Abstract: The experimental determination of the relationships between the stress distribution zone in the soil layer and the parameters of tillage working bodies is a labor-intensive process. Therefore, preliminary mathematical modeling of this process is recommended to mi-nimize the total number of experiments. The research was conducted using the principles of classical mechanics and soil mechanics. Using an equation proposed by J. Boussinesq,a graphical-analytical method was developed to evaluate the stress state in the soil layer induced by a dihedral wedge. This method incorporates both the geometric parameters of the dihedral wedge and the physico-mechanical properties of the soil. A direct pro-portional relationshipwas established between the length of the dihedral wedge and the total area of the deformed soil mass. Specifically, increasing the length of the dihedral wedge by 83% (from 0.05 to 0.30 m) resulted in an 80% increase in the area of the de-formed soil mass (from 0.02 to 0.10 m²). The proposed graphical–analytical method can be employed in the design of tillage implements.

Abstract: Detecting distributional drift is central to reliable inference in evolving systems, yet existing approaches treat it as a static discrepancy measured by fixed divergence functionals. We introduce an information-geometric approach grounded in the Kerimov–Alekberli (KA) framework, where drift is modeled as a non-equilibrium trajectory on a curved statistical manifold. Three contributions are presented. First, we establish a formal impossibility result: no fixed divergence can achieve uniform optimality under non-stationary dynamics (Theorem 3.1). Second, we connect drift detection to entropy production, linking statistical inference with physical irreversibility (Theorem 4.1). Third, we introduce an asymmetric entropy operator A(θ) into the KA drift equation—a directional term that amplifies early entropy signals through a skew-symmetric perturbation of the gradient flow. Validated by 50 Monte Carlo runs on non-stationary Gaussian processes, adaptive divergence achieves mean detection at step 113 ± 11 vs 117 ± 13 for fixed KL (reference onset at step 100), demonstrating no-regret behaviour: the method is never significantly worse than the best fixed metric, while outperforming scalar baselines (ADWIN, Page-Hinkley) by wide margins (p < 10⁻⁴). These results motivate the principle of adaptive divergence, whereby the notion of distance between distributions must itself evolve in response to system dynamics.

Abstract: The increasing presence of microplastics (MPs) in freshwater ecosystems poses significant threats to aquatic biota; yet, species-level information on the presence of MPs in Indian riverine ecosystems is scarce. This study assessed 220 fish samples from twelve species and various trophic levels for MP ingestion, organ-level accumulation, polymer type, and ecological risk at four locations along the River Yamuna in India. MPs were detected in all the studied species and organs, indicating their widespread distribution across various ecological habitats and trophic levels. A total of 1,678 MPs were quantified, which were significantly higher in fish from urban Delhi compared to upstream regions. The gastrointestinal tract had the highest MP concentrations (751), followed by gills (605) and muscle tissues (322), thus confirming ingestion as the primary route of MP uptake and their subsequent translocation into internal organs. Fibers dominated the MP community (>78%), with transparent (44%) and blue (19.5%) particles being the most abundant. ATR-FTIR analysis revealed the presence of ten different polymers, with polyethylene (≈24%) and polypropylene (≈21%) contributing to approximately 45% of MPs. Significant organ-level correlations (r/ρ = 0.635-0.958) and spatial variability (Kruskal-Wallis, H = 11.03, p = 0.011) indicated coordinated MP accumulation influenced by urban pollution. The Polymer Hazard Index analysis revealed a high PHI value (Category IV), mainly contributed by the widespread distribution of highly toxic polymers such as polycarbonate and polyimide.

Abstract: Traditional European craft practices face dual pressures: the erosion of tacit knowledge held by aging practitioners, and the risk of cultural homogenization through uninformed digital adoption. This paper presents a comparative analysis of a structured design pilot conducted across five Representative Craft Instances (RCIs): glassblowing, tapestry, marble/silversmithing, porcelain, and woodcarving within the Horizon Europe CRAEFT project. Drawing on co-creative workshops, motion capture pipelines, physically based rendering (PBR), interactive simulation, and additive manufacturing, we analyze how context-specific digital tools performed as mediators rather than modernizers across heterogeneous craft domains. Cross-domain analysis reveals that digital tools achieve cultural legitimacy only when introduced through co-creative, practitioner-led cycles; that gesture and tacit knowledge are transferable via structured computational pipelines; and that methodological portability, not workflow replication, is the appropriate model for cross-context scaling. Implications are discussed for sustainable heritage policy, design education, and the development of craft-sensitive digital infrastructure in Europe.

Abstract: Background: Sickle cell disease (SCD) is the most common inherited blood disorder globally, affecting approximately 300,000 newborns annually. Cardiac remodeling, resulting from chronic anemia, vascular obstruction, and endothelial dysfunction, substantially contributes to morbidity and mortality in SCD. Apprehending these patterns is important for guiding clinical management and enhancing outcomes in pediatric patients. However, a comprehensive summary of cardiac remodeling patterns in pediatric and adolescent SCD, and their associations with genotype and clinical severity, is lacking. Methods: We conducted a systematic review following PRISMA 2020 guidelines, searching five databases for studies published from January 1978 to December 2024. Of 1,131 records, 37 studies met the inclusion criteria: 31 focused exclusively on children (Group A), and 6 included both children and adults (Group B). We analyzed cardiac remodeling patterns, genotype-specific findings, associations with disease severity, and imaging modalities. The 37 studies included approximately 4,847 patients from 18 countries, covering varied populations and imaging techniques. Left ventricular (LV) dilation was the most frequent finding (89.2%), followed by diastolic dysfunction (48.6%), pulmonary hypertension or elevated tricuspid regurgitant jet velocity (TRV; 37.8%), myocardial fibrosis (8.1%), and arrhythmia (2.7%). The HbSS genotype was associated with the most severe cardiac changes. Markers of disease severity, such as elevated lactate dehydrogenase (LDH), frequent acute chest syndrome (ACS), and increased hospitalizations, were strongly correlated with more pronounced cardiac remodeling. Variability in study design and imaging modalities underscores the requirement for standardized assessment protocols to enhance comparability and clinical translation. Cardiac remodeling in SCD starts early, even in infancy, and progresses with age. Recognizing this early can prompt healthcare professionals to prioritize rapid interventions and point out the importance of early monitoring and management.

Abstract: Early identification of corrosion-prone conditions remains a major maintenance challenge in closed, hard-to-access structural zones. This paper presents a multi-sensor data fusion approach for early warning of corrosion-prone conditions in selected closed zones of a medical rescue aircraft, as part of a structural health monitoring framework. The study combines sensor selection, installation in restricted-access compartments, and analysis of in-service data collected during helicopter operation. The workflow includes data acquisition, preprocessing, feature extraction, fused interpretation of multi-channel data, and assignment of warning levels linked to maintenance actions. Environmental, conductance, and electrochemical channels provide a first-stage early-warning layer that indicates persistent conditions favorable to long-term corrosion development, rather than direct proof of existing damage. Persistent warning states are intended to trigger staged follow-up diagnostics: PZT sensing localizes suspect subregions, while eddy-current sensing verifies and monitors the growth of local metallic degradation. Field inspection evidence of corrosion in hidden zones supports the practical relevance of this approach. Although demonstrated on an aircraft, the methodology is transferable to other closed or poorly accessible structural zones, including civil engineering applications.

Abstract: Mobile edge computing (MEC) enables computation-intensive and latency-sensitive tasks to be offloaded from mobile devices to nearby edge servers. Most existing MEC task offloading studies formulate offloading as a selection problem over a fixed or fully available set of candidate servers, which is restrictive in heterogeneous MEC scenarios with task-node eligibility constraints. Under such constraints, a task can be processed by an edge server only when task attributes, service requirements, link conditions, and node states jointly satisfy the corresponding eligibility conditions. The feasible action set therefore varies over time, while offloading decisions are further coupled with local queueing competition and long-term load evolution. To address this problem, this paper proposes RoSCo, a load-aware task offloading method with scheduling and constraint coordination for eligibility-constrained MEC systems. RoSCo constructs a dynamic feasible action set, applies eligibility-aware action masking to exclude infeasible offloading actions, introduces priority-driven local coordination to characterize service competition among heterogeneous tasks, and designs a load-responsive reward to guide congestion mitigation and load balancing. The offloading policy is learned using a dueling double deep Q-network (D3QN). Simulation results show that RoSCo reduces task drop rate and edge-node load imbalance while maintaining competitive task completion delay and energy consumption, especially under high-load and sparse-eligibility conditions.

Abstract: RNA-level regulation provides a therapeutically actionable bridge between genomic variation and disease mechanism. Although clinical genomics has traditionally prioritised variants that alter protein-coding sequence, many pathogenic variants act by disrupting RNA processing, stability, localisation, translation, editing, or surveillance. These mechanisms determine whether a transcript is correctly spliced, degraded, translated, toxic, or therapeutically recoverable. This review examines RNA-level mechanisms in human disease with emphasis on their implications for precision therapeutics. Splicing defects, pseudoexon inclusion, untranslated-region variants, premature termination codons, stop-loss variants, RNA-binding protein dysfunction, non-coding RNA dysregulation, altered codon usage, ribosome stalling, and surveillance pathways such as nonsense-mediated decay, nonstop decay, and no-go decay all create distinct therapeutic opportunities. Importantly, treatment selection must be mechanism-matched: splice defects may require antisense-mediated correction or small-molecule splice modulation; toxic transcripts may require antisense oligonucleotides (ASOs) or siRNA-mediated silencing; haploinsufficiency may require mRNA replacement, gene replacement, or transcript rescue; premature termination codons may be considered for readthrough only when the transcript and protein context are favourable; and regulatory RNA defects require network-aware approaches. RNA sequencing, long-read transcriptomics, allele-specific expression analysis, and functional assays are therefore essential not only for diagnosis but also for therapeutic stratification. An RNA-centric framework moves variant interpretation beyond descriptive classification toward mechanism-based intervention and precision RNA medicine.

Abstract: This article describes the development of a compact and affordable variable temperature NMRinstrument designed primarily to measure dynamic molecular motions in solids and liquids. The instrument consists of Lab-Tools’ Mk4 palm-top time-domain NMR spectrometer fitted with a Peltier-cooled variable temperature probe inside a shimmed Halbach magnet. Measurement of NMR relaxation times T₁, T₂, T_1ρ are possible over the temperature range −20^◦C to 70^◦C with cooling and heating rates, and data acquisition controlled from an integrated mini-PC. The overall footprint of the instrument is roughly that of a shoe box making both in-the-field and bench-top measurements possible. Applications of this instrument include measuring the pore size distribution in porous rocks, the viscosity of oils and tars trapped in porous rock, the properties of polymers, and the viscosity of the liquid components of foods (e.g. fruits, vegetables and seeds). Results of test measurements on calibrated oils and olive oil are presented together with measurements of the molecular mobility in a sold polymer.

Abstract: Professional development in discipline-specific pedagogy remains a persistent challenge in teacher education, particularly when addressed in fully online, asynchronous settings with in-service practitioners. This article describes and theorises the design of Specific Didactics of Biology and Geology, a 14-week professional development course offered at Universidade Aberta, Portugal’s public open distance university, within the framework of the institution’s Pedagogical Model. The course targets in-service Biology and Geology teachers enrolled in the first year of a teacher professionalisation programme. Rather than reporting a completed impact evaluation, the article offers a theoretically grounded, design-oriented practitioner inquiry account of course design, early implementation reflections, and transferable design principles for asynchronous online teacher professional development. Its conceptual core is the framing of teaching as didactic decision-making, integrating reflective practice, didactic transposition, and experiential learning theory into a coherent design philosophy that places participants’ existing professional experience at the centre of learning.The article argues that subject-specific professional development can be meaningfully designed in a fully asynchronous online format, provided that the e-activity repertoire is epistemically varied, experientially anchored, and coherent with the pedagogical principles under study. Three structural features are discussed in detail: a two-tier activity design distinguishing non-assessed preparatory tasks from assessed e-activities; a deliberately diverse typology of digital activity formats, each serving a distinct cognitive and social function; and an integrative culminating task, the Professional Development and Pedagogical Action Plan, developed progressively across the final three weeks of the course.By describing the course design in sufficient detail to enable adaptation to comparable contexts, the article contributes to research-informed approaches to online teacher professional development. It highlights the pedagogical potential of asynchronous learning environments when they are designed not as simplified substitutes for face-to-face provision, but as coherent spaces for reflective, situated, and discipline-specific professional learning.

Abstract: Relation extraction is an important task for structuring information from unstructured text. However, Romanian language still lacks dedicated datasets and benchmarks for this task. To address this gap, we introduce RoRED, a Romanian relation extraction dataset built by combining two complementary data construction strategies: translating existing high-quality English resources and applying distant supervision to native Romanian Wikipedia data. We leverage a powerful open-source large language model to automatically translate English examples into Romanian. For the native subset, we align Romanian Wikipedia entities with Wikidata relations to obtain naturally occurring Romanian examples. To better reflect real-world relation extraction scenarios, we also introduce synthetic negative examples generated using existing Romanian named entity recognition models. Finally, we validate the dataset by fine-tuning and evaluating multiple baseline models. Our strongest model, LUKE-RoRED, achieves a macro-F1 score of 0.8744 on the RoRED test set, demonstrating that the dataset can support relation extraction for Romanian. Overall, RoRED provides a strong first native benchmark for Romanian relation extraction.

Abstract: This project checks methods in wind power forecasting by comparing Gregorian calendar based on seasonal alignments with the vedic lunisolar calendar parallely. Rather than using timestamps like most forecasting methods, this project seeks to determine whether periodic cycles based on nature’s cosmos could reveal correlational patterns of wind activity surges and enhance accuracy. This study exploits the SOLETE dataset from SYSLAB, Denmark, which consists of 15 months of power generation alongside weather data. The dataset underwent processing with the CleanTS tool (an R package) and it was transformed into Gregorian and Vedic time frameworks. Within both time frameworks, the forecast approaches a hybrid forecasting model integrating “Variational Mode Decomposition (VMD) with Gaussian Process Regression (GPR)” was designed and assessed [11][12 ]. The Vedic forecasting approach is slightly better as it gives RMSE of 2.5519 and MAE of 2.0763, while the Gregorian forecasting approach gives RMSE of 2.6123 and MAE of 2.1424. The MAE correlation analysis over months revealed differing patterns within the two forecasting approaches with vedic giving better correlation than gregorian. This suggests that the Vedic calendar forecasting approach is better than the gregorian calendar system, which is based on natural cycles and is lunisolar, it is more accurate in capturing the chaotic signal of wind patterns than the arbitrary gregorian forecasting approach. This project helps in research, questioning the standard time representation in forecasting models which uses the gregorian timestamps and gives idea that if we put natural cycles through alternative calendar systems will it enhance the accuracy of energy predictions, potentially updating grid integration and operational planning.

Abstract: Infective endocarditis (IE) is a severe pathology with recent changes in its epidemiological profile, characterised by older patients with more comorbidities. The objective of this study is to describe the clinical and microbiological characteristics, as well as potential relations with mortality, of patients with IE in a tertiary academic centre. Material and Methods: Descriptive, retrospective, and observational study of patients over 18 years of age with a confirmed diagnosis of IE, conducted between 2021 and 2023 at the Dr Hernán Henríquez Aravena Hospital in Temuco, Chile. Biodemographic variables, risk factors, microbiology, echocardiographic findings, and complications were analysed using descriptive statistics and logistic regression models. Results: 119 patients were included (average age 60 years; 65.5% male; 28.5% rural). The most frequent risk factors were arterial hypertension (55%) and diabetes mellitus (29%). 18% were on haemodialysis (HD). Microbiological isolation was achieved in 78.1% of cases, with Streptococcus gallolyticus the most frequent isolate (16.8%), followed by Staphylococcus aureus (15.1%) and coagulase-negative Staphylococcus (15.1%). Complications were present in 69% of cases, mainly emboli (43%) and septic shock (23%)—59.6% required surgery. Global mortality was 44.5%, with a decreasing annual trend (from 58% in 2021 to 33% in 2023). Independent predictors of mortality were chronic renal failure on HD (OR 5.76; p = 0.001), heart failure (OR 3.13; p = 0.025), and septic shock (OR 3.31; p = 0.016). Conclusions: IE in this centre presents an aggressive profile and a high burden of comorbidities. The prevalence of S. gallolyticus stands out, possibly associated with high regional rurality. Mortality remains high, although it is improving.

Abstract: Quantum-like models of cognition account for order effects, conjunction and disjunction fallacies, and contextuality in human decision data using the Hilbert-space formalism without claiming literal quantum processes in the brain. Two decades of theoretical development have produced a mature mathematical apparatus, but its empirical foundation rests almost entirely on human-subject paradigms that are subject to linguistic priming confounds, demand characteristics, and replication concerns. This paper proposes that engineered brain-organoid preparations on multielectrode arrays—specifically the Cortical Labs CL1 and DishBrain-class systems—constitute the first substrate on which the structural commitments of quantum-like cognition can be tested without these confounds. I specify four operational signatures (sequential-stimulation order effects, Contextuality-by-Default cyclic-system inequalities, response replicability under non-invasive measurement, and interference effects in combined stimulation), and characterize, for each, the formal observable, the discriminating prediction against classical adaptive-learning baselines, and the substrate-level constraints imposed by current commercial wetware. The paper is offered as a theoretical specification, not an experimental protocol, and is calibrated for falsifiability rather than confirmation: a positive result on any signature would constrain classical models of organoid learning without confirming quantum-like dynamics; a fully negative result would narrow—though not conclusively delimit—the empirical scope of the quantum-like cognition program, with one natural reading being that these signatures depend more strongly on linguistic, pragmatic, or task-structured features of human-subject paradigms than on generic neural substrate dynamics.

Abstract: Deutsch's influential argument holds that the exponential speedup of quantum algorithms such as Shor's is best explained by computation distributed across ontologically real parallel branches of the wavefunction. This paper interrogates that claim by asking what minimal ontological commitments are actually required to underwrite observed quantum computational advantages. Drawing on the framework of final-state constraints and informational pruning developed in prior work, we argue that Deutsch's computational argument depends on an unpruned Everettian ontology in which all branches persist as computational substrates. We show that pruned-histories interpretations—in which boundary conditions or decoherence-based selection mechanisms restrict the space of ontologically realized branches—can preserve the empirical predictions of quantum computation while denying the parallel-universes inference. The argument requires three positive commitments: a records-based criterion for ontological commitment, a thermodynamically graded boundary between unitary computation and outcome-stabilization, and a positive account of computational speedup grounded in global Hilbert-space structure and entanglement rather than in a population of parallel worlds. We situate this result within the ontological models framework and recent observer-dependence theorems—including Frauchiger-Renner, Bong et al., and Walleghem et al.—and engage directly with Hewitt-Horsman's functionalist defense of computational branch realism. We conclude that the Deutsch argument, while rhetorically powerful, is interpretation-laden rather than interpretation-neutral.

Abstract: Surface modification of zinc oxide nanoparticles (ZnO NPs) with organosilane capping agents represents an effective strategy to control their physicochemical and biological properties. In this work, we report for the first time the use of halogenosilanes, namely (3- chloropropyl)trimethoxysilane (CPTMS), (3-bromopropyl)trimethoxysilane (BPTMS) and (3-iodopropyl)trimethoxysilane (IPTMS), for the surface functionalization of ZnO NPs obtained by chemical precipitation. Structural and morphological characterization (PXRD, TEM, SEM-EDX and FTIR) confirmed successful surface modification and revealed a significant particle size reduction from ~31 nm for unmodified ZnO to ~8 nm for BPTMS-modified ZnO (ZnO_b). The biological evaluation showed that halogenosilane-modified ZnO NPs exhibit enhanced cytotoxic activity against prostate cancer cell lines (PC3 and 22Rv1), with ZnO_b displaying the highest activity, likely associated with improved cellular uptake and increased reactive oxygen species (ROS) generation. In contrast, antimicrobial assays revealed only moderate bactericidal effects against Escherichia coli and Staphylococcus aureus at relatively high concentrations (≥1250 µg mL⁻¹), while no significant activity was observed against Pseudomonas aeruginosa, Burkholderia contaminans or Candida spp. within the tested range. These findings suggest that halogenosilane functionalization modulates the biological profile of ZnO nanoparticles by enhancing anticancer effects while also influencing microbiocidal activity, highlighting the role of surface chemistry in tuning biological selectivity. The present study supports the concept that rational surface engineering of ZnO-based nanoplatforms can be exploited to favor tumor-targeted activity over broad-spectrum antimicrobial effects, providing new perspectives for the design of application-oriented nanomaterials.

Abstract: Background/Objectives: Search for the new drugs capable of suppressing the development of drug resistance in tumor cells is extremely important for clinical practice. Cell signaling pathway inhibitors that control cell proliferation and death can be used in the complex therapy of malignant tumors. Methods: Cell cycle assay by flow cytometry, In Vitro Cell Viability Assay Cells chemosensitivity was analyzed by direct cell counting after trypan blue staining using microscope. Results: In the present work, we have shown that the combined action of doxorubicin and XMU-MP-1, the inhibitor of the MST1/2 kinase in the Hippo signaling pathway, prevents the development of drug resistance in Namalwa cells and significantly slows it down in K562 cells. and restores the sensitivity of resistant K562 cells to doxorubicin. We have shown that the combined action of doxorubicin and XMU-MP-1, causes a significant decrease in cell division rate and leads to the death of hematological tumor cells the Burkitt's lymphoma Namalwa, and myeloma K562 cells compared to monotherapy. Cell cycle analysis has demonstrated that the combined action of XMU-MP-1 and doxorubicin results in a catastrophic disruption of the cell cycle, and a significant increase in the number of cells undergoing apoptosis containing fragmented DNA. Conclusions: Thus, XMU-MP-1 can potentially be used in combination with anthracy-clines for the treatment of hematological malignancies and, in particular, the drug-resistant forms of cancer.

Abstract: This research presents a novel approach for enhancing retinal fundus images to detect anomalies better and diagnose retinal diseases. The work is divided into two stages: image representation and enhancement. Fundus images are represented in a Clifford color space, a 3D color model based on the RGB system, where colors are stored as multivectors that preserve color information and luminance. A rotation operation is applied to correct the image's illumination by adjusting brightness and color deviations, with the rotation angle and axis being critical for accurate enhancement. The gray-level axis serves as the rotational plane and the rotational angle of with a grayscale bivector axis, determined via discrete entropy (DE), optimally corrects image illumination. Following this, the green channel is extracted and enhanced using the CLAHE technique before being recombined with the other channels, and the image is reverse-rotated to its original color space. The effectiveness of the proposed method is evaluated using PSNR, DE, and SSIM on the MESSIDOR and DRIVE datasets, showing superior image quality and information preservation compared to existing methods. This enhanced technique is particularly beneficial for retinal landmark and lesion detection, improving diagnostic accuracy in retinal imaging.

Abstract: Background: Categorical TNM staging evaluates anatomical mass in a vacuum, creating artificial prognostic "cliffs" and ignoring cellular kinetics. The Ghartey et al Frugal Scalar Model introduced a kinetic paradigm to oncology—1 / ([Mass] x [Velocity]) —demonstrating that multiplying mass by proliferative velocity significantly improves prognostication. However, its initial iterations remained mathematically constrained by ordinal stage approximations. This study aims to evaluate the Ghartey Scalar against baseline TNM staging and computationally validate the integration of continuous spatial mass to mitigate these categorical limitations. Methods: Clinical data and exact continuous tumor dimensions (mm) were extracted from the METABRIC breast cancer cohort (N = 1,300). Model discrimination and probabilistic calibration were evaluated strictly against 5-year Disease-Specific Survival (DSS, 240 events). Traditional TNM staging was compared against the Ghartey et al Frugal Scalar Model, as well as continuous 1-Dimensional (winsorized diameter) and 3-Dimensional (log-volume) spatial derivatives. Model stability was verified via nonparametric bootstrapping (1,000 resamples). Results: The continuous 1-D Scalar definitively outperformed traditional categorical TNM staging, achieving an AUROC of 0.7364 and a Brier score of 0.1358. To evaluate clinical triage utility within the cohort (N=1,300; 240 events), a "Kinetic Maximum" threshold for early-stage disease was established using the largest, fastest-growing tumor within the Stage I boundary (RSD=1.0, Diameter=2.0 cm, Velocity=3.5; TmdRxResCoef (%) cut off = 14.28%). At 5 years, this threshold demonstrated a Negative Predictive Value (NPV) of 89.3% and a Sensitivity of 75.8% for predicting disease-specific death. Furthermore, longitudinal analysis out to 10 years revealed a significant stage-migration effect, as the Positive Predictive Value (PPV) of the Red Zone increased from 32.8% to 46.0%, confirming that the scalar successfully flags intrinsic biologic momentum prior to delayed clinical failure. Conclusion: Substituting ordinal tumor stage with exact continuous spatial dimensions entirely eliminates the stage-cliff. The Ghartey Frugal Scalar provides an exceptionally safe exclusionary triage threshold (NPV ~90%) while isolating high-velocity phenotypes that require immediate systemic velocity-braking.

Abstract: Federated learning (FL) is an attractive learning paradigm for privacy-preserving edge intelligence because it allows distributed devices to train a shared model without moving raw data to a central server. This feature is especially relevant to 5G and emerging 6G networks, where ultra-low latency, dense connectivity, and edge-native computing are expected to support large-scale intelligent services. Nevertheless, practical FL deployment remains difficult in heterogeneous wireless environments because client devices differ in processing capability, battery budget, data volume, and channel quality. These differences create stragglers, increase round latency, and waste scarce communication resources when client participation is scheduled naively. This study develops a deployment-oriented framework for dynamic client selection and resource allocation in heterogeneous edge environments. We formulate each FL round as a latency-constrained optimization problem that jointly captures computation time, uplink transmission time, and minimum participation requirements. On this basis, we propose a Dynamic Client Selection and Resource Allocation (DCS-RA) method that ranks clients using a weighted score combining computational capability, channel quality, and a fairness term, followed by a greedy radio-resource allocation procedure that prioritizes the largest marginal reduction in estimated completion time. Using the reported simulation setting with 100 clients and 20 resource blocks, DCS-RA reduces average round-completion time from 1.92 s to 1.55 s on MNIST and from 2.02 s to 1.57 s on CIFAR-10, corresponding to improvements of 19.39% and 22.47%, respectively. The results indicate that lightweight joint scheduling can substantially improve wall-clock efficiency for FL over heterogeneous 5G/6G edge networks.