Abstract: The Job Shop Scheduling Problem (JSSP) is a paradigmatic and strongly NP-hard combinatorial optimisation problem that underpins production planning in modern manufacturing systems, and constraint programming (CP) has become one of the leading methodologies for tackling it. However, comparative studies of CP solvers for the JSSP have so far been restricted to a single benchmark family, a single instance-size range, or a single hardware setting, which limits the practical guidance they offer to both researchers and practitioners. This paper presents a controlled empirical evaluation of four state-of-the-art CP solvers—IBM ILOG CP Optimizer, Google OR-Tools (CP-SAT), Hexaly, and OptalCP—on the makespan-minimisation JSSP. The four engines are run with default parameters and a uniform 600-second wall-clock time budget on 332 instances drawn from nine canonical benchmark families (Fisher–Thompson, Lawrence, Adams–Balas–Zawack, Applegate–Cook, Yamada–Nakano, Storer–Wu–Vaccari, Taillard, Demirkol–Mehta–Uzsoy, and Da Col–Teppan), spanning sizes from 6 × 5 up to 1000 × 1000 operations. OptalCP emerges as the most robust engine overall, certifying optimality on 57.5% of the instances with the smallest average optimality gap (3.55%), while Hexaly dominates on industrial-scale problems and produces the bulk of 31 new best-known upper bounds and one new best-known lower bound reported here. Solver competitiveness depends sharply on instance size and on the n/m ratio, with square instances confirmed as the hardest case. These findings support an instance-aware approach to CP solver selection in industrial scheduling.

Abstract: We propose a framework that isolates a precise complexity-theoretic bottleneck between counting complexity and the Birch--Swinnerton-Dyer conjecture (BSD) via Tunnell's theorem. The framework rests on two number-theoretic conjectures: a \emph{Reduction Conjecture} asserting the existence of a polynomial-time reduction from any \#P-complete problem to the counting of integer representations $D_n = \#\{(x,y,z) : n = 8x^2 + 2y^2 + 16z^2\}$ (with counts preserved up to a polynomial factor), and a \emph{Solution Density Conjecture} asserting that the values $\{D_n : n \text{ even square-free congruent}\}$ are sufficiently densely distributed (within the Eichler--Deligne ceiling $D_n = O(n^{1/2+\varepsilon})$) to support iterated polynomial descent. We do \emph{not} claim that $\text{P} = \text{NP}$ implies $\#\text{P} = \text{FP}$ (the natural binary-search route fails because the threshold predicate $[\#I \geq k]$ is PP-complete, not in NP, and PP is not known to collapse under $\text{P} = \text{NP}$). Instead, we prove a structural equivalence: under the two conjectures, BSD, and $\text{P} = \text{NP}$, $\#\text{P} \subseteq \text{FP}$ if and only if the specific family TunnellCount $:= \{n \mapsto D_n\}$ is in FP. The framework thus does not resolve the $\#\text{P} \stackrel{?}{=} \text{FP}$ question; it converts it into a concrete, falsifiable arithmetic question about the polynomial-time tractability of representation counts on one specific ternary quadratic form. We identify three concrete open problems---parsimony in Matiyasevich representations, the distribution of weight-$3/2$ Fourier coefficients via Waldspurger's formula, and the FP-tractability of $D_n$ itself---whose resolution would substantiate or refute the framework.

Abstract: Inflammasomes arise from complex protein assembly mechanisms and play a fundamental role in managing inflammation and the innate immune response. The molecules that trigger inflammasome assembly and activation are molecules derived from pathogens or DNA fragments released following cellular damage. The phenomena resulting from inflammasome activation range from the activation of caspases, such as caspase-1, to the secretion of pro-inflammatory cytokines, to cellular death by apoptosis or pyroptosis. Various pathologies have been linked to aberrant inflammasome activation, including several autoimmune diseases, leading scientists to direct experiments toward identifying the mechanisms responsible for aberrant inflammasome activation to develop new therapeutic strategies. In this review, we summarize the assembly mechanisms and involvement of two specific inflammasomes, NLRP3 and AIM2, in the autoimmune disease Sjögren's syndrome (SjD); NLRP3 and AIM2 aberrant activations appear to be involved in the exacerbation of inflammation, which becomes chronic, leading to dry mouth and dry eye and to an increased risk of developing B-cell non-Hodgkin's lymphoma in these patients. Understanding how different inflammasomes contribute to the pathogenesis of SjD could be fundamental to understanding the complex molecular mechanisms underlying this disease.

Abstract: Phishing detection models often report strong benchmark performance, yet their reliability under realistic deployment conditions remains uncertain. This study examines this problem by investigating three failure modes of cross-dataset phishing email detection: corpus generalisation failure, asymmetric prevalence-shift failure, and artifact-driven spurious learning. Using six public email corpora, CEAS_08, Enron, Ling, Nazario, Nigerian Fraud, and SpamAssassin, the study evaluates Term Frequency (TF) and Inverse Document Frequency (IDF)-based Logistic Regression and Linear Support Vector Classifier (SVC) models across pooled baseline testing, single-corpus cross-dataset transfer, leave-one-corpus-out pooled training, prevalence-shift simulation, training-prevalence manipulation, dataset-identification analysis, top-feature inspection, artifact-removal ablation, and targeted artifact masking. The findings show that single-corpus models are unstable under cross-dataset transfer, with F1-scores varying substantially across source–target combinations. In contrast, leave-one-corpus-out pooled training improves robustness, with Logistic Regression achieving sustained F1-scores between 0.8201 and 0.8994, and Linear SVC achieving F1-scores between 0.7607 and 0.8910 across unseen corpora. Prevalence-shift experiments reveal that failure is asymmetric and threshold-dependent. High-prevalence-trained models maintain high recall under fixed thresholds but suffer sharp recall degradation when operational alert-budget constraints are imposed. Conversely, low-prevalence-trained models become overly conservative in high-threat environments, producing high precision but substantially lower recall and poorer calibration. Artifact analyses further show that source corpus identity is highly learnable, with dataset-identification accuracy reaching 0.9722 for Logistic Regression and 0.9806 for Linear SVC. Top-feature and masking analyses indicate that models rely partly on corpus markers, date tokens, URL/domain terms, headers, and other artifact-like features rather than only general phishing indicators. The study contributes a deployment-aware and adversary-aware evaluation framework for phishing detection. It shows that benchmark accuracy alone is insufficient for assessing real-world robustness and that reliable phishing detection requires cross-corpus validation, prevalence-aware thresholding, and systematic testing for artifact-driven spurious learning.

Abstract: Earth blocks are attractive for low-cost housing because they use local soil, require less firing energy, and can provide good thermal mass, but their adoption in humid tropical regions is limited by moisture sensitivity. This review examines how agricultural waste ash stabilizers, with emphasis on palm kernel shell ash and related pozzolanic residues, influence moisture durability, dry/wet compressive strength behavior, and practical suitability of earth blocks for affordable housing. The paper synthesizes evidence from compressed earth block literature, pozzolanic material standards, and studies on ash-modified earthen masonry. It argues that wet-to-dry strength retention is a more realistic durability indicator than dry compressive strength alone because low-cost walls are exposed to wind-driven rain, capillary rise, damp surfaces, and imperfect maintenance. The review shows that ash stabilizers can improve particle bonding and pore refinement when properly processed, proportioned, compacted, and cured, but excessive ash, poor soil selection, or inadequate detailing can increase water absorption and reduce field reliability. The paper proposes a moisture-durability framework connecting material chemistry, block production, wall detailing, and tropical housing performance. It concludes that agricultural waste ash stabilized earth blocks are promising only when laboratory strength gains are tied to water-resistance testing and moisture-conscious architectural detailing.

Abstract: Long Range (LoRa) is a chirp spread spectrum (CSS) physical-layer technology that has become a leading candidate for low-power wide-area network (LPWAN) connectivity in the Internet of Things (IoT). At the receiver, the standard demodulator multiplies the incoming signal with a conjugate reference chirp and applies a one-dimensional discrete Fourier transform (DFT), reducing symbol detection to peak search in the frequency domain. While this non-coherent baseline is simple and robust under additive white Gaussian noise (AWGN), its symbol error rate (SER) degrades significantly in frequency-selective multipath channels, where parasitic spectral peaks distort the dominant tone. This paper presents a unified comparative study of seven LoRa detectors for spreading factor seven, six of which share a common one-dimensional DFT engine while a matched-filter bank operates directly in the time domain, with the six DFT detectors differing in their per-bin frequency-domain weighting and decision rule. The detector set spans the standard non-coherent DFT, a non-coherent matched filter bank, two coherent equalizers in the frequency domain (zero-forcing and minimum mean-square error), a phase-only equalizer, a maximal-ratio combiner with non-coherent decision, and an exhaustive maximum-likelihood detector that serves as a near‑optimal reference under the same preamble‑based CSI. To mitigate inter-symbol interference in the multipath case, every transmitted symbol is preceded by a cyclic prefix that converts the linear convolution with the channel into a circular convolution, enabling per-bin frequency-domain processing. Throughout the paper a deployment-realistic receiver model is adopted: the per-bin channel response is estimated by a frequency-domain least-squares estimator from a short preamble, and the noise variance is estimated blindly from the preamble residuals. The quality of the noise-variance estimator is reported separately as a diagnostic. Each detector is evaluated under both AWGN and a two-tap Rayleigh multipath channel through Monte Carlo simulation, and its execution time per call is recorded to provide a complementary view of computational cost. The framework introduced here clarifies how coherent processing, diversity combining, equalization, and exhaustive search trade detection performance against complexity within a single DFT-centric LoRa receiver architecture. The principal quantitative finding is that, under the two-tap Rayleigh multipath channel, the MMSE equalizer reaches SER ≈ 4.4×10⁻⁵ at SNR = −5 dB and tracks the exhaustive maximum-likelihood detector within 0.1 dB across the full SNR sweep, while costing only 1.26× the per-symbol time of the standard DFT receiver. Conversely, the standard non-coherent baseline hits an irreducible 16% error floor and the unregularized zero-forcing equalizer fails to reach the 10⁻² SER level at any SNR considered, isolating MMSE as the recommended choice in the multipath regime at every SNR for which a LoRa link is operationally viable.

Abstract: Electrochemical CO₂ reduction to ethanol is a promising route for circular-carbon fuel and chemical production, but practical implementation remains limited by coupled membrane, catalyst, transport and system-integration constraints. This Communication reassesses anion-exchange membranes (AEMs) and bipolar membranes (BPMs) using recent 2024–2026 literature. The central argument is that membrane selection is not a passive separation choice; it controls local pH, charge carriers, CO₂ availability, carbonate formation, water activity, proton/cation deliv-ery, product crossover and downstream techno-economic assessment (TEA) and life cycle assessment (LCA) burdens. AEM operation can create alkaline cathodic microenvironments that favor C–C coupling, but bicarbonate/carbonate formation imposes carbon-loss, salt-management and recovery penalties. BPM operation can improve pH separation and carbon management through water dissociation and bicarbonate acidification, but its viability depends on water-dissociation efficiency, co-ion exclusion, junction stability and voltage control. Recent ethanol-selective catalyst studies further show that copper oxidation state, grain boundaries, sub-surface dopants, ionomers, interfacial wettability and dynamic operation interact strongly with membrane-imposed microenvironments. The Communication pro-poses a membrane-centered decision framework linking AEM/BPM selection with ethanol selectivity, single-pass carbon utilization, energy efficiency, durability, TEA/LCA boundaries and future reactor design.

Abstract: Background: African Union (AU) guidance identifies decentralized diagnostics as central to epidemic preparedness. However, the epidemiological role of self-testing across epidemic-prone diseases remains underexplored. Drivers for potential impact of self-testing impact were examined conceptually using a transmission model. Methods: A deterministic SEIR model compared standard-of-care testing with additional self-testing. Global sensitivity analysis using Latin Hypercube sampling and partial rank correlation coefficients (PRCCs) examined parameters influencing reductions in peak disease prevalence (mitigation). Dynamics were illustrated using AU pathogen archetypes (Ebola, Influenza A, Cholera, Coronavirus, and Mpox), estimating the number needed to self-test (NNST) to avert one death as a measure of marginal efficiency. Results: Epidemic mitigation was minimal (median 1.9%; IQR: 0.4%–5.8%); correlated with isolation adherence (PRCC = 0.784), self-testing intensity (PRCC = 0.617), lower R0 (basic reproductive number; PRCC = -0.607) and greater duration of infectiousness (PRCC = 0.370). Achieving a 10% reduction in peak prevalence at R₀ = 1.1 required 34 self-tests per 10,000 people per day, exceeding AU COVID-19 operational benchmark of 10 per 10,000 per week. High-mortality, moderate-transmission archetypes (e.g., Ebola) were most responsive to mortality reductions (Median 1,512 NNST/death averted) compared to Mpox (Median 355,708 NNST/death averted). Adherence to post-test isolation exerted greater epidemiological impact than diagnostic accuracy. Conclusions: The epidemiological value of untargeted self-testing depends on pathogen characteristics and post-test behavioral adherence. Epidemic mitigation effects were generally limited under constrained health-system capacity. However, future demonstration studies evaluating rapid, early decentralized self-testing deployment during Ebola-archetype outbreaks may help identify operationally feasible targeted deployment strategies to support mitigation and mortality reduction.

Abstract: Recent work on neural scaling demonstrates consistent performance gains with increased data and model capacity, yet these improvements are typically assessed using surface-level metrics that do not capture factual reliability. In multi-document summarization (MDS), this limitation is particularly acute, as scaling has been shown to amplify hallucination and content distortion. In this paper, we investigate the empirical scaling behaviour of faithfulness-aware transformers under tightly controlled conditions, using LSHT as a fixed architectural and training baseline. Rather than proposing new scaling laws, we analyze how summarization quality, faithfulness and efficiency evolve as dataset size and model capacity are independently increased, while holding architecture, optimization, decoding and hardware constant. All experiments are conducted exclusively on the Multi-News benchmark to avoid cross-dataset confounds. Across ROUGE, coverage, repetition and faithfulness-oriented metrics, we show that lexical overlap and factual consistency follow distinct scaling dynamics. Faithfulness improves most rapidly during early data scaling (approximately 3–4% relative gain from 3k to 12k samples) but exhibits diminishing marginal returns at larger scales, whereas ROUGE continues to increase more smoothly. We further show that faithfulness is more sensitive to data diversity than to volume alone and identify practical scaling regimes that maximize faithfulness gains relative to computational cost. These results establish empirical expectations for scaling faithfulness-aware MDS systems and provide actionable guidance for reliable summarization under realistic resource constraints.

Abstract: Far-UVC radiation for disinfection in the presence of people in public indoor spaces through unshielded open radiation sources has been promoted for several years, claiming to be a simple solution to reduce infections from airborne pathogens such as bacteria and viruses. This literature review summarizes the existing research on the effectiveness of far-UVC radiation for inactivating pathogens, as well as potential risks to skin and eyes associated with exposure to far-UVC radiation. Further, it discusses radiation protection aspects of using far-UVC radiation in the presence of people, and addresses possible effects of far-UVC radiation on the human environment as well. The literature review shows that despite its antimicrobial and antiviral effectiveness, there is so far no sufficient evidence that far-UVC radiation can be used for disinfection in the presence of people in public indoor spaces without risks for humans and the environment. There are particular concerns about the safety of vulnerable groups such as children, the elderly and people with pre-existing medical conditions. The authors recommend further and extended studies in this field concerning potential risks of far-UVC radiation.

Abstract: Maintenance management of stationary combustion engines in the agricultural sector faces critical challenges owing to a reliance on manual methods, which increases the risk of unplanned downtime. This study developed a machine learning-based predictive model to anticipate failures within a 60-day horizon, facilitating the transition from a reactive to a proactive maintenance approach. Following the CRISP-DM methodology and drawing on a historical dataset of 2,250 records from 59 engines, feature engineering techniques were applied to derive 48 predictor variables, while K-Means clustering was employed to identify operational load profiles. The performance of two ensemble algorithms was then evaluated; LightGBM outperformed Random Forest under five-fold cross-validation with a 60/40 temporal split. The proposed model achieved an area under the ROC curve (AUC) of 0.91, a precision of 92.9%, and a recall of 76.1% for detecting actual failures. The findings indicate that gradient boosting techniques are highly effective for optimizing maintenance planning and reducing operating costs in the Ecuadorian agricultural context.

Abstract: Permutation flow shop scheduling is an important production planning problem handled in different contexts. Just-in-time measures have been significant in the optimization of real problems and one is specifically addressed here: the total earliness and tardiness of jobs. The most used approach in the literature to mathematically express this measure is to sum them up using unit weights thus obtainning a mono-objective function. In this paper it is shown that this is a simplification of a problem that is inherently multi-objective, highlighting how a more comprehensive approach can better support decision-making. A bi-objective mathematical optimization model and tools capable of analyzing the mono-objective solution within the multi-objective perspective are proposed. A computational study to analyze the benefits and difficulties of the solution using the bi-objective approach is presented. The results show that for large scale instances in which the tardiness factor is small, the conflict between the objectives of minimizing the total earliness and minimizing the total tardiness of jobs increases significantly. Therefore, the multi-objective approach has a greater potential to support decision-makers. Furthermore, we show that the choice of the solution method must be carefully considered, since the Pareto frontier associated with most instances has many non-supported points.

Abstract: With the aim of finding a generalization of the principle of general covariance applicable to certain models for a {\it fundamental dynamical}, a categorical approach to dynamics is developed. Given a category, the notion of pre-dynamics is introduced. The dynamics is then defined as a {\it finest pre-dynamics}. The {\it categorical recurrence principle}, which constitutes a categorical generalization of the principle of general covariance, is formulated. Natural consequences of the theory are the deterministic character of the fundamental dynamics and the evolving character of the physical dynamical law itself. One main problem considered is the way how time parameters can be defined in terms of the elements of the category and set theory, as a part of the implementation of the categorical recurrence principle. Aimed to address this question, a construction is presented showing how time parameters are defined from the elements of the theory recursively.

Abstract: Occupational exposure to hazardous chemicals remains a major concern across several industrial sectors, including, among others, manufacturing, agriculture, mining, and healthcare. Conventional exposure assessment methods, typically based on stationary environmental monitoring or periodic biological sampling, often fail to capture the dynamic and individualized nature of workplace exposures. In this context, wearable chemical sensors have emerged as promising tools for continuous and real-time exposure monitoring. This exploratory review provides an overview of recent advances in wearable chemical sensing technologies and discusses their potential applications in occupational exposure monitoring. Attention is given to electrochemical, optical, and hybrid sensing systems, as well as to the role of nanomaterials and flexible electronics in improving sensor sensitivity, selectivity, and wearability. Current applications in occupational settings, including monitoring of volatile organic compounds, toxic gases, and heavy metals, are discussed. The review also examines key challenges limiting the practical implementation of wearable sensors in occupational health programs, including long-term stability, environmental interference, power consumption, data integration, regulatory acceptance, and user compliance. Although wearable chemical sensors represent a rapidly evolving field with significant potential for improving occupational risk assessment and preventive strategies, further interdisciplinary research and long-term field validation studies are required before widespread implementation can be achieved.

Abstract: Managing safety and operational efficiency in large-scale events requires tools capable of capturing complex crowd dynamics while supporting rapid and informed decision-making. This paper presents a Generative AI-powered digital twin framework that integrates agent-based crowd simulation, an API-based execution pipeline, and a Large Language Model (LLM)-driven conversational interface within a unified system. The proposed approach enables dynamic configuration, execution, and analysis of crowd scenarios under varying operational conditions, including high-demand and emergency evacuation contexts. Experimental results demonstrate the system’s ability to reproduce nonlinear crowd dynamics, detect congestion patterns, and evaluate evacuation performance, providing actionable insights for planning and safety assessment. A key contribution lies in the introduction of an API-based execution paradigm that exposes the full simulation lifecycle (configuration, validation, execution, and output retrieval) through programmatic interfaces, enabling reproducible and scalable what-if analysis. Additionally, the integration of an LLM-based conversational interface allows non-technical users to interact with complex simulation models through natural language, significantly improving accessibility and usability. The framework is validated through a TRL-4 prototype, demonstrating robust performance, scalability, and interaction reliability. Overall, the proposed system advances digital twins from static analytical tools to executable, interactive, and user-centric platforms for decision support in complex urban environments.

Abstract: Nowadays, the organisational landscape aiming to provide value through their product and service offerings relies on having the infrastructure necessary to deliver at the expected service levels, as well as contributing to business continuity and organisational resilience in the face of modern organisational performance disruptions. This requires appropriate adaptation of existing frameworks, methods, and models to their business models which have generated consistent deliverables across time and industries. The same is applicable for the Romanian Information Technology (IT) organisations, which face increasing pressure to deliver within the expected quality, time, and budget parameters. Therefore, this paper aims to assess how stakeholder relationship management components, viewed through the Malcolm Baldrige National Quality Award (MBNQA) excellence framework, with impact on organisational quality and its contribution to business continuity and organisational resilience in Romanian IT organisations. This is a pilot-type study with a sample of N = 52 participants, to explore the applicability of the MBNQA framework within the Romanian IT sector. The results suggest that the four components of MBNQA focused on stakeholders (Leadership and Governance, Workforce, Customers and Markets, Community Engagement) may be suitable to be considered in assessment tools on the Romanian IT market. The “Workforce” variable emerges as the strongest area to focus on for achieving quality in stakeholder relationship management (SRM). Given its pilot delimitation, this study provides can be seen as providing an initial foundation for applying MBNQA in a specific regional IT context. While limited by sample size and geographic focus, the findings justify expanding the research to include broader population segments. Future research could transition from this correlational design to longitudinal frameworks to validate the associations across other multiple geographical markets.

Abstract: Previous studies on rats showed a deterioration of left ventricular (LV) function and myocardial injury characterized by oxidative/nitrosative stress, PARylation, and apoptosis in the heart after three days of hypoxia. In the present study on rats, we investigated whether a three-day recovery period in normoxia can reverse myocardial injury and dysfunction. Further, we studied the effects of norepinephrine (NE) administration as a model of strong sympathetic activation on hypoxia-induced LV dysfunction and myocardial damage, as well as their reversibility. Three days of normobaric hypoxia (10% O2) significantly decreased LV systolic function. Contrary to our expectations, NE infusion even aggravated the depression in LV function. These dysfunctions were completely reversed after three days of normoxic recovery. In contrast, nitrotyrosine as a marker of oxidative/nitrosative stress receded only partially, and poly-ADP-ribose (PAR) increased even further during the recovery period. Apoptosis-inducing factor receded at least partially indicating that PAR-related apoptosis (parthanatos) is not a major cause of hypoxia-induced LV dysfunction. Additional administration of NE mildly aggravated oxidative/nitrosative stress but did not significantly intensify PARylation and consequently, parthanatos. The findings demonstrate that hypoxia-induced LV dysfunction is reversible suggesting that subchronic hypoxia and subsequent reoxygenation has a better prognosis for the LV than classical ischemia/reperfusion injury.

Abstract: The pronounced aero-hydrodynamic coupling effects of modern Wind-Assisted Propulsion System (WAPS) ships challenge the applicability of traditional stability frameworks, which are predicated on hydrostatic energy balance, in satisfying the dynamic constraints of the Second Generation Intact Stability Criteria (SGISC). This paper systematically reviews the methodological evolution of dynamic stability assessments for WAPS ships under extreme and damaged conditions. By introducing a "Hierarchy of Evidence" evaluation framework, this study delineates the applicability boundaries of aerodynamic Reduced-Order Models (ROM), extended 3/4-DOF maneuvering equations, and 6-DOF time-domain hybrid architectures, defining the role of high-fidelity CFD-VPP in establishing calibration benchmarks. The review also discusses the damping distortion mechanisms induced by multiphase flow sloshing under damaged conditions. Synthesized findings indicate that transitioning towards a 6-DOF time-domain coupled architecture provides clear advantages for capturing unsteady aerodynamic hysteresis and nonlinear interference. Meanwhile, surrogate models, such as Physics-Informed Neural Networks (PINNs), offer a potential pathway to mitigate the computational demands associated with long-term extreme value extrapolations. Ultimately, this review provides a methodological reference for the high-fidelity assessment of WAPS and the development of Digital Twin systems.

Abstract: Aging bridge infrastructure and limited inspection resources have created an urgent need for automated and reliable bridge condition assessment systems. Most existing deep learning-based inspection approaches detect damage types from images without considering the structural member on which the damage occurs, limiting their practical utility for maintenance decision-making. This study proposes a structure-aware deep learning framework for automated bridge inspection that integrates structural member segmentation, multiclass damage detection, and spatial damage-to-member association within a unified pipeline. A SegFormer-based semantic segmentation model was trained on a custom bridge inspection dataset comprising 1,339 images to identify three primary structural member classes — main girder, deck slab, and abutment — achieving a test mean Intersection over Union (mIoU) of 0.851. Boundary refinement using the Segment Anything Model (SAM) in mask-prompt mode was applied to improve mask precision during training data preparation. A YOLOv8s object detection model was trained on a custom bridge damage dataset of 6,531 images to detect two damage classes — crack and corrosion — achieving a mean Average Precision (mAP50) of 0.445 at a confidence threshold of 0.30. The framework associates detected damage with segmented structural members using a region-based spatial assignment strategy, enabling structure-aware outputs such as “crack on main girder” and “corrosion on deck slab.” Manual evaluation on 100 bridge inspection images demonstrated a damage detection accuracy of 70.0% fully correct and 84.0% fully or partially correct, and a member assignment accuracy of 62.0% fully correct and 87.0% fully or partially correct. The main girder class achieved the highest combined accuracy for both damage detection (90.9%) and member assignment (93.9%). These results demonstrate the potential of the proposed framework for practical automated bridge inspection and infrastructure monitoring applications.

Abstract: 1) Background: Coumestrol is a bioactive compound that inhibits HASPIN activity and prevents tau and H3 phosphorylation. Oral ingestion of CBSs increases blood testosterone levels, which decline with age causing late-onset hypogonadism. Oral ingestion of coumestrol-rich bean sprouts (CBSs) has been shown to suppress the onset of Alzheimer’s disease in 5xFAD mice and the onset of colon cancer in APC^min/+ mice. 2) Methods: We investigated the effect of oral ingestion of CBSs on the progression of aging in male senescence-accelerated prone 1 (SAMP1) mice allowed voluntary exercise or no exercise. The SAMP1 mice were divided into two groups fed either a standard diet or a diet including bean sprouts from 12 to 18 weeks of age. Each group was divided into two groups with voluntary exercise or no exercise. 3) Results: Voluntary exercise accelerated aging-related declines in blood testosterone levels, nerve growth factor levels, and spatial working memory, and oral ingestion of CBSs suppressed these age-related phenotypes, regardless of exercise. 4) Conclusion: Ingestion of CBSs prevented aging-related phenotypes in the experimental mice. A detailed analysis of the molecular mechanisms of coumestrol will be useful for understanding aging and preventing age-related diseases such as cancer, Alzheimer’s disease, and LOH.