Search Results (2,573)

This study explores fine-tuning methods and dataset structures for multilingual neural machine translation using the No Language Left Behind model, with a case study on Kazakh, English, and Russian. We compare single-stage and two-stage fine-tuning approaches, as well as triplet versus non-triplet dataset configurations, to improve translation quality. A high-quality, 50,000-triplet dataset in information technology domain, manually translated and expert-validated, serves as the in-domain benchmark, complemented by out-of-domain corpora like KazParC. Evaluations using BLEU, chrF, METEOR, and TER metrics reveal that single-stage fine-tuning excels for low-resource pairs (e.g., 0.48 BLEU, 0.77 chrF for Kazakh → Russian), while two-stage fine-tuning benefits high-resource pairs (Russian → English). Triplet datasets improve cross-linguistic consistency compared with non-triplet structures. Our reproducible framework offers practical guidance for adapting neural machine translation to technical domains and low-resource languages. Full article

In this paper, we propose a fuzzy formulation of the classic Frankot–Chellappa algorithm by which surfaces can be reconstructed using normal vectors. In the fuzzy formulation, the surface normal vectors may be uncertain or ambiguous, yielding a fuzzy Poisson partial differential equation that requires appropriate definitions of fuzzy derivatives. The solution of the resulting fuzzy model is approached by adopting a fuzzy variant of the discrete sine transform, which results in a fast and robust algorithm for surface reconstruction. An adaptive defuzzification strategy is also introduced to improve noise handling in highly uncertain regions. In experiments, we demonstrate that our fuzzy Frankot–Chellappa algorithm achieves accuracy on par with the classic approach for smooth surfaces and offers improved robustness in the presence of noisy normal data. We also show that it can naturally handle missing data (such as gaps) in the normal field by filling them using neighboring information. Full article

Microglia, the brain’s resident innate immune cells, play a fundamental role in maintaining neural homeostasis and mediating responses to injury or infection. Upon activation, microglia undergo morphological and functional changes, including phenotypic switching between pro- and anti-inflammatory types and the release of different inflammatory mediators. These processes contribute to neuroprotection and the pathogenesis of various central nervous system (CNS) disorders. Mast cells, although sparsely located in the brain, exert a significant influence on neuroinflammation through their interactions with microglia. Through degranulation and secretion of different mediators, mast cells disrupt the blood–brain barrier and modulate microglial responses, including alteration of microglial phenotypes. Notably, mast cell-derived factors, such as histamine, interleukins, and tryptase, activate microglia through various pathways including protease-activated receptor 2 and purinergic receptors. These interactions amplify inflammatory cascades via various signaling pathways. Previous studies have revealed an exceedingly complex crosstalk between mast cells and microglia suggesting a bidirectional regulation of CNS immunity, implicating their cooperation in both neurodegenerative progression and repair mechanisms. Here, we review some of the diverse communication pathways involved in this complex interplay. Understanding this crosstalk may offer novel insights into the cellular dynamics of neuroinflammation and highlight potential therapeutic targets for a variety of CNS disorders. Full article

In the agricultural domain, the availability of labelled data for disease recognition tasks is often limited due to the cost and expertise required for annotation. In this paper, a novel semi-supervised learning framework named TOM-SSL is proposed for automatic tomato leaf disease recognition using pseudo-labelling. TOM-SSL effectively addresses the challenge of limited labelled data by leveraging a small labelled subset and confidently pseudo-labelled samples from a large pool of unlabelled data to improve classification performance. Utilising only 10% of the labelled data, the proposed framework with a MobileNetV3-Small backbone achieves the best accuracy at 72.51% on the tomato subset of the PlantVillage dataset and 70.87% on the Taiwan tomato leaf disease dataset across 10 disease categories in PlantVillage and 6 in the Taiwan dataset. While achieving recognition performance on par with current state-of-the-art supervised methods, notably, the proposed approach offers a tenfold enhancement in label efficiency. Full article

Demand Side Management (DSM) plays a crucial role in modern energy systems, enabling more efficient use of energy resources and contributing to the sustainability of the power grid. This study examines DSM strategies within a multi-environment context encompassing residential, commercial, and industrial sectors, with a focus on diverse appliance types that exhibit distinct operational characteristics and user preferences. Initially, a single-objective optimization approach using Genetic Algorithms (GAs) is employed to minimize the total energy cost under a real Time-of-Use (ToU) pricing scheme. This heuristic method allows for the effective scheduling of appliance operations while factoring in their unique characteristics such as power consumption, usage duration, and user-defined operational flexibility. This study extends the optimization problem to a multi-objective framework that incorporates the minimization of CO₂ emissions under a real annual energy mix while also accounting for user discomfort. The Non-dominated Sorting Genetic Algorithm II (NSGA-II) is utilized for this purpose, providing a Pareto-optimal set of solutions that balances these competing objectives. The inclusion of multiple objectives ensures a comprehensive assessment of DSM strategies, aiming to reduce environmental impact and enhance user satisfaction. Additionally, this study monitors the Peak-to-Average Ratio (PAR) to evaluate the impact of DSM strategies on load balancing and grid stability. It also analyzes the impact of considering different periods of the year with the associated ToU hourly schedule and CO₂ emissions hourly profile. A key innovation of this research is the integration of detailed, category-specific metrics that enable the disaggregation of costs, emissions, and user discomfort across residential, commercial, and industrial appliances. This granularity enables stakeholders to implement tailored strategies that align with specific operational goals and regulatory compliance. Also, the emphasis on a user discomfort indicator allows us to explore the flexibility available in such DSM mechanisms. The results demonstrate the effectiveness of the proposed multi-objective optimization approach in achieving significant cost savings that may reach 20% for industrial applications, while the order of magnitude of the trade-offs involved in terms of emissions reduction, improvement in discomfort, and PAR reduction is quantified for different frameworks. The outcomes not only underscore the efficacy of applying advanced optimization frameworks to real-world problems but also point to pathways for future research in smart energy management. This comprehensive analysis highlights the potential of advanced DSM techniques to enhance the sustainability and resilience of energy systems while also offering valuable policy implications. Full article

Water treadmill (WT) exercise is used for horses’ rehabilitation and training. Given that each training needs to be individualized for each horse, the goal is to assess whether infrared thermography (IRT) can serve as a non-invasive tool for daily monitoring of individual training and rehabilitation progress in horses undergoing WT exercise. Fifteen Polish Warmblood school horses were subjected to five WT sessions: dry treadmill, fetlock-depth water, fetlock-depth water with artificial river (AR), carpal-depth water, and carpal-depth water with AR. IRT images, collected pre- and post-exercise, were analyzed for the mean temperature (Tmean) and maximal temperature (Tmax) across 14 regions of interest (ROIs) representing the body surface overlying specific superficial muscles. While on a dry treadmill, Tmean and Tmax increased post-exercise in all ROIs; wetting of the hair coat limited surface temperature analysis in ROIs annotated on limbs. Tmax over the m. brachiocephalicus, m. trapezius pars cervicalis, m. triceps brachii, and m. semitendinosus increased during walking in carpal-depth water, which therefore may be suggested as an indirect indicator of increased activity related to forelimb protraction and flexion–extension of the limb joints. Tmax over the m. latissimus dorsi and m. longissimus increased during carpal-depth WT exercise with active AR mode, which may be suggested as an indicator of increased workload including vertical displacement of the trunk. Full article

The alloy Pr₂Fe_16.75Ni_0.25 has been examined to investigate its structural properties, critical behavior, and magnetocaloric effects. Rietveld’s refinement of X-ray diffraction patterns has revealed a rhombohedral structure with an $R\overline{3}m$ space group. Pr₂Fe_16.9Ni_0.25 also demonstrates a direct magnetocaloric effect near room temperature, accompanied by a moderate magnetic entropy change ($-\mathrm{\Delta }{S}_M^{\max }$ = 5.5 J kg⁻¹ K⁻¹ at ${\mu }_0\mathrm{\Delta }H=5$ T) and a broad working temperature range. Furthermore, the Relative Cooling Power (RCP) is approximately 89% of the widely recognized gadolinium (Gd) for ${\mu }_0\mathrm{\Delta }H=2$ T. This compound exhibits a commendable magnetocaloric response, on par with or even surpassing that of numerous other intermetallic alloys. Critical behavior was analyzed using thermo-magnetic measurements, employing methods such as the modified Arrott plot, critical isotherm analysis, and Kouvel-Fisher techniques. The obtained critical exponents ($\beta$, $\gamma$, and $\delta$) exhibit similarities to those of the 3D-Ising model, characterized explicitly by intermediate range interactions. Full article

Background/Objectives: Infectious endophthalmitis is a vision-threatening complication of intraocular surgery and intravitreal injections. Standard treatment involves intravitreal antibiotics; however, concerns regarding multidrug resistance and vancomycin-associated hemorrhagic occlusive retinal vasculitis (HORV) highlight the need for alternative antimicrobial strategies. This study aimed to evaluate the clinical efficacy and safety of a protocol combining intravitreal injection of 1.25% povidone-iodine (PI) with intraoperative irrigation using low concentrations of vancomycin and ceftazidime. Methods: We retrospectively analyzed 11 eyes from patients diagnosed with postoperative or injection-related endophthalmitis. Six of the eleven cases received an initial intravitreal injection of 1.25% PI, followed by pars plana vitrectomy with irrigation using balanced salt solution PLUS containing vancomycin (20 μg/mL) and ceftazidime (40 μg/mL). A second intravitreal PI injection was administered at the end of surgery in all cases. Additional PI injections were administered postoperatively based on clinical response. Clinical outcomes included best-corrected visual acuity (BCVA), microbial culture results, corneal endothelial cell density, and visual field testing. Results: All eyes achieved complete infection resolution without recurrence. The mean BCVA improved significantly from 2.18 logMAR at baseline to 0.296 logMAR at final follow-up (p < 0.001). No adverse events were observed on specular microscopy or visual field assessment. The protocol was well tolerated, and repeated PI injections showed no signs of ocular toxicity. Conclusions: This combination protocol provides a safe and effective treatment strategy for infectious endophthalmitis. It enables rapid and complete infection resolution while minimizing the risks associated with intravitreal antibiotics. These findings support further investigation of this protocol as a practical and globally accessible alternative to standard intravitreal antimicrobial therapy. Full article

Addressing the limitations of existing path planning algorithms for mobile robots in complex environments, such as poor adaptability, low convergence efficiency, and poor path quality, this study establishes a clear connection between mobile robots and real-world challenges such as unknown environments, dynamic obstacle avoidance, and smooth motion through innovative strategies. A novel multi-strategy fusion white shark optimization algorithm is proposed, focusing on actual scenario requirements, to provide optimal solutions for mobile robot path planning. First, the Chaotic Elite Pool strategy is employed to generate an elite population, enhancing population diversity and improving the quality of initial solutions, thereby boosting the algorithm’s global search capability. Second, adaptive weights are introduced, and the traditional simulated annealing algorithm is improved to obtain the Rapid Annealing Method. The improved simulated annealing algorithm is then combined with the White Shark algorithm to avoid getting stuck in local optima and accelerate convergence speed. Finally, third-order Bézier curves are used to smooth the path. Path length and path smoothness are used as fitness evaluation metrics, and an evaluation function is established in conjunction with a non-complete model that reflects actual motion to assess the effectiveness of path planning. Simulation results show that on the simple 20 × 20 grid map, the fusion of the Fused Multi-strategy White Shark Optimisation algorithm (FMWSO) outperforms WSO, D*, A*, and GWO by 8.43%, 7.37%, 2.08%, and 2.65%, respectively, in terms of path length. On the more complex 40 × 40 grid map, it improved by 6.48%, 26.76%, 0.95%, and 2.05%, respectively. The number of turning points was the lowest in both maps, and the path smoothness was lower. The algorithm’s runtime is optimal on the 20 × 20 map, outperforming other algorithms by 40.11%, 25.93%, 31.16%, and 9.51%, respectively. On the 40 × 40 map, it is on par with A*, and outperforms WSO, D*, and GWO by 14.01%, 157.38%, and 3.48%, respectively. The path planning performance is significantly better than other algorithms. Full article

End-to-end relation extraction (E2ERE) generally performs named entity recognition and relation extraction either simultaneously or sequentially. While numerous studies on E2ERE have centered on enhancing span representations to improve model performance, challenges remain due to the gaps between subtasks (named entity recognition and relation extraction) and the modeling discrepancies between entities and relations. In this paper, we propose a novel Label Annotation Interaction-based representation enhancement method for E2ERE, which institutes a two-phase semantic interaction to augment representations. Specifically, we firstly feed label annotations that are easy to manually annotate into a language model, and conduct the first-round interaction between three types of tokens with a partial attention mechanism; Then we construct a latent multi-view graph to capture various possible links between label and entity (pair) nodes, facilitating the second-round interaction between entities and labels. A series of comparative experiments with methods of various transformer-based architectures currently in use show that LAI-Net can maintain performance on par with the current SOTA in terms of NER task, and achieves significant improvements over existing SOTA models in terms of RE task. Full article

Background: Surgical management of large full-thickness macular holes (MHs) remains challenging, particularly when aiming for both rapid visual recovery and consistent anatomical closure without inducing retinal trauma. This retrospective single-center study evaluated the efficacy of activated conditioned plasma (ACP) as an intraoperative coadjuvant supporting ILM (internal limiting membrane) peeling and air tamponade in the treatment of idiopathic MHs measuring 400–800 µm, under real-time intraoperative optical coherence tomography (i-OCT) guidance. Methods: Seventy eyes from fifty patients underwent pars plana vitrectomy with intraoperative ACP application. ACP, a leukocyte-poor autologous platelet concentrate, was used intraoperatively as a coadjuvant to ILM peeling and air tamponade. It facilitated the formation of a transparent fibrin membrane over the retinal surface, supporting edge approximation and promoting retinal healing. Results: The primary outcome was complete MH closure confirmed by OCT; the secondary outcome was improvement in BCVA on postoperative day 7 and during a 12-month follow-up. Anatomical closure was achieved in 98.6% of cases. On day 7, 78.6% of eyes showed a ≥ three-line BCVA improvement, with mean BCVA increasing from 0.25 ± 0.21 to 0.69 ± 0.20 (p < 0.001). These outcomes remained stable throughout the follow-up. No significant intraoperative or postoperative complications were observed. Conclusions: The combination of ACP and i-OCT appears to be a safe and effective strategy for anatomical foveolar reconstruction, enabling early visual recovery while minimizing inflammation and fibrotic scarring associated with conventional techniques. Full article

Carbon dioxide (CO₂) emissions due to human activities are responsible for approximately 80% of the drivers of global warming, resulting in a 1.1 °C increase above pre-industrial temperatures. This study quantified the CO₂ assimilation and productivity of the brown macroalgae Lessonia spicata in the central Pacific coast of Chile, across seasonal and daily cycles, under different environmental stressors, such as temperature and solar irradiance. Measurements were performed using an infra-red gas analysis (IRGA) instrument which had a chamber allowing for precise quantification of CO₂ concentrations; additional photophysiological and biochemical responses were also measured. CO₂ assimilation, along with the productivity and biosynthesis of proteins and lipids, increased during the spring, coinciding with moderate temperatures (~14 °C) and high photosynthetically active radiation (PAR). Furthermore, the increased production of photoprotective and antioxidant compounds, including phenolic compounds, and carotenoids, along with the enhancement of non-photochemical quenching (NPQ), contribute to the effective photoacclimation strategies of L. spicata. Principal component analysis (PCA) revealed seasonal associations between productivity, reactive oxygen species (ROSs), and biochemical indicators, particularly during the spring and summer. These associations, further supported by Pearson correlation analyses, suggest a high but seasonally constrained photoacclimation capacity. In contrast, the reduced productivity and photoprotection observed in the summer suggest increased physiological vulnerability to heat and light stress. Overall, our findings position L. spicata as a promising nature-based solution for climate change mitigation. Full article

This study aimed to evaluate the impact of preoperative anti-VEGF injections on pars plana vitrectomy (PPV) outcomes in patients with proliferative diabetic retinopathy (PDR). Material and methods: We analysed 232 eyes with proliferative diabetic vitreoretinopathy treated with posterior vitrectomy. There were 112 women and 120 men. The patients were divided into two groups of 116 eyes each. In 116 eyes (study group), an anti-VEGF injection was administered 3 to 5 days before vitrectomy. The control eyes were not injected with anti-VEGF due to systemic contraindications to anti-VEGF treatment or lack of patient consent. All participants underwent pars plana vitrectomy with silicone oil injection. The oil was removed within 2–3 months after PPV. Results: At 2 years of observation, after removal of silicone oil, visual acuity (VA) was 0.24 ± 0.27 logMAR in the study and 0.37 ± 0.45 logMAR in the control group (p = 0.003). Intraocular pressure was 16.84 ± 6.25 mmHg in the study group and 17.78 ± 6.22 mmHg in the control group (p = 0.04). The mean duration of surgery was 47.62 ± 9.87 and 50.05 ± 9.41 min in the study and control groups, respectively (p = 0.02). The size of intraoperative haemorrhage was 0.97 ± 0.86 dd in the study group and 1.51 ± 1.22 dd in the control group (p = 0.003). The frequency of surgery-induced retinal breaks was 0.34 ± 0.56 in the study group and 0.56 ± 0.76 in the control group (p = 0.003). The recurrence rate of retinal detachment was 0.05 ± 0.22 in the study group and 0.1 ± 0.31 in the control group (p = 0.15). Conclusions: Preoperative anti-VEGF therapy shortens the duration of surgery, reduces complications, and improves long-term outcomes in terms of visual acuity and maintenance of normal eye function. Full article

Background: A precise radiographic evaluation of adolescent idiopathic scoliosis (AIS) is essential for effective treatment planning and follow-up. The pelvic axial rotation (PAR) and horizontal balance of the pelvis are critical factors to consider throughout the treatment and monitoring of AIS. While some previous studies have examined spinal curvature in relation to PAR direction and the direction of the major curve (DMC) in AIS patients, this study aims to explore the relationship between scoliosis morphology, pelvic axial rotation (PAR), and the direction of the major curve in patients with adolescent idiopathic scoliosis. Methods: Radiographic images of 397 patients diagnosed with AIS between 2023 and 2024 at a Tertiary Referral Hospital were retrospectively evaluated. Morphological and morphometric measurements, including sex, Lenke and Risser classifications, lower leg discrepancy, Cobb angle, PAR direction, and major curvature direction, were performed. Results: The mean age of the 397 patients (246 female, 151 male) was 14.47 ± 2.29. There is no significant correlation between PAR and DMC (p = 0.919). No significant differences were found in terms of sex (p = 0.603). Regardless of the PAR direction, major curvature was more common on the left side (57.7%). Furthermore, a positive correlation was noted between the Cobb angle and LLD. Conclusions: Our study contributes to a growing body of literature questioning the deterministic role of PAR in AIS. While previous reports have emphasized the directional correlation between the pelvis and spinal curvature, our findings suggest that pelvic rotation may not be a reliable indicator of curve direction in all patients. This highlights the complexity of AIS biomechanics and underscores the need for individualized radiographic and clinical evaluation rather than a reliance on generalized compensatory models. Full article

Machine learning algorithms are commonly employed for classification and interpretation of high-dimensional data. The classification task is often broken down into two separate procedures, and different methods are applied to achieve accurate results and produce interpretable outcomes. First, an effective subset of high-dimensional features must be extracted and then the selected subset will be used to train a classifier. Gradient Boosted Trees (GBT) is an ensemble model and, particularly due to their robustness, ability to model complex nonlinear interactions, and feature interpretability, they are well suited for complex applications. XGBoost (eXtreme Gradient Boosting) is a high-performance implementation of GBT that incorporates regularization, parallel computation, and efficient tree pruning that makes it a suitable efficient, interpretable, and scalable classifier with potential applications to medical data analysis. In this study, a Monte Carlo Gradient Boosted Trees (MCGBT) model is proposed for both feature reduction and classification. The proposed MCGBT method was applied to a lung cancer dataset for feature identification and classification. The dataset contains 107 radiomics which are quantitative imaging biomarkers extracted from CT scans. A reduced set of 12 radiomics were identified, and patients were classified into different cancer stages. Cancer staging accuracy of 90.3% across 100 independent runs was achieved which was on par with that obtained using the full set of 107 radiomics, enabling lean and deployable classifiers. Full article