Correction: Artificial intelligence assessment of valvular disease and ventricular function by a single echocardiography view

This article aims to analyze the effects of platformization in education and its implications for teaching praxis, based on a bibliographic review (2019-2024). We examine how the integration of digital infrastructures reconfigures pedagogical practices from the perspective of the attention economy and behavioral modulation, considering the context of acceleration of these processes intensified by the COVID-19 pandemic. To this end, we conducted a content analysis of 36 academic texts, organized into four categories: (1) Discourses and Commodification of Education; (2) Technologies, Data, and Strategies; (3) Teaching Praxis; and (4) Critical Assessments of Big Data, Artificial Intelligence, and Platforms in Education. The results highlight the consolidation of business models based on data extraction, driven by Big Techs and facilitated by public policies that normalize the adoption of these technologies. The discussion demonstrates that, despite discourses focused on technical efficiency, teaching praxis faces the challenge of imposed algorithmic logics that threaten pedagogical autonomy and theoretical pluralism. It is concluded that platformization is not a neutral phenomenon, requiring the strengthening of digital sovereignty, the development of critical public policies, and reflective teacher training to resist data colonialism and ensure the integrity of educational processes.

Efficient gas drainage is crucial for safe coal mine production and the clean utilization of gas resources. Despite recent advances, complex geological conditions and unstable system operation limit the effectiveness of traditional monitoring in underground borehole-pipe-pump systems. This study conducts controlled experiments to analyze the operational behavior of the gas drainage network under various leakage scenarios, quantitatively revealing characteristic patterns in negative pressure and flow rate. Based on these insights, an intelligent gas-drainage performance evaluation model using a Convolutional Neural Network (CNN) is developed to automate classification of drainage effectiveness. Experimental results using 10,000 samples from Xinfa Coal Mine show that the CNN model achieves optimal performance with a learning rate of 0.1 and batch size of 256, reaching classification accuracies of 100% for Classes I-III, 93% for Class IV, and 50% for Class V. The proposed approach integrates experimental simulation, leakage characterization, and deep-learning-based evaluation into a unified framework, providing an effective solution for real-time monitoring and intelligent assessment of gas drainage systems. This study offers technical support for improving gas extraction efficiency, enhancing mine safety, and promoting the clean and efficient utilization of coal-mine gas.

The advent of generative artificial intelligence (GenAI) is already impacting pedagogical strategies and assessment methodologies in higher education, particularly in the biological sciences, which have traditionally relied heavily on written assessments. GenAI's rapid and plausible text generation capabilities challenge traditional written assessments and prompt a shift toward more authentic assessment types. This article explores innovative applications of GenAI in biology education through case studies presented at a recent workshop. These case studies illustrate how GenAI has the potential to enhance academic activities, from developing learning resources to fostering student engagement through active learning strategies. The discussion highlights a shift from product-oriented assessments to process-oriented approaches that prioritize continuous interaction, iteration, and reflection among learners. Despite GenAI's reliance on preexisting data, raising concerns about originality and contextual accuracy, and its limitations in tasks requiring high creativity and deep understanding, it has the potential to enhance educational practices when applied with awareness of its constraints. The article concludes with a balanced analysis of the transformative impact and inherent challenges of integrating GenAI into biology education, advocating for thoughtful implementation to ensure it augments rather than replaces traditional teaching methods.NEW & NOTEWORTHY Generative artificial intelligence (GenAI) is transforming higher education by enabling rapid learning resource development, enhancing student engagement, and supporting authentic assessment. Our workshop-based case studies highlight GenAI's ability to foster interactive, process-oriented learning in biosciences while addressing challenges with creativity and originality. From creating tailored quizzes to promoting active learning and ethical AI use, these strategies empower educators to integrate AI responsibly, ensuring it enriches teaching and learning in bioscience education while maintaining academic integrity.

Towards Intelligent Safety Assessment Under Blast Loading: A Large Language Model-Powered Multi-Agent Framework

Municipal solid waste incineration state recognition system based on deep convolutional stochastic configuration machine.

An assessment and prediction system for the quality of public water networks was developed, using Timișoara, Romania, as a case study. This was implemented on a Google Firebase cloud storage system and comprised twelve ML algorithms applied to test samples for drinkability and used in predictions of upcoming samples. The system compares 17 water quality parameters to the World Health Organization and public reports of Timișoara drinking water standards for 804 samples. The system provides real-time data storage, drinkability prediction for the reservoir water system, and rule-based critical water recommendations for elementary treatment in samples. The most accurate and best-calibrated against random forest, gradient boosting, and Logistic Regression algorithms was the decision tree algorithm of the ML models. The experimental findings also determine the regions of the worst and best water quality and propose respective treatment. In contrast to previous research and structures, the paper demonstrates an approved stable solution for smart water monitoring, correlating practical deployment with sophisticated data-based conclusions. The results contribute to enhancing public health, enhancing water management measures, and upscaling the system for larger-scale applications.

Background/Objectives: Robotic-assisted radical prostatectomy (RARP) is a standard treatment for localized and locally advanced prostate cancer; however, optimizing oncologic control while preserving urinary continence and erectile function remains challenging. Advances in preoperative imaging, molecular diagnostics, artificial intelligence (AI), and intraoperative assessment have the potential to refine surgical planning and execution. This review summarizes contemporary evidence on advanced imaging and intraoperative technologies used to optimize RARP outcomes. Methods: A narrative literature review was conducted of English-language studies published between 2015 and 2025 using PubMed/MEDLINE, Scopus, and Google Scholar. Studies evaluating multi-parametric and bi-parametric MRI, prostate-specific membrane antigen-based positron emission tomography/computed tomography (PSMA PET/CT), AI-assisted tumor modeling, and intraoperative histologic or molecular imaging techniques in the context of robotic-assisted radical prostatectomy were included. Evidence from randomized controlled trials, prospective and retrospective studies, technical feasibility reports, and expert consensus statements was reviewed. Results: MRI remains central to anatomic mapping and local staging but consistently underestimates true tumor extent, with implications for margin control. AI-assisted platforms improve tumor contouring accuracy and may meaningfully influence surgical decision-making. PSMA-based imaging enhances detection of extra-prostatic extension and nodal disease and shows early promise for ex vivo and intraoperative guidance. Intraoperative margin assessment techniques are supported by randomized evidence demonstrating improved functional outcomes without compromising short-term oncologic safety and emerging digital histologic technologies offer scalable alternatives for real-time margin evaluation. Conclusions: Integration of advanced anatomic, molecular, and intraoperative imaging technologies represents an evolving multimodal paradigm in RARP. Combined use of MRI, PSMA-based imaging, AI-assisted modeling, and rapid histologic assessment may enable more precise, individualized surgery that balances oncologic control with functional preservation. Further validation is required to define optimal implementation in routine clinical practice.

Small and micro enterprises (SMEs) play a critical role in economic development, yet their access to credit is often constrained by inadequate risk assessment frameworks. Traditional credit scoring models struggle to capture the non-linearity, feature sparsity, and class imbalance inherent in SME financial data. To address these challenges, the authors propose HECRO (Heterogeneous Ensemble Credit Risk Optimizer), a multi-layered framework that integrates kernel-based and heuristic feature selection, ensemble base learners, and a Bayesian-optimized meta-stacking classifier. HECRO leverages KFS-MCLOC and BOWOA-KS for robust feature extraction, followed by GA-BPNN, SMOTE-XGBoost, Wide & Deep, and BD-LR models as diverse predictors, culminating in a BO-XGBoost meta-learner. SHAP-based interpretation enhances post-hoc transparency. These results demonstrate HECRO's superiority in both predictive accuracy and robustness. The study offers a practical and scalable solution for SME credit evaluation, providing new insights into the design of intelligent financial risk assessment systems.

Milk tea is a globally popular new-style tea beverage product. In recent years, the industry has achieved rapid development in terms of scale expansion and quality iteration and upgrading. The flavor quality and product stability have become the focus of attention and research hotspots in this field. The chemical foundation of milk tea flavor, processing methods, and flavor quality evaluation approaches are thoroughly elaborated. The chemical basis of tea-based, milk-based, and milk tea flavors is systematically summarized, primarily including the analysis of key flavor compounds and the interactions between tea-based and milk-based substances. Subsequently, the tea-based production methods, mixed processing techniques, and factors influencing storage and preservation of milk tea are discussed. Furthermore, evaluation methods for milk tea flavor quality, including traditional sensory evaluation and intelligent assessment techniques are systematically outlined. This review not only summarizes the recent research progress but also looks forward to the interdisciplinary work that needs to be carried out in the future. These efforts aim to provide information on the transformation from the research stage of tea milk product formulas to the development of solutions with controllable quality. Thus, they offer valuable theoretical guidance for the formation and regulation of tea milk flavor and quality as well as the development of new products. This work aims to provide theoretical insights and technical support for the translation from laboratory formulations to quality-controlled industrial solutions.

This study aimed to develop an intelligent quality assessment system for Codonopsis Radix based on machine learning modeling. First, Codonopsis Radix samples from six origins were grouped based on pharmacological and chemical indicators, integrating pharmacodynamic evaluations using impaired spleen and lung function animal models with compositional analysis of the alcohol-soluble extract and polysaccharide contents. Subsequently, an electronic nose was employed to objectively quantify their odor profiles. A machine learning-based modeling framework was constructed by integrating feature extraction, feature selection, and pattern recognition techniques. The classification model built by combining electronic nose data with machine learning algorithms demonstrated highly effective discriminatory capability in cross-validation. SHapley Additive exPlanations analysis identified sensors S8, S15, S16, and S18 as critical variables for classification. Concurrently, regression models were established to predict the alcohol-soluble extract and polysaccharide contents. Given the limited sample size, feature expansion and data augmentation strategies were applied exclusively to the training set to enhance model robustness. In summary, the proposed interpretable modeling approach, which integrates pharmacological efficacy, chemical composition, and electronic nose data, provides a referential technical pathway for similar studies.

To cultivate composite medical professionals capable of adapting to the development of intelligent healthcare,this consensus is grounded in the competency-based medical education,integrating the competency model and Miller's pyramid of clinical competence. A two-round Delphi method involving a multidisciplinary expert panel was conducted,combined with a systematic literature review,to develop a 21-indicator artificial intelligence(AI) literacy competency list for medical students across three domains:knowledge (8 indicators),skills (8 indicators),and attitudes (5 indicators). Furthermore,the consensus proposes a practical assessment system:standardized testing for the knowledge domain,situational judgment tests for the attitudes domain,and objective structured clinical examinations incorporating AI-related scenarios for the skills domain. In addition,a longitudinal assessment strategy spanning the phases of admission,preclinical training,and clinical training is recommended. The competency list and assessment framework established in this consensus demonstrate strong scientific rigor,authority,and practical applicability,and can serve as an important reference for medical schools seeking to advance the deep integration of AI and medical education and to cultivate composite medical talents suited to the era of intelligent healthcare.

Transcatheter edge-to-edge mitral valve repair is a key therapeutic option for patients with severe symptomatic mitral regurgitation at high surgical risk. This prospective study aimed to develop a novel end-to-end deep learning model for preoperative artificial intelligence assessment in transcatheter edge-to-edge mitral valve repair (TEERAI-pre) candidates using multiview, multimodal echocardiography. TEERAI-pre, a video vision transformer-based classification model, predicts morphological suitability for transcatheter edge-to-edge mitral valve repair from multiview, multimodal echocardiography. A transformer-based feature-level fusion module was designed in TEERAI-pre to integrate multiview, multimodal features for final prediction. An internal data set of 633 patients (7997 transthoracic echocardiographic videos; 766 pulsed-wave Doppler images) was split for 5-fold cross-validation. An external data set of 150 patients (1735 transthoracic echocardiographic videos; 169 pulsed-wave Doppler images) across 2 hospitals evaluated generalizability. Reference standards were provided by 2 experienced valvular cardiologists per international guidelines. On the internal data set, TEERAI-pre achieved 75.0% accuracy (95% CI, 71.7%-78.4%) for classifying red (unsuitable), yellow (challenging), and green (ideal) zones, with 77.1% precision, 75.5% recall, and 76.2% F1 score. External validation yielded 73.3% accuracy, 74.0% precision, and 74.0% recall. Multiview multimodal integration improved performance. Binary classification (red versus green) showed TEERAI-pre matched senior experts and outperformed intermediate/junior echocardiologists. Feature-level fusion outperformed output-level fusion and single-view model. Backbone selection and calibration analysis confirmed robust performance. TEERAI-pre demonstrates strong performance in transcatheter edge-to-edge mitral valve repair preoperative assessment using transthoracic echocardiographic videos and images, supporting more accurate patient selection and enhancing clinical workflow efficiency. URL: clinicaltrials.gov; Unique Identifier: NCT05508438.

The aim of this study is to analyze the challenges of assessing spiritual intelligence in Islamic Religious Education, focusing on the gap between cognitive and affective-spiritual aspects. The research uses a qualitative case study approach through interviews, observations, and document analysis. The findings show that assessment practices are still dominated by cognitive methods such as written tests, memorization, and oral examinations, while aspects of attitude, worship, and character are difficult to measure objectively. As a result, students demonstrate sufficient religious knowledge but lack spiritual and moral maturity. The study recommends developing holistic assessment instruments, providing teacher training, and utilizing digital technology to support more balanced assessment of knowledge and spiritual intelligence.

This review presents a comprehensive examination of intelligent software models designed for predictive risk assessment through the application of advanced artificial intelligence design principles. Predictive risk assessment has become increasingly critical across multiple domains including finance, healthcare, cybersecurity, manufacturing, and supply chain management. The integration of sophisticated AI methodologies including deep learning, ensemble methods, and neural architectures has revolutionized the capability to forecast, quantify, and mitigate risks before they materialize. This study synthesizes current literature on AI-driven risk prediction systems, analyzes their architectural foundations, evaluates design principles such as explainability, robustness, scalability, adaptability, fairness, and privacy, and identifies emerging trends and challenges. The findings indicate that successful implementation of intelligent risk assessment models requires a holistic approach combining advanced algorithms, robust data pipelines, ethical considerations, and domain-specific customization. This review provides valuable insights for researchers, practitioners, and policymakers seeking to leverage AI for enhanced risk management capabilities.

With the proliferation of data-driven methods in automotive noise, vibration, and harshness (NVH) analysis, the digital transformation of NVH performance evaluation has become increasingly imperative. However, in the actual testing process, signals are inevitably affected by abnormal interference, resulting in a decrease in the evaluation accuracy of NVH performance. To solve this problem, we propose a new adaptive anomaly detection and correction framework. The key methodological innovation lies in the IResNet–IFCNN collaborative architecture, which introduces an improved residual network (IResNet) for high-precision anomaly identification and an improved fully connected network (IFCNN) for adaptive multi-condition correction. The main contribution is a closed-loop “detection – matching – correction” mechanism, which dynamically selects specific weights based on the type of anomaly. Verified on the real vehicle test data, the correction accuracy reached 93.33%, significantly enhancing the robustness of the intelligent NVH assessment under interference.

Introduction: Botulinum toxin is widely used to treat upper facial wrinkles, and its efficacy is typically assessed through photographic comparisons and standardized scales. Artificial intelligence (AI) is increasingly being integrated into aesthetic dermatology for objective wrinkle evaluation. Objectives: This study aimed to compare human and AI-based assessments of pre- and posttreatment of upper facial wrinkles and evaluate their consistency and treatment effectiveness. Methods: A total of 228 individuals (204 females, 24 males) who received abobotulinum toxin for glabellar, forehead, and lateral canthal wrinkles were analyzed using pre- and posttreatment photographs. Wrinkles were assessed by four human raters using the 5-point Merz scale and Global Aesthetic Improvement Scale (GAIS). AI evaluations were conducted using Haut.AI Face Skin Metrics 2.0, a pre-trained machine learning platform. Results: AI had better error rates than humans for age prediction. The AI and human assessments showed high agreement for static and dynamic wrinkle evaluations (P<0.001). Posttreatment analysis indicated significant wrinkle reduction in both the human and AI assessments (P<0.001). Human assessment of GAIS scores was negatively correlated with wrinkle reduction (P<0.001). The treatment effects measured by AI and human raters showed a weak-to-moderate correlation. Conclusion: AI-based assessments align well with human evaluations and can detect posttreatment improvements. However, the treatment effect did not correlate well with human evaluations. AI can serve as an objective tool for evaluating botulinum toxin treatment outcomes and complementing human assessments. However, there is still a need for a gold standard method to evaluate aesthetic improvement and harmony.

Until 1995, patients with newly diagnosed germinoma received 40-60 Gy of radiation to the primary site with or without chemotherapy (regimen A). After 2000, treatment shifted to chemotherapy followed by 24 Gy of whole-ventricle radiation therapy (WVRT) (regimen B). This study compares long-term intelligence outcomes between the two treatment regimens. This retrospective analysis included 151 patients diagnosed with germinoma between 1983 and 2021. Intelligence was assessed using the Wechsler Adult Intelligence Scale (revised or 3rd edition) and the Wechsler Intelligence Scale for Children (3rd edition). Patient backgrounds were also collected. A total of 55 and 69 patients were treated with regimens A and B, respectively. The number of patients who underwent at least one longitudinal neurocognitive assessment was 35 and 29 for regimen A and 53 and 22 for regimen B, respectively. The median interval from initial treatment to the last neurocognitive assessment was 120 months. In the longitudinal intelligence assessments, the median intervals were 58 months from treatment to the first evaluation and 83 months from the first to the final assessment. Full-Scale Intelligence Quotient (FSIQ) scores declined in regimen A but were maintained in regimen B according to analysis of covariates and generalized linear mixed model analysis. Chemotherapy followed by 24 Gy of WVRT appears to be associated with a smaller decline in FSIQ over a long-term follow-up.

Affective computing has emerged as a pivotal field in human–computer interaction. Recognizing human emotions through electroencephalogram (EEG) signals can advance our understanding of cognition and support healthcare. This study introduces a novel subject-independent emotion recognition framework by integrating multiple EEG emotion databases (DEAP, MAHNOB HCI-Tagging, DREAMER, AMIGOS and REFED) into a unified dataset. EEG segments were transformed into feature vectors capturing statistical, spectral, and entropy-based measures. Standardized pre-processing, analysis of variance (ANOVA) F-test feature selection, and six machine learning models were applied to the extracted features. Classification models such as Decision Tree, Discriminant Analysis, Support Vector Machine (SVM), K-Nearest Neighbor (K-NN), Naive Bayes, and Artificial Neural Networks (ANN) were considered. Experimental results demonstrate that SVM achieved the best performance for arousal classification (70.43%), while ANN achieved the highest accuracy for valence classification (68.07%), with both models exhibiting strong generalization across subjects. The results highlight the feasibility of developing biomimetic brain–computer interface (BCI) systems for objective assessment of emotional intelligence and its cognitive underpinnings, enabling scalable applications in affective computing and adaptive human–machine interaction.

A model for anomaly detection in user behavior within corporate databases was presented, based on a hybrid LSTM-Autoencoder architecture. It is emphasized that the proposed approach integrates both structural and temporal behavioral factors, allowing effective detection of contextual and sequential deviations. For validation, real ERP-oriented data from the SALT (Sales Autocompletion Linked Tables) dataset were used, comprising more than 2.3 million records reflecting transactions, clients, and logistics processes. The data were aggregated into temporal windows of length m = 20 queries with 87 features, formalizing the dynamics of user activity. Training was conducted on a Tesla T4 GPU (16 GB) using the Adam optimizer with a learning rate of 1e−3, batch size 128, and 50 epochs, during which the loss function stabilized at MSE = 0.0023. The threshold value dynamically adapted to the current risk distribution, reducing false positives to 3.1%. The mean reconstruction error for normal windows was Lrec = 0.0017, while for anomalous windows it was 0.0079, providing more than a fourfold separation between clusters. The model achieved Precision = 0.946, Recall = 0.931, F1-score = 0.938, and AUC = 0.972, outperforming classical methods such as Isolation Forest and One-Class SVM by 7–15%. The results show that that the dynamic threshold mechanism θt enables the system to adapt its sensitivity to varying workloads, maintaining balance between accuracy and robustness. Experimental results confirm the model’s ability to distinguish between structural and behavioral anomalies, including sudden shifts in query types, actions inconsistent with user roles, and unusual geographic sources of access. Thus, the proposed method forms the basis for an intelligent real-time behavioral risk assessment system for corporate databases, capable of integration into existing DBMS environments through a middleware interface without compromising performance.