$ m, while autoregressive residuals with a larger correction GRU reduced it to $4.157$ m; removing the gate increased it again to $4.946$ m. On the full Argoverse 2 validation split, the final configuration achieves a minimum average displacement error of ${\mathrm{minADE}}_6=1.21$ m and a minimum final displacement error of ${\mathrm{minFDE}}_6=2.78$ m. The reported diagnostics show that the compact model generates a useful six-mode set, but still needs better probability ranking for top-1 selection. Full article

We extend the phase-tagged fluctuation framework for cumulative shock models from a fixed failure threshold to a linearly moving boundary $u_0+c{\tau }_n$, the premium drift regime that underlies insurance ruin theory. The moving boundary turns the first-passage problem from a fixed-level crossing into a crossing of a drifting walk, for which the partial sum truncation of the fixed-threshold theory no longer applies. We resolve this with a drift-aware ladder apparatus: a rank-one drift ladder matrix${\mathit{G}}^{\uparrow }\left(c\right)=g\left(c\right)\mathbf{e}\mathit{\alpha }$ whose scalar $g\left(c\right)$ solves a discrete Lundberg equation $a\left(z\right)=z^c$, a closed-form ascending ladder-height law obtained from the roots of the Lundberg polynomial, and a compound-geometric (Pollaczek–Khinchine) representation of the ruin probability. Building on these, we derive in closed form the full five-variable moving-boundary reliability functional ${\mathbf{\Phi }}_{\nu }^{\mathrm{ruin}}(\xi ,u,v,\vartheta ,\theta )$, the drift analogue of the fixed-threshold phase-tagged functional, and show it lies in the same two-dimensional matrix subspace $\mathrm{span}\left\{H\right(\theta ),H(\omega \left)\right\}$: the premium drift deforms the scalar coefficients through a first-passage transform while leaving the matrix structure invariant. As the principal application, we obtain the phase-resolved Gerber–Shiu expected-discounted-penalty function, with the joint transform of the time of ruin, the deficit at ruin, and the surplus prior to ruin. The classical scalar Gerber–Shiu function and the fixed-threshold functional are recovered as projections and as the $c\to 0$ limit, respectively. All closed forms reduce to root-finding on a single polynomial for rational model primitives, and every structural result is verified against Monte Carlo simulation and exact recursion. Full article

Galloping of iced transmission lines under complex microclimates poses a severe threat to power grid security, whereas traditional full-span de-icing strategies suffer from excessive energy redundancy and limited spatial precision. To address the spatial uncertainty of actual ice accretion, a three-dimensional nonlinear aeroelastic finite element model is established by considering geometric nonlinearity and eccentric ice-induced added stiffness. A state-space Monte Carlo framework is then used to evaluate the galloping probability under different wind speed regimes and spatially non-uniform ice distributions. The results reveal a distinct non-monotonic instability characteristic: the galloping probability decreases to 33.0% at 8.0 m/s, forming a clear probability trough and indicating an aerodynamic self-stabilization effect associated with the shift in the baseline effective angle of attack. To map spatial ice heterogeneity to global dynamic instability, a galloping sensitivity index (GSI) based on the Spearman rank correlation coefficient is proposed to identify the dominant sensitive sections responsible for inducing galloping-prone responses. Based on this index, a GSI-guided targeted ultrasonic de-icing decision strategy is constructed. Under the assumption of identical rated power for each section, the proposed strategy activates only 40% of the physical sections and reduces the number of activated sections, as well as the associated operational energy demand, by 60% compared with the full-span de-icing strategy. This framework provides a quantitative basis for linking stochastic ice distribution, galloping probability evaluation, and energy-efficient targeted de-icing decisions. Full article

Large language models (LLMs) built upon the Transformer architecture have achieved remarkable performance in natural language understanding, text generation and logical reasoning, while their internal working mechanisms remain poorly interpreted. This paper establishes a systematic mathematical analysis framework tailored for decoder-only Transformer LLMs, based on linear algebra, tensor analysis, probability theory, information theory, optimization dynamics and geometric deep learning. We conduct rigorous mathematical modeling and theoretical deduction on core modules including word embedding, position encoding, self-attention, feed-forward networks, training optimization and generalization reasoning, and explore the mathematical nature of semantic representation, contextual correlation, knowledge storage and logical inference within models. In this paper, we strictly distinguish between classic established Transformer theories and our original mathematical derivations and conclusions. Distinct from existing fragmented theoretical studies, this work presents six targeted novel contributions beyond conventional Transformer theories: (1) we construct the first full-process unified mathematical framework covering all core modules and the entire lifecycle of Transformer-based LLMs; (2) we provide strict mathematical proof to verify that single-head self-attention is essentially a kernel weighted average operation in reproducing kernel Hilbert space and derive the low-rank and sparse properties of attention weights; (3) we establish a high-dimensional non-convex optimization dynamics model for pre-training and mathematically prove that model training converges to flat local minima; (4) we derive a tighter upper bound of generalization error and quantify the quantitative relationship among model parameters, sequence length, training data scale and generalization performance; (5) we characterize the latent space as a low-curvature smooth Riemannian manifold and model logical reasoning as geometric transformation on this manifold; (6) we design multi-group controlled experiments on mainstream datasets to quantitatively validate all above theoretical conclusions. This paper further summarizes the inherent mathematical limitations of current Transformer LLMs and proposes feasible theoretical optimization paths, referring to state-of-the-art research published from 2021 to 2026. The outcomes of this research can provide solid mathematical theoretical support for improving model interpretability, optimizing network structures and boosting practical performance, and facilitate the transition of LLM research from empirical engineering practice to theory-driven development. Full article

Existing post hoc XAI (Explainable Artificial Intelligence) methods produce numerical attributions without symbolic structure (SHAP, LIME), low-coverage local rules (Anchors), or crisp tree surrogates without an interpretable rule-level uncertainty proxy. We present a fuzzy rule-based explanation framework for tabular black-box classifiers, extracting global IF–THEN rules with linguistic labels. This was validated on a 13-dataset benchmark with four model families (Wilcoxon, Friedman, TOST equivalence): (i) prediction-boundary-aware fuzzy partitioning raises mean fidelity from a vanilla Wang–Mendel baseline of 0.736 to 0.893 (+10.4 pp excluding the Breast Cancer outlier; +15.7 pp aggregate, both transparently reported); (ii) fired-rule consequent entropy provides a zero-cost rule-level uncertainty proxy (Spearman ρ = 0.420 with model prediction entropy, significant on 11/12 datasets—moderate by Cohen’s convention, with a 4/12 weak-correlation tail; complementary to probability-entropy and margin baselines). Fidelity is statistically equivalent to tree surrogates on classification (TOST p = 0.002, δ = 0.05) at ≈100% coverage. SHAP/LIME are excluded from the formal stability ranking because the perturbation metric measures the wrapped black-box rather than the attribution vector; cross-explainer comparison is reported in grouped form (full-coverage surrogates vs. local-coverage methods). On continuous regression (California Housing fidelity 0.422 vs. TreeSurrogate 0.840) and XOR-type multi-feature interactions, the framework is structurally weaker, addressed by a planned TSK extension. Full article

Contemporary vulnerability management relies on the Common Vulnerability Scoring System (CVSS) and the Exploit Prediction Scoring System (EPSS), both of which evaluate Common Vulnerabilities and Exposures (CVE) entry in isolation, disregarding the network topology in which vulnerable components operate. We present the Dynamic Security Resistance Distance (DSRD) framework, which parses Docker Compose, GNS3, and Containerlab configuration files into weighted attack graphs where edge conductance reflects EPSS exploitability. A version-aware filtering stage matches discovered CVEs against the software versions declared in container image tags, reducing version-irrelevant CVE matches by up to 97%. Kirchhoff effective resistance, computed via the Moore-Penrose pseudoinverse of the graph Laplacian, yields a structural compromise affinity—a monotone score guaranteed not to increase upon patching. Four algorithms—Ant Colony Optimization, Physarum, Fungal Network Growth, and Greedy Kirchhoff-rank vulnerabilities by their structural impact on network-wide risk. Evaluation on nine representative topologies derived from public IaC artifacts, spanning six Docker Compose and three GNS3 deployments, with 895 version-relevant vulnerability nodes from cvelistV5 shows that graph-aware prioritization reduces structural risk by up to $5.62\times {10}^{-2}$ after ten patches, whereas EPSS-only ordering achieves at most $1.28\times {10}^{-2}$ on the same topology. EPSS-only targets high-probability CVEs on entry points that do not lie on critical paths; graph-aware methods instead prioritize CVEs on high-resistance paths toward critical assets. The advantage depends on infrastructure heterogeneity and topology structure: topologies with diverse vendors and well-defined structural bottlenecks benefit most, while densely connected or homogeneous environments show marginal improvement. We release the full pipeline as open-source software. Full article

Angular misalignment of spherical roller bearings in wind turbine main shafts is a known cause of premature failure. Manufacturing and assembly tolerances introduce unavoidable variability in this misalignment—a source of uncertainty typically neglected in deterministic life models, thereby creating a gap between installation quality and system reliability. A probabilistic framework combining a Hertzian contact model, the Ioannides–Harris fatigue theory, and Monte Carlo simulation is developed to predict the fatigue life of double-row spherical roller bearings under uncertain misalignment. The sensitivity of eight geometric parameters, selected based on manufacturing tolerances, is quantified using Sobol indices for global sensitivity analysis, allowing their relative importance to be ranked. Application to a 950-series wind turbine main bearing under nominal and extreme loads shows that even with centered installation a non-negligible failure probability persists under nominal conditions. The strongly asymmetric bearing response requires asymmetrical installation tolerances to ensure high reliability. Global sensitivity analysis identifies the misalignment angle as the dominant source of uncertainty, followed by the roller contour radius. Under extreme loads, the bearing is under-dimensioned relative to the 20-year design life required for wind turbine main bearings, leading to a fatigue failure probability that approaches unity regardless of installation quality. The interaction between misalignment and radial clearance becomes pronounced under extreme overloads. Overall, the proposed framework provides a quantitative basis for reliability-based tolerance specification and emphasizes the necessity of considering the full load spectrum—including assembly variability—in bearing design. Full article

Neltuma juliflora (Sw.) Raf. (syn. = Prosopis juliflora (Sw.) DC.) is among the world’s most aggressive woody invaders, yet its ecological impacts remain poorly quantified in hyper-arid environments, where soils are calcareous and ecosystems recover slowly from disturbance. In this study, we tested two hypotheses: (1) the presence of N. juliflora changes native plant diversity, as well as soil and key physicochemical properties in hyper-arid Qatar, and (2) agricultural farms act as primary sources of N. juliflora invasion. Using a comparative observational design across 62 sites (45 invaded and 17 non-invaded), we applied a generalised additive model (GAM) and a generalised linear mixed model (GLMM) to quantify invasion drivers and the impact of invasion on perennial species diversity, respectively. Additionally, we used the Wilcoxon rank-sum test to compare the soil properties in the invaded and non-invaded sites. Our results indicate that N. juliflora is positively associated with farms, with the probability of occurrence declining by ca. 20% for each kilometre farther away from agricultural farms. This pattern suggests substantial propagule pressure from agricultural farms. Perennial species richness declined from 7.5 species at 0% N. juliflora cover to 4.8 species at full cover (36% reduction). Invaded sites were characterised by higher amounts of coarse sand (16%); reduced silt–clay fractions (5%); and elevated salinity indicators, including electrical conductivity (0.744 dS m⁻¹) and total dissolved solids (476 mg L⁻¹), while major N–P–K pools remained unchanged. These findings demonstrate measurable invasion-related changes in soil conditions and native perennial diversity in hyper-arid ecosystems and highlight the role of agricultural land use as a key driver of biological invasion. From a sustainability perspective, early detection, targeted control near agricultural and grazing zones, and integration of invasive species monitoring into land-use planning frameworks are essential to prevent further ecosystem degradation, protect biodiversity, and enhance the resilience of desert landscapes under increasing climate and land-use pressures. Full article

The paper presents the Failure Mode and Effects Analysis (FMEA) method applied to a process-based case study, together with an approach for implementing the AIAG & VDA harmonized FMEA standard by using modern digital tools. While classical FMEA is widely used in the industry, risk assessment based on the Risk Priority Number (RPN) often leads to the inconsistent ranking of failures and unclear prioritization of corrective actions. This paper explores the shift from the traditional Risk Priority Number (RPN) approach to the Action Priority (AP) concept introduced in the AIAG & VDA FMEA Handbook and explains why this change leads to clearer, more consistent risk-based decisions. Rather than focusing only on the methodological differences, the paper also outlines a practical framework for full implementation, showing how Industry 4.0 technologies can strengthen traceability, improve response time, and ensure greater consistency in PFMEA development. It also examines how Artificial Intelligence (AI) and Large Language Models (LLMs) can support engineers in everyday practice—for example, by helping identify potential failure modes, standardizing documentation, and guiding the definition of prevention and detection controls. In parallel, IoT-based monitoring and real-time data collection can provide valuable feedback to validate occurrence and detection ratings. Over time, this data-driven feedback loop can improve the accuracy and reliability of risk assessments. The proposed framework contributes to improved responsiveness in process optimization activities, reduces the probability of recurring failures, and supports continuous quality improvement in manufacturing organizations. The solution is discussed in relation to classical FMEA practices and recent trends in the digital transformation of quality management systems. Full article