Abstract:

This study aimed to develop a comprehensive typology of Sudanese sorghum-farming households within their food security status to inform targeted agricultural policy and rural development strategies. Using survey data from 392 households across 11 Sudanese states, the research captures the structural, socio-economic, and geographical diversity of farming systems and scrutinizes the relationship between socioeconomic characteristics of farmer households and related probability of constituting a specific farmer type. To assert this, Principal Component Analysis (PCA), hierarchical clustering, and Multinomial logistic regression analysis were applied. Through PCA and hierarchical clustering, three types of farmers were identified: The first type (Vulnerable Farmers), characterized by low education levels, small landholdings, high food insecurity, and reliance on subsistence farming; The second type (Well-off Remote farmers), operating larger landholdings meant for commercial purposes, yet facing challenges related to geographic isolation and limited market access; The third type (Educated Farmers with access to urban areas), consisting of households with higher education, diversified income sources, and proximity to markets, though still experiencing persistent food insecurity. Multinomial logistic regression analysis confirmed that household size, age, education, land size, market distance, and income structure are significant predictors of respective types of farmers. Thus, the study stands as a tool to enlighten intended/future policies, in providing input support and credit for vulnerable farmers, infrastructure and market access for remote commercial farmers, and land tenure security with innovative-geared incentives for farmers interacting with urban areas to foster inclusive, adaptive agricultural policies, and sustainable development across Sudan’s diverse farming communities.

Abstract: WRKY transcription factors (TFs) are pivotal in plant stress responses, yet their roles in Camellia oleifera, an economically important oil crop, remain poorly understood. We identified 192 WRKY genes in the tetraploid C. oleifera genome, and classified them into three groups (I, II and III) based on conserved domains. Chromosomal distribution revealed uneven localization of the WRKY genes, with the highest density (25 WRKY genes) on the Chromosome 10. RNA-seq analysis on anthracnose-resistant (CL150) and susceptible (CL102) cultivars inoculated with Colletotrichum gloeosporioides identified 1,822 differentially expressed genes (DEGs) and 109 DEGs dependent to CL150, including 11 core DEGs shared between the cultivars. Notably, one WRKY gene (YC.08G0001620-1A, Type I) exhibited significant upregulation in CL150, suggesting its role in disease resistance. Functional enrichment linked the DEGs to oxidative stress and metabolic pathways. This study provides a comprehensive WRKY family analysis in C. oleifera and highlights YC.08G0001620-1A as a promising candidate for molecular breeding to enhance anthracnose resistance in this economically important oil crop.

Abstract: Lung cancer is one of the most commonly diagnosed cancers worldwide and remains the leading cause of cancer-related death in both men and women. In 2022, approximately 2.5 million new cases of lung cancer and 1.8 million deaths due to the disease were estimated. Historically, lung cancer has been more frequent in men, although the difference between sexes has been decreasing, with tobacco use remaining the main etiological factor. Survival rates vary considerably depending on the stage at diagnosis and other factors, and overall prognosis remains poor, with a relatively low five-year survival rate compared to other types of cancer. In this work, the objective is to present current approaches to lung cancer diagnosis through the study of multiple genetic alterations and biomarkers, mainly detected by next-generation sequencing (NGS), which has significantly transformed cancer diagnostics by enabling highthroughput and cost-effective genomic analysis. In the context of lung cancer, NGS plays a crucial role in improving molecular characterization, guiding targeted therapies, and supporting personalized medicine strategies. Specifically, its relevance lies in the ability to provide a comprehensive genomic profile of the tumor, identify driver mutations, predict treatment response, detect co-occurring alterations, and assist in therapeutic stratification. A real-world case study was conducted including 101 patients diagnosed with lung cancer between 2023 and 2025 in a reference laboratory, whose tumors were analyzed using NGS. The most frequently altered genes identified were KRAS, EGFR, and ALK, together with other less common but clinically actionable alterations, as well as the evaluation of programmed death-ligand 1 (PD-L1) expression by immunohistochemistry. In summary, next-generation sequencing represents a fundamental tool in the diagnostic workflow of lung cancer, enabling comprehensive molecular profiling that supports personalized treatment selection and contributes to improved clinical management of patients.

Abstract: This study was inspired by Alcantara-Bode’s equivalent to the Riemann Hypothesis published in 1993, the equivalent formulation consisting in the injectivity of an integral operator connected to Riemann Zeta function. Surprisingly, the research on this line has not continued, an explanation would be the lack of criteria for the injectivity of integral operators. This paper aims to fill this gap by proposing a functional-numerical analysis solution exploiting the operator positivity properties on dense sets. The main theorem says that a linear, bounded operator strict positive definite on a dense set of a separable Hilbert space, has its null space containing only the null element, equivalently, it is injective. Having in mind to obtain a generic and useful criterion, we gradually changed the hypothesis of the strict positivity of the operator on a dense set to the involvement at the end, of the associated Hermitian operator that is semi positive on the whole space requesting additional properties related to the positivity of operator approximations on finite dimension subspaces. Then, in order to apply the criterion for Hermitian Hilbert-Schmidt operators, we choose an adequate dense set allowing to obtain operator sparse matrix representations. The criterion applied to the associated Hermitian of the Alcantara-Bode integral operator, showed that the equivalent holds, so the Riemann Hypothesis is true.

Abstract:

This study presents a multi-index performance system that is systematically used to assess the binder synergy and fly ash reactivity of eco-sustainable cementitious composite (ESCC) using the Strength Activity Index (SAI) as a reference in line with ASTM C618. The partial replacements of fly ash with high and low calcium fly ash (HCFA and LCFA) were added to the fly ash to sand (FA/S) ratios of 0, 10, 20, and 30% with a constant mix parameter, such as a 50% ratio of water to slag and a 20% ratio of activator to slag. Initial Flow Index (IFI) and Flow Retention Index (FRI) were used to measure fresh-state performance, and compressive-, tensile-, and flexural-based indices, i.e., SAI, Tensile Strength Index (TSI), and Flexural Strength Index (FSI), were used to measure mechanical performance. The results indicate that flowability and workability retention decrease with an increase in FA/S ratio, with LCFA-based mixtures having better flow retention than HCFA systems. The optimum mechanical performance at a replacement level of 20% FA/S produced the maximum SAI values of about 112% HCFA and 110% LCFA with a consistent increase in TSI and FSI values at 28 days. When the replacement levels were increased (30% FA/S), all strength indices decreased with the effect of dilution and decreased the packing efficiency of the binder. Comparisons of SAI with the respective TSI and FSI values through correlation analysis showed that the quantitative relationship between compressive, tensile, and flexural behavior was definite and showed that compressive strength alone is not enough to extrapolate mechanical performance. Collectively, the proposed framework provides a reasonable performance-based basis for the manner in which fly ash could be utilized in the most effective way in eco-sustainable cementitious compositions.

Abstract: Reliable estimation of kinetic parameters in molecular dynamics (MD) requires distinguishing physical phenomena from numerical artifacts. Standard MD workflows often mask integration errors through empirical damping, potentially obscuring rare configurational transitions. We introduce a calibration framework employing intentionally conservative numerical parameters—including reduced timesteps (0.10 fs) and attenuated intermolecular forces—to establish a numerical fidelity baseline. This approach isolates integration artifacts from force-field complexities, providing a reference against which production MD methods can be benchmarked. By demonstrating stable integration under maximally challenging conditions, we provide a methodology for validating the numerical foundations of kinetic inference in drug discovery applications.

Abstract: We develop an informational extension of spacetime thermodynamics in which local entropy production is coupled to spacetime curvature within an effective covariant framework. Spacetime is modeled as a continuum limit of finite-capacity information registers, giving rise to a coarse-grained entropy field whose gradients define an informational flux. Within a nonminimally coupled scalar–tensor formulation, the resulting field equations imply that the local divergence of this flux is sourced by the Ricci scalar, establishing a direct relation between curvature and entropy production. The corresponding integral form links cumulative entropy generation to the integrated spacetime curvature over a causal region. In stationary limits, the framework reproduces the Bekenstein–Hawking entropy of horizons, while in homogeneous expanding cosmologies it yields monotonic entropy growth consistent with the observed arrow of time. The construction remains compatible with unitarity at the microscopic level and with holographic entropy bounds in the stationary limit. Numerical solutions in flat FLRW backgrounds are used as consistency checks of the coupled evolution equations and confirm the expected curvature–entropy behavior across cosmological epochs. Overall, the results provide a thermodynamically consistent interpretation of curvature as a geometric source of irreversible information flow, without modifying the underlying gravitational field equations.

Abstract:

District heating systems are central to Europe’s decarbonisation efforts and its 2050 climate-neutrality target. However, given the deep embedding of district heating in the socio-economic system and built environment, meeting policy targets at the local level gives rise to a range of technical, infrastructural and socio-economic challenges. This is due to the high complexity and multidimensionality of the process, as well as the scarcity of local resources (e.g. land, surface waters, waste heat, etc.). In Bucharest, Romania, the largest district heating system in the European Union, the process of decarbonisation represents a particularly complex challenge. The system is characterised by high technical wear, heavy dependence on natural gas, significant heat losses and complex governance structures. This paper presents a strategic planning exercise for aligning the Bucharest system with the Energy Efficiency Directive 2023/1791. Drawing on system data, investment modelling and local resource mapping from the LIFE22-CET-SET_HEAT project, it evaluates scenarios for 2028 and 2035 that shift generation from natural gas to renewable, waste heat and high-efficiency sources. Options include large-scale heat pumps, waste-to-energy, geothermal and solar heat. Heat demand profiles and electricity price dynamics are used to evaluate economic feasibility and operational flexibility. The findings show that technical decarbonisation is possible, but financial viability hinges on phased investments, regulatory reforms and access to EU funding. The study concludes with recommendations for staged implementation, coordinated governance and socio-economic measures to safeguard affordability and reliability.

Abstract:

Forest ecosystems are vital to global carbon cycling as sinks or sources, while fast-growing, adaptable pines such as P. kesiya and P. oocarpa are central to national carbon sequestration efforts. This study was aimed at determining biomass accumulation variations and carbon stock dynamics between these two species at the age of 16 years in the Viphya Plantations, a prominent timber producing area in northern Malawi. Following the systematic sampling, forest inventory data was collected from 20 circular plots of 0.05 ha each. Above and below ground biomass was estimated using generic allometric models for pine species. Findings indicate that there were significant (P<0.001) differences in biomass accumulation and carbon sequestration between P. oocarpa and P. kesiya plantations. P. oocarpa accumulated more biomass (298.86±12.09 Mgha^-1) than P. kesiya (160.13±23.79 Mgha^-1). Furthermore, P. oocarpa plantation had a higher annual carbon sequestration (32.22±1.30 tCO₂e/ha/yr) as compared to P. kesiya plantation (17.26±2.56 tCO₂e/ha/yr). In addition, the uncertainty was less than 1% and fit within the IPCC’s recommended range (<15%). Therefore, the study has demonstrated that species selection should match management objectives: P. oocarpa maximizes short-to-medium term carbon sequestration and productivity, while P. kesiya supports long-term soil carbon stability. Hence, integrating both optimizes carbon benefits.

Abstract:

Background/Objectives: Rural and underserved adults face barriers to hypertension (HTN) self-management, and in-person lifestyle education programs in academic medical settings may have limited reach This pilot study evaluated a publicly available HTN self-management app (iOS/Android) with respect to feasibility, perceived usefulness, user satisfaction, and user-entered metrics relevant to HTN and lifestyle management. Methods: We conducted an internet-based, single-arm pilot of a mobile app available in commercial app stores. Adults aged ≥19 years who downloaded the free app and reported HTN self-enrolled via in-app registration and electronic consent; no direct recruitment or compensation was provided. Outcomes included an in-app questionnaire (HTN history, perceived BP status, concern, and a key self-management behavior) and app engagement/health-entry data (registration counts; use of tracking features; distributions of user-entered metrics). Results: From June 2020–July 2025, 819 users completed the in-app questionnaire; five were excluded as spam (N=814). Responses clustered in 2021 (76.8%), and completion time was brief (median 91 s; IQR 65–131). Most respondents reported hypertension for >2 years (57.3%; 21.5% unsure). Perceived BP was “normal” (42.1%), “borderline” (24.8%), or “high” (15.2%), with 15.0% unsure. For a key self-management behavior, only 21.8% reported measuring their blood pressure “usually/always,” while 24.8% reported never measuring their blood pressure. More than half were at least somewhat concerned about their BP (56.6%). Conclusions: In a largely rural, southeastern context, this publicly available HTN app demonstrated feasible low-touch uptake and captured user-entered self-management data, though sustained tracking occurred in a subset of users. Findings support further pragmatic testing focused on engagement, equity, and integration into nurse-led care workflows.

Abstract: We develop Real–Now–Front (RNF) cosmology, a generative framework in which spacetime arises dynamically as an advancing physical present aligns a pre-geometric chronon medium. Chronons are alignment degrees of freedom, not quanta of time; their coherent ordering induces Lorentzian geometry, causal structure, and operational rods and clocks. The dynamics are governed by the Temporal Coherence Principle (TCP), a local alignment and relaxation rule that reconstructs matter patterns and selects a preferred coherence density, so that spacetime symmetries emerge as stable operational properties rather than being postulated. Because each RNF advance encounters a metric-free layer, TCP enforces geometric rescaling to restore coherence, yielding kinematic cosmic expansion without vacuum energy and a local, self-tuning Hubble flow. Under-coherent regions expand, over-coherent regions shrink and collapse, and near-equilibrium regions evolve GR- and FRW-like, with vacuum-dominated regions generically producing late-time acceleration. Chronon microphysics further imposes a universal curvature bound through the Chronon Exclusivity Principle (CEP), leading to finite-density, nonsingular cores with R_core ∝ M^1/3. Small cores (Micro Chronon Condensates) provide a natural cold dark matter candidate, while larger cores reproduce general-relativistic black-hole exteriors with CEP-regulated interiors. RNF cosmology also predicts a mild two-metric structure, yielding small but testable distance–redshift deviations while qualitatively reproducing the large-scale phenomenology of ΛCDM.

Abstract: Underground mining environments are complex and hazardous operations where emergencies continue to happen. Post-incident investigations consistently identify training gaps in human related factors such as situational awareness and decision-making under stress. Conventional mine emergency training largely relies on instruction-based approaches which provide insufficient exposure to the cognitive and behavioral demands of real underground emergency situations. There has been an identified need to train miners for knowledge, skills, abilities, and other characteristics (KSAOs). This study proposes an adaptive immersive training framework (AITF) for miner self-escape readiness integrating immersive technology, situational awareness theory, KSAOs, and cognitive task analysis (CTA). The AITF aligns NIOSH-identified self-escape competencies with immersive training scenarios designed to assess and develop cognitive readiness and decision-making. CTA of historical mine accidents is introduced as a foundational design method for translating accident investigation findings into simulation scenarios and performance metrics. CTA of 2006 Darby Mine No. 1 explosion is presented as a proof of concept. The proposed framework supports individualized assessment, iterative scenario refinement, and data-driven feedback. The AITF advances miner training toward cognitive preparedness during mine emergencies and provides a foundation for future training systems that leverage digital tools, digital twins, and artificial intelligence for the mines of the future.

Abstract: Intimate relationships among contemporary emerging adults frequently manifest as situationships, characterized by emotional closeness in the absence of explicit commitment. Shaped by digital culture and evolving social norms, these relationships reflect heightened uncertainty and psychological tension within modern intimacy. The present study conceptualizes situationship as a multidimensional psychological construct, including commitment ambiguity, avoidance of emotional investment, and anxiety related to relationship uncertainty. Associations with attachment anxiety, trust, and subjective well-being are also investigated.To examine these dynamics, an integrated scale development and validation methodology was employed. The results indicated a stable three-factor structure. Structural equation modeling demonstrated that experiences of situationships were positively associated with attachment anxiety and psychological distress, and negatively associated with trust and well-being. Importantly, attachment anxiety partially mediated the relationship between relational ambiguity and relationship-related well-being.These findings establish relational ambiguity as a measurable psychological construct. The study contributes to positive psychology by enhancing understanding of relationship health and emotional regulation within contemporary intimate contexts. The results suggest that interventions promoting commitment clarity and emotional openness may enhance psychological well-being in emerging forms of intimate relationships.

Abstract: This paper presents a systematic comparison of neural architectures for English-to-Spanish machine translation. We implement and evaluate five model configurations ranging from vanilla LSTM encoder-decoders to Transformer models with pretrained embeddings. Using the OPUS-100 corpus (1M training pairs) and FLORES+ benchmark (2,009 test pairs), we evaluate translation quality using BLEU, chrF, and COMET metrics. Our best Transformer model achieves a BLEU score of 20.26, closing approximately 65% of the performance gap between our strongest LSTM baseline (BLEU 10.66) and the state-of-the-art Helsinki-NLP model (BLEU 26.60). We analyze the impact of architectural choices, data scale, and pretrained embeddings on translation quality, providing insights into the trade-offs between model complexity and performance.

Abstract:

This article examines how technological asymmetries—understood as differences in access to advanced digital tools, AI capabilities and IT infrastructure—shape the financial stability and market performance of enterprises of various sizes. The study integrates comparative analyses of 100 industrial joint-stock companies from multiple countries, including technologically advanced large corporations and innovative SMEs, to assess how disparities in digitization and AI implementation influence financial resilience. Using multivariate regression models and index-based financial metrics such as MC, EV, P/E, PEG, P/S, P/B, EV/R and EV/EBITDA, the research identifies relationships between technological advancement, operational efficiency and risk exposure. The findings indicate that companies with higher levels of digitization and AI adoption demonstrate stronger resistance to market disruptions, more effective risk management and more favorable capital structures than SMEs with limited technological resources. However, restricted access to detailed operational data for smaller firms may affect the precision of comparative assessments. The study concludes that investments in digital competences and international cooperation enhance financial stability and support strategic decision-making, while SMEs play an important complementary role by providing outsourcing services that facilitate AI implementation in larger corporations.

Abstract: Rural defensive heritage sites are highly vulnerable assets that require decision-making under conditions of limited data and high uncertainty, particularly in the context of large-scale infrastructure projects and accelerating environmental processes. This study proposes a modular decision-support model for defining conservation priorities in a transparent, traceable, and data-sensitive manner, based on four selected fortress sites in the Yusufeli district of Artvin, Türkiye. The model employs a risk-based approach to quantify anthropogenic risks (AR) through the combined assessment of impact (I) and probability (P). Topographic and contextual vulnerability (TC) is structured through sub-indicators including visual dominance disruption, access discontinuities, landscape fragmentation, and microclimatic exposure, while material and intervention compatibility (MS) is evaluated as a distinct compatibility–risk component. These three modules are integrated through normalization and weighted aggregation into a single Priority Index (PI). In addition, the study introduces a Data Completeness Index (DCI) to explicitly address heterogeneity and gaps in field data, allowing prioritization outcomes to be interpreted with an associated confidence level. Laser-scanning-based documentation, deterioration mapping, and photographic records support the evidence-based construction of indicators. The proposed framework offers a transferable approach for generating intervention and monitoring priorities for rural defensive heritage under rapid landscape transformation, while explicitly managing data uncertainty rather than obscuring it.

Abstract: Background: There is a growing interest in the effects of coffee consumption on the human health. This study was conducted to identify a causal relationship between the coffee consumption and the risk of metabolic syndrome (MetS). Methods: We analyzed the data of the 5th Korea National Health and Nutrition Examination Survey in 2010 for the current study. Results: The risk of MetS, high triglyceride (TG) and low high-density lipid-cholesterol (HDL-C), was significantly lower in the female subjects with a daily amount of coffee consumption of ≥ 3 cups as compared with those with a daily amount of coffee consumption of < 1 cup. There was a significant dose-response inverse correlation between the amount of coffee consumption and the risk factors of MetS (high TG and low HDL-C) after the adjustment of multiple confounding factors (P=0.015 and 0.011, respectively). There was also a modest dose-response relationship between the amount of coffee consumption and MetS (P=0.056). There was no significant correlation between the amount of coffee consumption and MetS in the male subjects. Conclusions: The coffee consumption might have a beneficial effect in lowering the risk of MetS. The current results suggest that it would be mandatory to consider individuals’ recognition of health impacts of coffee consumption.

Abstract: We study the local infrared content of four-dimensional Palatini gravity in the projective equivalence class, with the observable sector defined by a scalar PT projection. Restricting to a strictly local, curvature-linear two-derivative truncation, we (i) give an explicit and complete basis for all PT-even, projectively admissible bulk scalars in the trace/scalar channel and (ii) define the admissible equivalence relations that preserve the posture, including IR closure of the matter sector and tensorsector locking diagnostics used as auditable admissibility tests. A key structural consequence of full one-form projective invariance, \( \Gamma^\lambda{}_{\mu\nu}\to \Gamma^\lambda{}_{\mu\nu}+\delta^\lambda_{\mu}\xi_\nu \), is the appearance of a projectively invariant residue one-form \( \mathcal{T}_\mu \equiv T_\mu-A_\mu \), where \( A_\mu \) is a compensator transforming as \( A_\mu\to A_\mu+3\xi_\mu \). We then prove a conditional local no-go: within the closed two-derivative operator class and modulo admissible equivalences, there exists no reformulation that removes \( \mathcal{T}_\mu \) from the bulk dynamics in the observable trace/scalar channel while simultaneously (a) preserving IR closure of the minimal matter-coupling posture and (b) preserving the tensor-sector locking diagnostics (in particular luminality). Any attempted bulk removal is necessarily exhausted by a small set of controlled failure modes, including collapse to a trivial residue-free branch, departure from the admissible operator class / IR non-closure, or locking failure. On admissible domains one may restrict to longitudinal representatives, where a scalar \( \epsilon \) parameterizes the physical longitudinal content of \( \mathcal{T}_\mu \); \( \epsilon \) is not a compensator and cannot be eliminated as a pure gauge artifact. We summarize the exclusion logic in a compact diagnostic table and provide a minimal counterexample on standard backgrounds, thereby making the local IR residual (“IR island”) operationally auditable.

Abstract: Wedevelop a fully gauge-invariant analysis of gravitational-wave polarizations in metric f(R) gravity with a particular focus on the modified Starobinsky model f(R) = R +αR² −2Λ, whose constant curvature solution R_d = 4Λ provides a natural de Sitter background for both early- and late-time cosmology. Linearizing the field equations around this background, we derive the Klein–Gordon equation for the curvature perturbation δR and show that the scalar propagating mode acquires a mass m_ψ² = 1/(6α), highlighting how the same scalar degree of freedom governs inflationary dynamics at high curvature and the ropagation of gravitational waves in the current accelerating Universe. Using the scalar–vector–tensor decomposition and a decomposition of the perturbed Ricci tensor, we obtain a set of fully gauge-invariant propagation equations that isolate the contributions of the scalar, vector, and tensor modes in the presence of matter. We find that the tensor sector retains the two transverse–traceless polarizations of General Relativity, while the scalar sector supports a massive breathing/longitudinal mode determined by the massive scalar propagating mode. Through the geodesic deviation equation—computed both in a local Minkowski patch and in fully covariant de Sitter form—we independently recover the same polarization content and identify its tidal signatures. The resulting framework connects the extra scalar polarization to cosmological observables, providing a unified, gauge-invariant link between gravitational-wave phenomenology and the cosmological implications of metric f(R) gravity.

Abstract: The computational analysis of therapeutic communication presents fundamental challenges in multi-label classification, severe class imbalance, and heterogeneous multimodal data integration. We introduce a comprehensive bidirectional framework that addresses patient emotion recognition and provider behavior analysis through advanced data mining techniques. For patient-side emotion recognition, we employ ClinicalBERT fine-tuned on human-annotated CounselChat comprising 1,482 counseling interactions across 25 emotion categories exhibiting class imbalance ratios reaching 60:1. Through frequency-stratified class weighting combined with dynamic per-class threshold optimization, we achieve macro-F1 of 0.74, representing a six-fold improvement over baseline multi-label approaches. Recognizing that patient emotion detection alone provides insufficient analytic utility, we extend our framework to provider-side behavior recognition using real-world psychotherapy sessions. We process 330 YouTube therapy sessions through an automated pipeline incorporating speaker diarization, automatic speech recognition, and temporal segmentation, yielding 14,086 annotated 10-second communication segments. Our provider-side architecture combines DeBERTa-v3-base for contextual text encoding with WavLM-base-plus for self-supervised audio representation learning, integrated through cross-modal attention mechanisms that learn content-dependent prosodic associations. On controlled human-annotated HOPE data comprising 178 sessions with approximately 12,500 utterances, the provider model achieves macro-F1 of 0.91 with Cohen's kappa of 0.87, comparable to inter-rater reliability reported among trained human annotators in psychotherapy process research, outperforming simple concatenation-based fusion by 12 percentage points. On automatically annotated YouTube data, the model achieves macro-F1 of 0.71, demonstrating feasibility of analyzing naturalistic clinical communication at scale while highlighting the performance gap between controlled and real-world scenarios.