Search Results (11,882)

Edible lily bulbs (Lilium lancifolium Thunb.) are highly valued for their rich nutritional profile and significant medicinal properties. To further enhance these beneficial attributes, this study investigates a novel, green post-harvest enhancement strategy using low-dose ⁶⁰Co-γ irradiation (2–5 Gy) followed by immediate freeze-drying. Experimentally, fresh bulbs were exposed to varying irradiation doses, after which physiological parameters (enzyme activities, phytochemical contents) and metabolic profiles (via LC-MS/MS) were systematically evaluated. The results demonstrated that 4 Gy represented the optimal dose; compared to the non-irradiated control group (0 Gy), the 4 Gy treatment significantly enhanced the antioxidant defense system by eliciting oxidative stress responses, increasing superoxide dismutase (SOD) and catalase (CAT) activities by 1.55- to 1.86-fold, and elevating total phenolics and flavonoids by 1.31- and 2.69-fold, respectively. Untargeted metabolomic analysis identified 42 differential metabolites, revealing that 4 Gy irradiation primarily upregulated the phenylpropanoid and flavonoid biosynthetic pathways. Key antioxidants, specifically rosmarinic acid and the newly detected alkaloid jurubine, exhibited significant accumulation. Integrated correlation analysis and molecular docking simulations elucidated the underlying mechanisms, providing theoretical insights into their transient interactions with DPPH radicals. In conclusion, low-dose ⁶⁰Co-γ irradiation serves as a sustainable, green post-harvest processing strategy to improve the nutritional quality of edible lily bulbs by activating intrinsic metabolic defense mechanisms. Full article

Collagen functional fragments retain selected activities of parent collagens while allowing greater flexibility in sequence design and precision fermentation. Although recent reviews have covered recombinant collagen production technologies, expression platforms, purification strategies, quality control, and biomedical applications, fragment selection, host–process matching, production, and characterization have received less integrated attention. This review focuses primarily on collagen-derived functional fragments, while collagen-mimetic peptides and collagen-like proteins are discussed as related design systems. The biological basis for fragmentation includes receptor-recognition motifs, matrikines and matricryptins, and basement membrane-derived fragments. The review further examines how motif context, Gly-X-Y organization, stabilizing sequence features, protease susceptibility, post-translational modification requirements, and host compatibility influence fragment stability, expression performance, production feasibility, and product integrity. Microbial production using Escherichia coli, Komagataella phaffii, and Saccharomyces cerevisiae is discussed from the perspectives of construct–host matching, secretory or intracellular production, prolyl 4-hydroxylase configuration, fermentation optimization and scale-up, and product characterization. Finally, we discuss AI-assisted, quality-guided design-build-test-learn workflows that integrate computational prediction, curated structural, extracellular-matrix, interaction, and protease resources, two-tier candidate evaluation, and format-appropriate experimental testing to support iterative sequence, host, and process optimization. The development of collagen functional fragments therefore depends on coordinated optimization of biological function, molecular design, microbial host performance, fermentation processes, and product characterization. Full article

In modern process industries, data-driven soft sensors have become indispensable for monitoring critical process variables that are inaccessible to direct measurement. Nevertheless, the accurate modeling of industrial processes remains challenging due to their intrinsic complexities, such as time-series behaviors, measurement outliers, redundant variables, and potential concept drift. Existing approaches can address subsets of these challenges but generally lack a unified mechanism that integrates robust offline modeling, variable-importance analysis, and efficient online adaptation. To address these issues, this study proposes an online-updating robust soft sensor framework based on bidirectional long short-term memory (BiLSTM) with integrated gradients (IG) and smoothed quantile loss (SQLoss). During offline modeling, a soft-sensing model is constructed using a BiLSTM, and the proposed SQLoss is introduced to reduce the influence of outliers; the IG method is then employed to evaluate the importance of input variables, enabling input variable selection. During online operation, model parameters associated with significant variables are selectively updated based on IG-derived variable importance, thereby addressing concept drift. Finally, experimental results on an industrial desulfurization process demonstrate that, compared with the best-performing competing basic learner, the proposed SQLoss-BiLSTM-IG reduces the average root mean squared error (RMSE) and mean absolute percentage error by 5.26% and 1.87%, respectively, while increasing the average correlation coefficient by 1.94%; in the online evaluation, the proposed updating strategy achieves a mean RMSE of 2.441, demonstrating its effectiveness in handling concept drift. Moreover, the analysis of key variable importance is consistent with field experience, offering valuable insights for optimizing the desulfurization control system. Full article

Nonlinear Continuous global Optimization Problems (NCOPs) are well-known problems that arise in many applications, from engineering to robotics. The Branch & Bound method is a widely used approach for solving NCOPs to global optimality, often interleaving techniques like bisection and filtering. A key aspect of this approach is identifying feasible solutions early in the search process, which enables effective pruning of the search tree and avoids unnecessary computational effort. Inner linear relaxation techniques, such as the AbsTaylor strategy, have proven effective for identifying feasible regions; however, they heavily rely on a heuristically chosen expansion point (often the box midpoint), which directly impacts solution quality and relaxation success. In this work, we propose a novel gradient-based strategy to dynamically optimize the selection of this expansion point. By employing Gradient Descent to minimize a Mean Squared Error (MSE) objective formulated exclusively over the active constraints, we systematically guide the expansion point safely away from boundaries and into a strictly feasible interior region. To manage computational overhead, we evaluate restricted iteration budgets alongside algorithmic variants, specifically introducing a point inheritance strategy for warm-starting and comparing Batch versus Incremental gradient updates. Experimental results on a well-known benchmark set demonstrate that this gradient-based approach minimizes the probability of relaxation failure, significantly enhancing pruning effectiveness and overall solver efficiency compared to the original strategy. Full article

Flexible electronic textiles hold potential applications across various fields, yet current functionalization methods frequently suffer from poor coating uniformity and severe aesthetic alteration. This study addresses these challenges by establishing a multi-analytical approach, based on Raman spectroscopy together with DLS, Zeta potential and XPS measurements, to optimize graphene oxide (GO) precursor selection prior to electrostatic deposition onto cotton fabrics using a polyethyleneimine linker. This coating strategy was inspired by layer-by-layer deposition technique and centrifugation was used to partition a heterogeneous commercial GO precursor into three distinct homogeneous fractions. Raman spectroscopy revealed that centrifugation acts not merely separating GO based on its size but effectively sorts GO sheets based on their chemical functionalization degree. Consequently, this approach allows for the identification of the optimal precursor fraction, balancing sheet dimensions with defect density, to ensure strong functionalization. Overall, this work established a foundational spectroscopic methodology for precursor selection, deposition monitoring and process optimization, which can also be extended to the characterizing and comparison of different commercial GO batches from different industrial suppliers. Finally, micro-Raman mapping and XPS were preliminary applied to verify the textile fiber functionalization. Full article

As a noninvasive liquid biopsy approach, extracellular vesicle (EV)-based detection offers significant advantages in reflecting real-time tumor dynamis and overcoming the limitations of conventional tissue biopsy. EVs, nanoscale vesicles secreted by cells, carry diverse biomolecules such as proteins and nucleic acids, playing key roles in tumor progression, metastasis, and immune evasion, and have emerged as promising biomarkers for cancer liquid biopsy. Surface-enhanced Raman spectroscopy (SERS), characterized by high sensitivity, resistance to photobleaching, minimal sample consumption, and multiplexing capability, has shown great potential in EV analysis. This review systematically summarizes current methods for EV isolation, characterization, and storage, with a focus on label-free and label-based SERS detection strategies for early cancer diagnosis, treatment response monitoring, and prognosis evaluation. Furthermore, the integration of SERS with machine learning and deep learning algorithms has substantially improved diagnostic accuracy and cancer subtyping. Despite remaining challenges, such as optimization of SERS substrate performance, intelligent processing of Raman spectral fingerprints, and clinical translation, EV-based SERS technology holds great promise for precision oncology. Full article

Sodium chloride (NaCl) serves multiple functions in meat product processing, including flavor enhancement, preservation, and texture regulation. However, excessive sodium intake significantly increases the risks of chronic diseases, including cardiovascular disease, hypertension, and gastric cancer. Globally, sodium intake among the general population consistently exceeds the daily upper limit recommended by the World Health Organization (less than 2000 mg sodium per day, equivalent to <5 g salt per day). In certain regions, processed meat products are an important source of dietary sodium intake. Consequently, the development of low-sodium meat products has emerged as a critical priority in both the food industry and public health. This review reviews and summarizes the multifunctional roles of sodium chloride in meat products and the underlying mechanisms of these functions, and evaluates mainstream sodium reduction strategies, namely direct sodium reduction, physical modification, salt substitutes, flavor enhancement, odor-induced saltiness enhancement (OISE), and non-thermal processing. The analysis indicates that individual strategies exhibit limitations in terms of sensory quality, safety, or cost. Future efforts should focus on achieving effective sodium reduction through the synergistic application of multiple strategies, without compromising product quality or safety. This review further proposes a product-type-oriented strategy matrix, and multi-strategy synergy combined with AI optimization which is put forward as a promising potential pathway for the industrialization of sodium reduction in meat products, which remains to be validated by further research. Full article

Cross-view geo-localization (CVGL) between unmanned aerial vehicle (UAV) imagery and satellite imagery is a key technique for autonomous UAV navigation in Global Navigation Satellite System (GNSS)-denied environments. However, most existing methods rely on energy-intensive Artificial Neural Networks (ANNs), making them difficult to deploy on resource-constrained edge computing platforms. Spiking Neural Networks (SNNs) provide a promising alternative for energy-efficient inference, but their application to CVGL still faces two challenges that remain insufficiently addressed. First, the isotropic computation used by existing SNN backbones is mismatched with the directional characteristics of spike activations. Spike activations tend to form oriented aggregation patterns along elongated geographic structures, and isotropic computation can therefore dilute directional signals. Second, the limited representational capacity of SNNs further increases the sensitivity during training optimization. However, the standard triplet loss adopts a static weighting strategy and assigns the same weight to all triplets that violate the margin constraint, which is unfavorable for learning from hard negatives. To address these challenges, we propose a framework with two core contributions. At the feature extraction level, the Directional Adaptive Convolution Module (DACM) processes spike feature maps by sequentially performing horizontal strip convolution and vertical strip convolution, thereby capturing a more complete geometric structure of directional spike clusters. At the training supervision level, we propose a Dual-dimensional Progressive Reweighting (DPR) loss, which jointly characterizes sample difficulty from pairwise difficulty and positive-pair quality difficulty. A learnable fusion parameter is used to adaptively balance these two types of difficulty information. Experimental results on the University-1652 and SUES-200 benchmarks show that the proposed framework, when equipped with the same representation learning head as its ANN counterparts, achieves competitive and, in many settings, superior performance. In terms of energy efficiency, its estimated theoretical energy consumption is over $8.8\times$ lower than that of published ANN methods under their original configurations. Under a more rigorous matched ANN control that shares the identical architecture, the estimated energy is reduced from 29.84 mJ to 6.36 mJ, an approximately $4.7\times$ reduction obtained at a cost of only 2.29 percentage points in R@1. Full article

With the rapid evolution of Large Language Models (LLMs) in natural language understanding and generation, Relation Extraction (RE) has achieved substantial milestones in low-resource and open-domain scenarios. However, prevailing LLM-based RE methodologies predominantly rely on one-shot prompting or static workflows, which lack autonomous evaluation and iterative optimization mechanisms. Consequently, these approaches are prone to issues such as missing relations, type confusion, and factual hallucinations when navigating complex relational contexts. To address these limitations, this paper proposes DyFIRER (Dynamic Feedback Iterative Relation Extraction Framework), a multi-agent framework characterized by dynamic feedback. By constructing three functionally complementary agents—Extraction, Verification, and Optimization—the framework models the RE task as a closed-loop iterative process consisting of “extraction-verification-feedback-optimization,” thereby enabling dynamic adjustment and continuous refinement of extraction strategies. Experimental results on the DuIE 2.0 open relation extraction extension subset demonstrate that DyFIRER achieves an F1-score of 80.5%, modestly but statistically significantly outperforms GPT-4 (p = 0.014), a result that holds on both the augmented and non-augmented test sets and mainstream static methods (yielding a 10.3% improvement over Qwen-7B). Ablation studies further substantiate the critical role of the dynamic feedback iterative mechanism and the strategy retrieval module in mitigating complex relation omissions and factual hallucinations. The framework requires no additional annotated data or fine-tuning, suggesting potential applicability to low-resource settings, though this was not directly evaluated in the current study. Full article

Machine learning (ML) provides new opportunities to model the nonlinear relationships among composition, processing, microstructure, defects, properties, and in-service degradation of structural steels. This structured critical review examines ML applications to materials and process design, microstructural characterization, mechanical-property prediction, corrosion, fire and elevated-temperature performance, fatigue, fracture, and remaining-life assessment. Literature published up to 31 July 2026 was searched primarily through the Web of Science Core Collection and Scopus. A total of 110 publications were retained based on their relevance to structural steels, transparency of data and modeling procedures, and availability of information on validation or engineering applicability. The reviewed studies show that model suitability depends strongly on data modality, sample independence, feature representation, and validation strategy rather than on algorithm family alone. ML has progressed from property prediction toward process optimization, inverse materials design, environmental degradation assessment, and fatigue- and crack-related prognostics. However, independent cross-manufacturer, cross-laboratory, production-scale, and field validation remains limited, while uncertainty quantification and applicability-domain assessment are still inconsistently reported. These limitations are particularly important for corrosion, fire, fatigue, and remaining-life applications, where internally validated models should not be interpreted as substitutes for established physical models or design provisions. Future research should prioritize standardized multimodal data, physics-informed and uncertainty-aware modeling, prospective validation, and rigorously evaluated closed-loop monitoring and digital-twin frameworks for structural-steel life-cycle management. Full article

Sustainable management of biodegradable waste is a key component of the circular economy and resource recovery strategies. Composting is a complex biological process whose efficiency depends on numerous operational and physicochemical factors. Under industrial conditions, continuous laboratory monitoring of waste properties is often limited by time and cost constraints, necessitating reliable predictive tools to support process management. This study investigates the use of artificial neural networks (ANNs), decision trees (C&RT), and principal component analysis (PCA) for optimizing the composting of biodegradable waste under industrial-scale conditions. The research was conducted at a full-scale mechanical–biological treatment facility in Poland processing both the organic fraction mechanically derived from mixed municipal waste and separately collected biowaste. A dataset containing 23 records was developed from operational parameters (airflow, water addition, turning frequency, and process duration) and physicochemical properties of composted waste, including moisture content (MC), loss on ignition (LOI), total organic carbon (TOC), respiration activity (AT4), and higher heating value (HHV). The best-performing neural model achieved a predictive accuracy of 0.999 (coefficient of determination R² in the test set). For each of the neural networks, goodness of fit indices were also determined: MAE and RMSE. PCA confirmed strong relationships among key waste properties, while decision tree analysis identified airflow as the dominant operational factor affecting MC, LOI, and TOC, whereas turning frequency had the strongest influence on AT4. The results demonstrate that machine learning tools can effectively support industrial composting optimization by predicting operational parameters required to achieve desired waste stabilization characteristics, providing practical decision-support solutions for composting plant operators. It is recommended to implement single-output MLP models for dynamic, real-time process control and C&RT rules as emergency procedures. This study aligns with circular economy principles and the Sustainable Development Goals by demonstrating the potential of artificial intelligence to enhance sustainable biodegradable waste management, resource recovery, and industrial composting performance. Full article

Phytoene, a colorless precursor of carotenoids, has attracted increasing attention because of its favorable bioavailability, antioxidant activity, and potential applications in functional foods, nutraceuticals, and animal nutrition. However, its industrial utilization remains limited by low natural abundance and the dependence of current production strategies on genetic engineering or metabolic pathway manipulation. In this study, we identified and characterized a naturally occurring Brevundimonas vesicularis strain (Bv-xms2024) exhibiting pronounced phytoene accumulation without genetic modification. The strain was comprehensively characterized using morphological, biochemical, molecular, genomic, metabolomic, and transcriptional analyses. Quantitative LC–MS/MS analysis demonstrated that Bv-xms2024 accumulated phytoene to 420.42 ± 98.11 μg/g dry biomass after 96 h of cultivation, substantially exceeding the levels of downstream carotenoids, including β-carotene and astaxanthin. Optimization of cultivation parameters identified 25 °C, pH 7.0, and 96 h as the optimal conditions for phytoene accumulation, while serial passaging confirmed stable production over 20 generations. Genome annotation identified the carotenoid biosynthetic gene repertoire, while RT-qPCR analysis revealed a temporal shift from early upregulation of crtE and crtB to later upregulation of downstream pathway genes, consistent with the observed phytoene-dominant carotenoid profile. Short-term tolerance evaluations in mice and chickens revealed no observable adverse effects under the tested conditions. Collectively, these findings identify Bv-xms2024 as a promising natural microbial resource for phytoene production and provide a basis for further process development and strain-level safety evaluation. Full article

With the advancement of healthy city development and community renewal practices, the influence of perceived community environmental quality on residents’ psychological well-being has attracted increasing attention. Previous research has shown that the built environment is closely related to residents’ well-being; however, most studies have focused on objective spatial elements or facility provision and have paid insufficient attention to the mechanisms through which subjective environmental perceptions operate in the process by which the environment affects psychological well-being. From the perspective of Homo Urbanicus theory, this study takes 36 communities in Yangpu District, Shanghai, as cases and uses data from 882 valid questionnaires. Structural equation modeling is employed to examine the relationships among perceived community environmental quality, perceived environmental restorativeness, and residents’ psychological well-being, and to further explore age-related patterns among young-old, middle-old, and oldest-old groups. The results show that perceived community environmental quality has a significant positive effect on residents’ psychological well-being and produces a significant indirect effect through perceived environmental restorativeness. Exploratory age-stratified analyses further suggested different pathway patterns among the young-old, middle-old, and oldest-old groups. On this basis, the study further proposes a four-quadrant model of spatial contact opportunities and a five-category accessibility classification and accordingly develops community environment optimization strategies for older adults of different ages. The findings provide a theoretical basis and practical reference for healthy-community development and age-friendly community renewal. Full article

The production of second-generation bioethanol from lignocellulosic biomass requires efficient enzymatic hydrolysis and fermentation processes that remain effective at industrially relevant solids loadings. In this study, bleached Pinus taeda kraft pulp was evaluated as a model substrate for bioethanol production at bench scale (4 L reactor) under high-consistency conditions (12.5–13.9% solids). Three process configurations were compared: separate hydrolysis and fermentation (SHF), simultaneous saccharification and fermentation (SSF), and pre-hydrolysis followed by simultaneous saccharification and fermentation (pSSF). Enzymatic hydrolysis in the SHF and SSF configurations stabilized between 54% and 58%, indicating that hydrolysis was the main process bottleneck under the evaluated conditions. In contrast, Saccharomyces cerevisiae efficiently fermented the available glucose, achieving nearly complete conversion of glucose. Among the evaluated strategies, pSSF showed the highest ethanol yield and volumetric productivity, achieving an ethanol yield of 61.8% and a productivity of 0.61 g L⁻¹ h⁻¹. While laboratory-scale SSF experiments conducted at 2% solids achieved complete conversion, the ethanol yield decreased to approximately 58% at the bench scale, highlighting the impact of high-solids operation on process performance. The lower performance observed at high solids may be associated with factors commonly reported during scale-up, including increased slurry viscosity, reduced mixing efficiency, limited enzyme accessibility, and mass-transfer constraints. Overall, the results manifest the need to enhance hydrolysis performance through improved reactor design, more effective mixing strategies, and optimized high-solids processing to facilitate the scale-up of lignocellulosic bioethanol production. Full article

The ReFuelEU Aviation Regulation introduces mandatory targets for sustainable aviation fuels (SAF) from 2025 to 2050. However, hydrotreated esters and fatty acids (HEFA) technology based on waste oils alone is insufficient to meet targets beyond 2030, highlighting the need for alternative biocrude feedstocks to increase SAF production in the EU. Microalgae are promising feedstocks due to their biochemical composition and CO₂-utilization potential, although their high moisture content and nitrogen and oxygen levels require energy-efficient conversion technologies. Hydrothermal liquefaction (HTL) is a suitable process for converting wet microalgal biomass into biocrude, with an optimal temperature window of approximately 300–330 °C and typical biocrude yields ranging from 20 to 70 wt%, depending on feedstock composition and operating conditions. However, microalgal HTL remains at TRL 5–7 and faces challenges related to the high heteroatom content of the resulting biocrude. Hydrodeoxygenation (HDO) is a key upgrading step for converting biocrude into drop-in aviation fuels and commonly operates at approximately 250–400 °C and 10–30 MPa H₂ pressure. Nevertheless, few studies have addressed the HDO of microalgae-derived biocrude. This review examines microalgal HTL, pilot and demonstration facilities, biocrude yields and quality, and upgrading strategies for producing synthetic drop-in aviation biofuels. Full article