Abstract: Hantaviruses are emerging zoonotic pathogens that represent a significant global threat due to their expanding geographic distribution, broad host range, and potential to cause severe disease in humans. These viruses are primarily transmitted via aerosolized excreta from infected rodents, although insectivores and bats have also been identified as potential reservoirs. Human infections can lead to two main clinical syndromes: hemorrhagic fever with renal syndrome (HFRS), which is predominantly reported in Europe and Asia, and hantavirus cardiopulmonary syndrome (HCPS), primarily occurring in the Americas. Several factors contribute to the rising incidence and spread of hantavirus infections worldwide, including climate change, environmental disturbances, urbanization, habitat alteration, and increased human-animal interactions. This mini-review synthesizes current understanding of hantavirus epidemiology, pathogenesis, diagnosis, treatment, and prevention, highlighting their growing importance within the One Health framework.

Abstract: Both temperature and solar radiation cause variations in photovoltaic output power. Nev-ertheless, each variation has a maximum power point, which represents the photovoltaic output's maximum efficiency. Photovoltaic power must be managed at the highest point in order to achieve optimal efficiency. This can be accomplished by employing a converter to regulate the photovoltaic output voltage at the maximum power. This study proposes a quadratic boost converter (QBC) to control photovoltaic output power by using the Deep Recurrent Neural Network (DRNN) algorithm. The goal of DRNN is to decrease ripple at the maximum point and speed up time to reach the maximum power point. QBC is de-signed to obtain a higher DC output voltage than a regular boost converter, so it can elim-inate the use of a step-up transformer if the photovoltaic is connected to an inverter. The proposed method is applied to a 50 Wp solar panel with an Arduino microcontroller as the controller device. The experimental results demonstrate that the DRNN algo-rithm-based QBC has effectively controlled the solar panel output power at the maximum point with a smoother ripple and a faster response. QBC has also been able to produce higher voltage output according to its characteristics.

Abstract: In the era of data-driven decision-making, the pursuit of competitive excellence in professional football has evolved beyond instinct and tradition. This research explores the question: What makes a football team successful? — by adopting a team-centric machine learning approach grounded in performance analytics. Using a comprehensive dataset of Premier League player statistics from 1992 to 2019, the study aims to develop predictive models that can identify the key performance indicators (KPIs) that drive team success over time. Chapter I establishes the research background, problem statement, and objectives, emphasizing the growing relevance of artificial intelligence in modern football analysis. Chapter II presents a critical review of existing literature on sports analytics and machine learning, highlighting methodological gaps in explainable, team-focused success modelling. Chapter III details a structured methodology based on the CRISP-DM framework, encompassing data preprocessing, feature engineering, performance tier formulation, feature selection strategies, and supervised learning model development. Three supervised classification models-Logistic Regression, Random Forest, and Gradient Boosting—were implemented and evaluated using metrics including Accuracy, F1-Score, ROC-AUC, and confusion matrices. Ensemble learning techniques, including voting and stacking, were further explored to enhance predictive robustness. Model stability was assessed through 5-fold stratified cross-validation, and paired t-tests on cross-validated F1-scores indicated no statistically significant performance differences between models (p > 0.05). Gradient Boosting demonstrated consistently strong performance (mean F1-score ≈ 1.00), low variance across folds, and superior interpretability, supporting its selection as the primary base learner within the final ensemble framework. To address model transparency, SHAP (SHapley Additive exPlanations) was applied at both team and player levels, enabling granular interpretation of feature contributions to success predictions. The findings reveal that attacking efficiency, defensive stability, and disciplinary control consistently influence successful team outcomes. Beyond predictive accuracy, the study proposes practical decision-support extensions, like performance tiering, highlighting the real-world applicability of the framework. This project ultimately aims not only to predict success but to uncover why certain teams win—offering insights that could inform coaching, scouting, and strategy. The outcome is a step forward in applying AI to assist the beautiful game to further evolve.

Abstract: Monopile offshore wind turbines are vulnerable to excessive vibration under coupled wind, wave, and seismic loading because of their slender and flexible structural characteristics. This study investigates a single-sided vibro-impact nonlinear energy sink (SSVI NES) installed inside the nacelle of a 5 MW monopile offshore wind turbine. A reduced-order ten-degree-of-freedom dynamic model is established using the Euler–Lagrange formulation, and turbulent wind, irregular wave, and seismic inputs are generated using TurbSim, the Kaimal and JONSWAP spectra, the Morison equation, and 15 PEER ground motions. The proposed controller is compared with an optimized tuned mass damper (TMD) under nominal and frequency-detuned conditions. The results show that the SSVI NES achieves vibration reduction comparable to that of the optimized TMD under the design condition while requiring a smaller absorber stroke. More importantly, it retains its control effectiveness more stably under frequency detuning, indicating improved robustness against structural-frequency variations. These findings suggest that the SSVI NES is a promising passive solution for enhancing the operational safety and multi-hazard resilience of monopile offshore wind turbines.

Abstract:

Corydalis ophiocarpa is a medicinally valuable plant, noted for its abundant alkaloid content. Despite its significance, the mitochondrial genome of this plant has not been characterized, which impedes both the phylogenetic understanding within the Corydalis genus and the comprehension of its full genetic potential. In this research, we have successfully assembled the complete mitogenome of C. ophiocarpa by employing a hybrid method that integrates Oxford Nanopore long reads with Illumina short reads. The assembled genome forms a circular structure of 600,064 bp, with a GC content of 46.49%, and includes 63 genes, comprising 40 unique protein-coding genes (PCGs), 20 tRNAs, and three rRNAs. Through assembly and coverage analysis, we identified a 6,383 bp forward repeat associated with a contig having approximately double the depth, indicating a repeat-mediated multipartite structure where the main circle may coexist with two smaller subgenomic forms. We discovered 775 C-to-U RNA editing sites across the 40 PCGs, with 95.4% being non-synonymous and favoring hydrophobic amino acid substitutions, particularly in Complex I subunits. Furthermore, we identified sixteen mt plastid DNA fragments constituting 2.43% of the mitogenome, a proportion more than double that found in the closely related C. saxicola. Phylogenetic analysis confirms that C. ophiocarpa is most closely related to C. saxicola, with C. pauciovulata as another close relative. This study presents the first complete mitogenome of C. ophiocarpa, providing a genomic basis for investigating the relationships between mt genome structure, post-transcriptional regulation, and energy-intensive specialized metabolism in the Corydalis genus.

Abstract: The emergence of the eukaryotic cell is regarded as a pivotal transition in the history of life on Earth. However, mounting evidence suggests eukaryogenesis was a specific, accidental event sparked by a syntrophic symbiosis between an Asgard archaeon and a bacterial endosymbiont. This prompts a fundamental counterfactual question: what if this symbiosis never occurred? The prevailing assumption is that life would remain perpetually microbial, constrained by the bioenergetic limits of prokaryotic cells. This article challenges that view by exploring the evolutionary potential of a unique group of bacteria: giant sulfur bacteria. These bacteria, driven by powerful selection pressure to bridge spatially separated pools of hydrogen sulfide and oxygen, have independently evolved remarkable sizes and different forms of complexity, including a form of eukaryote-like compartmentalization in Thiomargarita magnifica. Through the analysis of their novel bioenergetic solutions and conceptual modelling of an alternative evolutionary history, I propose that in an eukaryote-free world, giant sulfur bacteria represent a plausible starting point for the de novo evolution of complex, multicellular life. This thought experiment, albeit extremely speculative, offers new understanding of mechanisms of gaining complexity and could be useful for the analysis of the actual eukaryogenesis event, as well for the modelling of life complexity in astrobiological settings.

Abstract: This article considers several aspects of creating a combined bioreactor heating system without the use of external power grid power sources. It's known to intensify anaerobic digestion of organic waste, the bioreactor temperature is maintained within a specified range. A bioreactor heating system based on a "water jacket" , that heated by combustion of coal-water fuel has been developed. A technology for preparing and burning coal-water fuel using a radial circulation injection device is offered. Calculations are performed to determine the optimal temperature regime for the combustion process. The results obtained can contribute to the optimization of waste management technologies and ensure environmental sustainability by reducing carbon dioxide emissions and waste accumulation.

Abstract: The "spin-up" problem—where convection-permitting models require hours to develop realistic clouds from large-scale initial fields—critically limits short-term severe weather forecasting. Cloud analysis offers a potential solution by directly incorporating hydrome-teor information from remote sensing observations. In this study, we leverage multi-source remote sensing data, including three-dimensional mosaic radar reflectivity, hourly aver-aged FY-2G satellite black-body temperature (TBB), and FY-2G total cloud water products, within a stepwise cloud-analysis initialization scheme. The scheme is implemented in a convective-scale ensemble forecasting system (CMA-Meso, 3 km resolution) for a heavy rainfall event. For each ensemble member, three-dimensional hydrometeor increments are independently generated from these remote sensing retrievals and gradually introduced over the first ten time steps, ensuring smooth coordination with the model's dynam-ic-thermal framework. Results demonstrate that the remote sensing-driven cloud analysis substantially enhances ensemble system performance across multiple dimensions: (i) spin-up time is significant-ly reduced, with precipitation forecasts exhibiting reasonable structure from the initial forecast hour; (ii) deterministic forecast accuracy improves systematically, with reduced RMSE for geopotential height, temperature, and wind fields across all levels; (iii) proba-bilistic forecasting skill is enhanced, evidenced by improved CRPS and AROC for surface elements and precipitation thresholds; (iv) ensemble reliability is optimized, with spread better matching forecast errors. Mechanistic analysis reveals that these improvements stem from physically coordinated hydrometeor-latent heat initial perturbations and sub-sequent cloud-radiation feedbacks that continuously regulate thermal-dynamic structures. This study establishes that assimilating diverse remote sensing data via cloud analysis is an effective approach for addressing spin-up challenges in convective-scale ensemble prediction.

Abstract: Anxiety symptomatology and excess weight are associated with chronic low-grade inflammation. Olive Leaf Extract (OLE) contains polyphenols with antioxidant and anti-inflammatory properties that have shown anxiolytic like effects in experimental models; however, evidence in humans remains limited. This randomized double-blind placebo-controlled pilot trial evaluated the effects of OLE supplementation on anxiety symptomatology, inflammatory markers, and metabolic parameters in women with excess weight and mild to moderate anxiety symptoms. Participants received OLE (750 mg/day) or placebo for 12 weeks. Anxiety symptomatology was assessed using HAM-A, BAI, and STAI, while inflammatory and metabolic parameters were evaluated at baseline and post intervention. OLE supplementation was associated with a significant reduction in HAM-A scores, particularly psychic anxiety symptoms, together with lower TNF-α levels compared with placebo at the end of the intervention. No significant differences were observed in body composition, caloric intake, IL-6, hs-CRP, cortisol, or most metabolic parameters. Correlation analyses revealed positive associations between inflammatory markers, fat mass, and anxiety related measures. These findings provide preliminary evidence suggesting that OLE supplementation may exert beneficial effects on psychic anxiety symptomatology and inflammatory activity in women with excess weight. However, larger randomized clinical trials are necessary to confirm these observations and clarify the underlying mechanisms.

Abstract: The aim of the study was to examine the effect of respiratory muscle strength parameters obtained from endurance athletes on aerobic capacity levels. A total of 70 endurance athletes, 23 females and 47 males, voluntarily participated with the study. Respiratory muscle strength of the athletes were measured with a digital spirometer. Max VO₂ was assessed using the cardiopulmonary exercise testing system (Cosmed K5). As a result of the research; MIP and MEP values were determined to be related to PETCO₂ value at maximum load in female endurance athletes. In male endurance athletes, MEP values were determined to be related to PETCO₂ values at maximum load, PETO₂ values at maximum load, MaxVO₂ values, VO₂ values at RCP, and VO₂ values at VT. Additionally, in male endurance athletes, the MIP value was determined to be related to the VCO₂ value at RCP and the VTidal value at maximum load. Other Max VO₂ sub parameters examined were not associated with respiratory muscle strength. Research results reveal that there are relationships between maximal oxygen consumption which is the most important indicator of aerobic performance and its sub-parameters and respiratory muscles.

Abstract: We present a new cosmology model---the Eternal Universe Model (hereafter, EU-model)---that emerges from a subtle but consequential modification of the standard Friedmann--Lemaitre--Robertson--Walker (FLRW) framework. At first glance, the EU-model resembles the familiar ∧CDM concordance model; its departure, however, is philosophically and physically decisive: we relinquish the assumption of temporal homogeneity. Specifically, we allow the rate at which time progresses---encoded in the 00-component [g₀₀ = a²_t (r) c²₀ ] of the spacetime metric tensor---to vary systematically with radial position throughout the infinite expanse of the Universe. This single and seemingly banal alteration in the temporal architecture of spacetime gives rise to a remarkably new and rich cosmology. It introduces the continuous creation of matter and energy; it permits the variation of Fundamental Natural Constants (FNCs); it accommodates non-ponderable negative matter as a natural substrate for antimatter; and it endows the Universe with a fixed, absolute spatial centre from which all motion may be referenced. Furthermore, this framework offers natural explanations for the Hubble tension and the Cosmological Axis of Evil. The Universe that emerges is temporally and spatially infinite, globally unchanging, and truly eternal---with no beginning and no end.

Abstract: The dynamics of organic matter, nitrogen status, and biological activity in soils in southern Kazakhstan under various land-use systems were studied. A key feature of the research is the comprehensive comparison of humus status, nitrogen state, and biological activity of virgin and arable dark Kastanozem, Gleyic Calcisol, and Haplic Calcisol, as well as identification of their correlation with signs of functional depletion of organic component. The assessment was conducted using set of agrochemical and biological methods, including determination of humus content, available nitrogen forms, C/N ratio, microbial population, and enzymatic activity. It has been determined that the highest humus content is typical for dark chestnut soils under natural vegetation, while plowing of them is accompanied by decrease in humus content due to increased mineralization processes. Gleyic Calcisol - are characterized by more stable humus state, in some cases with increased organic matter content under arable conditions. Minimum humus values were found in Haplic Calcisol, due to arid conditions and limited supply of organic residues. It is shown that arable soils are characterized by a decreased C/N ratio and increased rates of organic matter transformation. Soil biological activity is linked to mineralization processes, as confirmed by microbial population dynamics and enzymatic activity. Additional assessment using digital tools reveals signs of functional depletion of organic component in agrocenoses. The obtained results indicate the need to consider biological indicators when assessing soil conditions and developing sustainable land management systems in arid climates.

Abstract: The Fermi Paradox (“Where is everybody?”) refers to the apparent contradiction between the visualisable abundance of extraterrestrial civilizations and the continued absence of confirmed detections. This work explores whether finite communicative lifetimes, combined with Galactic distance scales and the finite speed of light, can substantially suppress the probability of causal overlap between technological civilizations. Using a simplified stationary Galactic model (v = 0) within a Minkowski spacetime framework, technological civilizations are represented as finite world-line segments generating expanding “Information Shells” through electromagnetic signal propagation. Within this interpretation, successful detectability requires overlap between the communicative intervals of different civilizations in both space and time. For representative communicative lifetimes of order L ~ 10³ years, the effective causal reach of detectable signals remains small compared with typical interstellar separations expected in sparse-civilization scenarios. Using a heuristic overlap model, we estimate that for N = 100 contemporaneous civilizations distributed throughout the Milky Way, the effective causal-overlap probability remains below 1% . The analysis further considers long-term engineering limitations on autonomous probes and persistent signalling systems, including radiation damage, impact erosion, and power degradation, collectively described here as a “Hardware Filter.” In addition, the work distinguishes between the total biological lifetime of a civilization and its externally detectable communicative phase, suggesting that advanced civilizations may evolve toward increasingly low-leakage or radio-quiet technological states. Within this framework, the apparent “Great Silence” may emerge naturally from finite communicative windows, spacetime separation, and engineering constraints even if intelligent life itself is not intrinsically rare.

Abstract: Objective. Non-communicable diseases (NCDs) are major contributors to morbidity and mortality in Thailand, yet the effectiveness of lifestyle counselling within routine practice is underexplored. This rapid realist review examined how, for whom, and under what circumstances lifestyle counselling supports behaviour change among Thai adults. Design. Rapid realist review following guidance from the Realist and Meta-narrative Evidence Synthesis: Evolving Standards. Setting. Lifestyle counselling and health-coaching interventions for NCD prevention and management delivered in Thai primary care, community settings, or digitally supported programmes. Data sources. Six international and Thai databases (Scopus, Google Scholar, ProQuest, PubMed, EMBASE (Ovid), ThaiJo) were searched for studies published between 2005 and 2025. Eligibility criteria. Empirical studies involving adults (≥18 years) in Thailand that described lifestyle counselling or coaching interventions for NCD-related prevention or management and reported outcomes. Data extraction and synthesis. Data were extracted to identify contexts (C), mechanisms (M), outcomes (O), and equity considerations. These were synthesised into context–mechanism–outcome configurations (CMOCs) and helped to form programme theories. Two Thai doctoral students with community health experience provided public involvement feedback on cultural relevance and feasibility. Results. Thirteen studies were included. Nineteen explanatory configurations were identified across six mechanisms: self-efficacy, social support, motivation, accountability, emotional resilience, and relevance and engagement. Mechanisms were strengthened by family-centred education, routine self-monitoring with feedback, culturally or literacy-tailored materials, and brief stress-regulation strategies. Barriers included low health and digital literacy, conflicting norms, short programme duration, and rural workforce constraints. Facilitators included plain-language materials, low-tech or hybrid follow-up, co-designed dietary strategies, and task-sharing with village health volunteers and family members. Public contributors emphasised cultural alignment, feasibility, and equity. Conclusions. Lifestyle counselling in Thailand operates through six key mechanisms shaped by cultural norms, family dynamics, village health volunteers and service capacity. Effective programmes should prioritise long-term, low-intensity support; cultural and literacy tailoring; and hybrid low-tech maintenance. These findings provide theory-driven guidance for designing and implementing future lifestyle counselling interventions.

Abstract: Background: Social-cognition assessment often relies on endpoint measures such as accuracy, which provide limited information about how social stimuli are visually sampled. Eye-tracking can capture visual-sampling processes, but the meaning of gaze metrics depends on task structure. Objective: To examine the feasibility and preliminary informativeness of eye-tracking during two computerized social-cognition tasks in healthy adults. Methods: Nineteen healthy adults completed a full-face facial emotion-recognition task (TREC) and the Reading the Mind in the Eyes Test (RMET) while gaze was recorded. Measures included fixation count, cumulative fixation duration, and reaction time. TREC analyses examined gaze allocation across the eyes, nose, mouth, and facial hemifields. Analyses were exploratory and hypothesis-generating. Results: In the TREC, gaze was mainly allocated to the eyes and nose, with less sampling of the mouth. Higher TREC performance was accompanied by greater eye-region and left-hemiface viewing. Negative expressions elicited more fixations, and older participants showed greater eye-region sampling. In the RMET, participants showed higher fixation count, longer cumulative fixation duration, and longer response time than in the TREC, but gaze metrics were not clearly associated with demographic or performance variables. Conclusions: Eye-tracking was feasible and yielded coherent, task-dependent visual-sampling patterns in this small pilot sample. Full-face stimuli enabled spatially resolved gaze characterization, whereas eye-region stimuli mainly provided global inspection metrics. Findings are preliminary and should inform larger studies testing the clinical or mechanistic value of gaze-derived measures.

Abstract: We present ChestPathCT5-S100, an open dataset of 87 real-world chest CT examinations spanning five common thoracic pathologies: rib fracture, pleural effusion, lung mass, pulmonary embolism, and pneumothorax. The dataset was assembled from a retrospective single-centre cohort over a ten-year period, intentionally preserving acquisition heterogeneity and concomitant findings representative of routine clinical practice. Cases include both contrast-enhanced (arterial and venous phase) and non-contrast examinations, drawn from emergency, oncologic, and trauma settings. All imaging volumes are provided in NIfTI format. Technical validation by two radiologists confirmed correct pathology category assignment, image integrity, and unambiguous visibility of the dominant pathology for each case. A binary co-occurrence matrix of concomitant findings is provided to support multi-label research designs. ChestPathCT5-S100 is publicly available on Zenodo under a CC0 1.0 license, permitting unrestricted use and redistribution. The dataset supports classification, detection, weakly supervised learning, and multi-task learning paradigms.

Abstract: The volatile compounds(VOCs) evolution of wild ginseng (WG) across growth years is not a unidirectional process but a divergent remodeling of the chemical fingerprint. In this study, HS-GC-IMS combined with chemometrics was employed to characterize the dynamic changes of VOCs in WG at four growth stages(10, 15, 20, and 25 years; n≥15 per group). A total of 68 VOCs were tentatively identified and semi-quantified, encompassing terpenes, aldehydes, ketones, alcohols, esters, pyrazines, and other classes. Among them, terpenes and pyrazines exhibited the most pronounced directional trends, marking the divergent evolution: terpenes such as camphene, (E,E)-α-farnesene, and β-ionone accumulated progressively (increases of 242%, 74.6%, and 93.4% from 10 to 25 years, respectively), whereas pyrazines including 2,3,5-trimethylpyrazine and 2,5-dimethylpyrazine declined continuously (decreases of 58.2% and 53.3%, respectively). In contrast, the majority of compounds(66%) displayed non-monotonic patterns, including stage-specific metabolic peaks and environmentally driven fluctuations, underscoring the complexity of this divergent evolution. Partial least squares-discriminant analysis (PLS-DA) effectively distinguished samples across growth years (R²Y=0.997, Q²=0.993), with a 200-times permutation test confirming no overfitting(R²=0.136,Q² intercept=−0.505). Twenty-nine differential compounds with variable importance in projection (VIP)>1 were identified as potential chemical markers, and a multi-marker combinatorial system was tentatively established for discriminating three growth stages (10–15, 15–20, and 20–25 years). These findings provide chemical evidence that WG flavor quality evolves divergently over time, suggesting that VOCs fingerprint could serve as a supplementary tool for growth-year assessment, particularly for "high-quality but poor-shape" specimens that are undervalued by traditional morphology-based methods.

Abstract: UBE3A is a dosage-sensitive HECT E3 ubiquitin ligase whose neuronal expression is shaped by genomic imprinting at the 15q11.2–q13 locus. Opposite directions of UBE3A dosage imbalance contribute to distinct neurodevelopmental phenotypes: loss of maternal UBE3A underlies Angelman syndrome, whereas maternally derived duplications involving UBE3A contribute to Dup15q-associated syndromic autism phenotypes. This review synthesizes evidence across molecular architecture, isoform biology, neuronal imprinting, synaptic regulation, circuit excitability, and therapeutic development. The central argument is that UBE3A should not be interpreted as a general explanation for autism, but as a mechanistically informative model for a defined subset of neurodevelopmental disorders in which parent-of-origin effects and copy-number state are central. In Angelman syndrome, UBE3A loss disrupts proteostasis, synaptic plasticity, inhibitory circuit function, and neuronal excitability through distributed rather than single-pathway mechanisms. In maternal Dup15q syndrome, increased UBE3A dosage is strongly implicated in neuronal and synaptic abnormalities, although interval-wide dosage effects also contribute. Therapeutically, the direction of dosage change creates opposite translational requirements: restoration or paternal reactivation in Angelman syndrome versus dosage normalization in Dup15q-associated gain-of-function states. A dosage-directionality framework may therefore clarify how UBE3A biology connects molecular mechanism, developmental timing, and precision therapeutic design.

Abstract: Major depressive disorder (MDD) is clinically heterogeneous, and peripheral inflammatory biomarkers may help clarify early biological mechanisms before illness chronicity or pharmacologic treatment confound interpretation. This systematic review synthesized evidence on peripheral inflammatory biomarkers in first-episode, drug-naïve major depressive disorder (FEDN-MDD) compared with healthy controls and examined associations with clinical severity. Following PRISMA 2020, searches of PubMed/MEDLINE, Embase, PsycINFO, and Scopus from inception to March 19, 2026 identified 313 records; after screening, 16 publications were included in qualitative synthesis. Studies varied in age group, biological matrix, assay platform, and statistical reporting, precluding meta-analysis. The most frequently assessed biomarkers were IL-1β, TNF-α, IL-6, and CRP/hs-CRP. IL-6 showed the clearest recurrent tendency toward elevation in FEDN-MDD, whereas CRP/hs-CRP findings were partially positive but methodologically limited. TNF-α and IL-1β findings were mixed, and clinical correlations with depressive severity were sparse and inconsistent. Overall, the evidence supports heterogeneous early immune dysregulation in a subset of patients with FEDN-MDD rather than a single reproducible inflammatory signature. Peripheral inflammatory biomarkers should currently be considered research tools for biological stratification and mechanistic hypothesis generation, pending larger standardized longitudinal studies.

Abstract: To address the large model size, high computational cost, and limited deployment re-sources of keyword spotting models on edge platforms, this study proposes a collaborative multi-compression acceleration framework for lightweight deployment. Built on an end-to-end convolutional neural network for keyword spotting, the framework integrates adaptive structured pruning, hardware-friendly mixed-precision dynamic quantization, and quantization-aware multi-stage knowledge distillation into a unified compression pipeline. To eliminate the influence of inconsistent training budgets and data partitions across different compression branches, the results of quantization, pruning, distillation, and joint compression are reorganized under a unified evaluation protocol with mul-ti-seed mean ± std reporting. Under this protocol, the retrained baseline reaches 97.13% ± 0.85. Experimental results show that, in the quantization branch, MPDQ achieves 95.78% ± 1.69 with a compression ratio of 9.56×, demonstrating the most favorable balance be-tween accuracy and storage efficiency; in the pruning branch, AIASP reaches 95.63% ± 0.67 at 30% sparsity with a compression ratio of 1.43×, indicating a balanced compromise between accuracy retention and stability; in the distillation branch, PMKD, Multi-Teacher KD, and Fixed-T KD achieve 96.81% ± 0.69, 95.99% ± 1.18, and 96.70% ± 0.74, respectively, showing that the student model can maintain strong recognition performance under ap-proximately 4× structural compression; and the final joint compression scheme reaches 96.16% ± 0.53 with a trade-off score of 4.26 at a compression ratio of 9.89×. These results indicate that the main advantage of collaborative multi-compression lies in achieving a more balanced optimization among accuracy, model size, and compression efficiency un-der stringent deployment constraints.