Management and Science University

Polyethylene glycol (PEG) is extensively utilized and its persistence in wastewater, soils, and aquatic environments has raised concerns regarding environmental contamination and ecological impacts. Understanding and predicting PEG’s heat capacity is therefore not only critical for material design but also for assessing its stability, transformation, and degradation under environmental conditions. This study develops advanced machine learning models including Gradient Boosting Decision Tree (GBDT) combined with Evolutionary Strategies (ES), Bayesian Probability Improvement (BPI), Batch Bayesian Optimization (BBO), and Gaussian Process Optimization (GPO) to forecast PEG’s molar heat capacity, using a databank of 528 experimental observations. Analysis of correlations demonstrated that temperature dominates the predictive framework (score 0.69), while molar mass provides a secondary but meaningful effect (score 0.51). Monte Carlo-based outlier detection validated data reliability, while sensitivity and SHapley Additive exPlanations (SHAP) analyses reinforced the primacy of temperature in determining heat capacity. The comparative analysis demonstrated that GBDT-ES provided the most accurate predictions, achieving an R2 of 0.998 and an average absolute relative error of 12.825

Human Action Recognition (HAR) is a significant research area in computer vision with applications in surveillance, healthcare, and human–computer interaction. This study proposes a hybrid deep learning framework that integrates the lightweight MobileNetV1 architecture with the scalable EfficientNetB3 network to achieve a balance between computational efficiency and recognition accuracy. The hybrid design leverages the complementary strengths of MobileNetV1 for efficient low-level feature extraction and EfficientNetB3 for enhanced high-level feature representation, enabling improved discriminative capability while maintaining reduced computational complexity. The model is evaluated on the UCF101 dataset, comprising 13,320 video clips across 101 action classes. A frame-based classification strategy with uniform sampling is adopted to reduce computational overhead, and data augmentation techniques such as rotation, shifting, shearing, and brightness adjustment are applied to improve generalization. Experimental results demonstrate that the proposed model achieves an accuracy of 89.18

Early detection of vision problems in children is essential for preventing developmental delays and academic challenges. Conventional vision screening methods are widely implemented; however, various barriers limit their effectiveness and accessibility. This systematic review aims to identify and analyse the barriers and limitations associated with conventional vision screening methods in children. A systematic search was conducted across the Scopus, Web of Science, and PubMed databases using predefined keywords and Boolean operators. Eligible studies included empirical investigations that examined barriers and limitations in paediatric vision screening. The review adhered to PRISMA 2020 guidelines, and quality assessment was performed using the Mixed Methods Appraisal Tool (MMAT). Of the studies screened, 25 met the inclusion criteria. Employing thematic analysis, five key barriers and limitations were discovered: (i) methodological limitations, (ii) resource constraints, (iii) competency gaps, (iv) socioeconomic and psychological barriers, and (v) policy and systemic challenges. Future research should focus on evaluating novel screening approaches that can overcome current limitations and enhance early detection rates for a broader range of paediatric vision conditions.

Background: This review discusses the application of eye-tracking technology in the detection and monitoring of binocular vision anomalies among children. Methods: A scoping review using PRISMA guidelines was conducted through Scopus, ScienceDirect, and PubMed using the keywords “eye-tracking,” “binocular,” “vision,” “anomalies,” “paediatrics,” and “children” from 2015 to 2025. Studies excluded were not written in English, did not apply the eye tracker as a research tool, involved an ineligible population, or involved non-human subjects. Results: The search strategy identified 77 citations, yet only 14 studies met the inclusion criteria. This review revealed a variety of binocular vision anomalies detectable through eye-tracking systems, along with the specific models and parameters employed in these assessments. Application of eye-tracking technology in diagnosing conditions such as strabismus and amblyopia demonstrated potential for improved accuracy and early detection. Discussion: Eye-tracking technology demonstrates considerable potential for the detection and monitoring of binocular vision anomalies in children, particularly as a non-invasive method for early screening, thereby strengthening its clinical applicability. By assessing fixation stability, saccadic movements, and vergence responses, eye-tracking allows for the early detection of subtle visual anomalies, especially in the paediatric population. Conclusions: Eye-tracking technology represents a valuable advancement in paediatric vision care, enabling the more objective and earlier detection of binocular vision anomalies in the paediatric population.

Fibroblast activation protein (FAP), widely overexpressed in the tumor stroma of various cancers, has emerged as a promising target for cancer imaging. Radiopharmaceuticals designed to target FAP offer significantly improved tumor-to-background contrast compared to traditional imaging agents, thereby enhancing the precision of cancer detection. This review summarizes recent advancements in FAP-targeted diagnosis radiopharmaceuticals, highlighting their advantages and future potential.