SARS-CoV-2 is a novel virus that has been affecting the global population by spreading rapidly and causing severe complications, which require prompt and elaborate emergency treatment. Automatic tools to diagnose COVID-19 could potentially be an important and useful aid. Radiologists and clinicians could potentially rely on interpretable AI technologies to address the diagnosis and monitoring of COVID-19 patients. This paper aims to provide a comprehensive analysis of the state-of-the-art deep learning techniques for COVID-19 classification. The previous studies are methodically evaluated, and a summary of the proposed convolutional neural network (CNN)-based classification approaches is presented. The reviewed papers have presented a variety of CNN models and architectures that were developed to provide an accurate and quick automatic tool to diagnose the COVID-19 virus based on presented CT scan or X-ray images. In this systematic review, we focused on the critical components of the deep learning approach, such as network architecture, model complexity, parameter optimization, explainability, and dataset/code availability. The literature search yielded a large number of studies over the past period of the virus spread, and we summarized their past efforts. State-of-the-art CNN architectures, with their strengths and weaknesses, are discussed with respect to diverse technical and clinical evaluation metrics to safely implement current AI studies in medical practice.

This paper presents a systematic evaluation of large language models (LLMs) and retrieval-augmented generation (RAG) approaches for question answering (QA) in the low-resource Kazakh language. We assess the performance of existing proprietary (GPT-4o, Gemini 2.5-flash) and open-source Kazakh-oriented models (KazLLM-8B, Sherkala-8B, Irbis-7B) across closed-book and RAG settings. Within a three-stage evaluation framework we benchmark retriever quality, examine LLM abilities such as knowledge-gap detection, external truth integration and context grounding, and measures gains from realistic end-to-end RAG pipelines. Our results show a clear pattern: proprietary models lead in closed-book QA, but RAG narrows the gap substantially. Under the Ideal RAG setting, KazLLM-8B improves from its closed-book baseline of 0.427 to reach answer correctness of 0.867, closely matching GPT-4o’s score of 0.869. In the end-to-end RAG setup, KazLLM-8B paired with Snowflake retriever achieved answer correctness up to 0.754, surpassing GPT-4o’s best score of 0.632. Despite improvements, RAG outcomes show an inconsistency: high retrieval metrics do not guarantee high QA system accuracy. The findings highlight the importance of retrievers and context grounding strategies in enabling open-source Kazakh models to deliver competitive QA performance in a low-resource setting.

This study is devoted to the transcription of human speech in the Kazakh language in dynamically changing conditions. It discusses key aspects related to the phonetic structure of the Kazakh language, technical considerations in collecting the transcribed audio corpus, and the use of deep neural networks for speech modeling. A high-quality decoded audio corpus was collected, containing 554 h of data, giving an idea of the frequencies of letters and syllables, as well as demographic parameters such as the gender, age, and region of residence of native speakers. The corpus contains a universal vocabulary and serves as a valuable resource for the development of modules related to speech. Machine learning experiments were conducted using the DeepSpeech2 model, which includes a sequence-to-sequence architecture with an encoder, decoder, and attention mechanism. To increase the reliability of the model, filters initialized with symbol-level embeddings were introduced to reduce the dependence on accurate positioning on object maps. The training process included simultaneous preparation of convolutional filters for spectrograms and symbolic objects. The proposed approach, using a combination of supervised and unsupervised learning methods, resulted in a 66.7% reduction in the weight of the model while maintaining relative accuracy. The evaluation on the test sample showed a 7.6% lower character error rate (CER) compared to existing models, demonstrating its most modern characteristics. The proposed architecture provides deployment on platforms with limited resources. Overall, this study presents a high-quality audio corpus, an improved speech recognition model, and promising results applicable to speech-related applications and languages beyond Kazakh.

Filament identification became a pivotal step in tackling fundamental problems in various fields of Astronomy. Nevertheless, existing filament identification algorithms are critically user-dependent and require individual parametrization. This study aimed to adapt the neural networks approach to elaborate on the best model for filament identification that would not require fine-tuning for a given astronomical map. First, we created training samples based on the most commonly used maps of the interstellar medium obtained by Planck and Herschel space telescopes and the atomic hydrogen all-sky survey HI4PI. We used the Rolling Hough Transform, a widely used algorithm for filament identification, to produce training outputs. In the next step, we trained different neural network models. We discovered that a combination of the Mask R-CNN and U-Net architecture is most appropriate for filament identification and determination of their orientation angles. We showed that neural network training might be performed efficiently on a relatively small training sample of only around 100 maps. Our approach eliminates the parametrization bias and facilitates filament identification and angle determination on large data sets.

This paper presents a comparative analysis of large pretrained multilingual models for question-answering (QA) systems, with a specific focus on their adaptation to the Kazakh language. The study evaluates models including mBERT, XLM-R, mT5, AYA, and GPT, which were tested on QA tasks using the Kazakh sKQuAD dataset. To enhance model performance, fine-tuning strategies such as adapter modules, data augmentation techniques (back-translation, paraphrasing), and hyperparameter optimization were applied. Specific adjustments to learning rates, batch sizes, and training epochs were made to boost accuracy and stability. Among the models tested, mT5 achieved the highest F1 score of 75.72%, showcasing robust generalization across diverse QA tasks. GPT-4-turbo closely followed with an F1 score of 73.33%, effectively managing complex Kazakh QA scenarios. In contrast, native Kazakh models like Kaz-RoBERTa showed improvements through fine-tuning but continued to lag behind larger multilingual models, underlining the need for additional Kazakh-specific training data and further architectural enhancements. Kazakh’s agglutinative morphology and the scarcity of high-quality training data present significant challenges for model adaptation. Adapter modules helped mitigate computational costs, allowing efficient fine-tuning in resource-constrained environments without significant performance loss. Data augmentation techniques, such as back-translation and paraphrasing, were instrumental in enriching the dataset, thereby enhancing model adaptability and robustness. This study underscores the importance of advanced fine-tuning and data augmentation strategies for QA systems tailored to low-resource languages like Kazakh. By addressing these challenges, this research aims to make AI technologies more inclusive and accessible, offering practical insights for improving natural language processing (NLP) capabilities in underrepresented languages. Ultimately, these findings contribute to bridging the gap between high-resource and low-resource language models, fostering a more equitable distribution of AI solutions across diverse linguistic contexts.

Background: Cardiovascular diseases (CVDs) remain the primary cause of global mortality, with arterial hypertension, ischemic heart disease (IHD), and cerebrovascular accident (CVA) forming a progressive continuum from early risk factors to severe outcomes. While numerous studies focus on isolated biomarkers, few integrate multidimensional visualization with artificial intelligence to reveal hidden, clinically relevant patterns. Methods: We conducted a comprehensive analysis of 106 patients using an integrated framework that combined clinical, biochemical, and lifestyle data. Parameters included renal function (glomerular filtration rate, cystatin C), inflammatory markers, lipid profile, enzymatic activity, and behavioral factors. After normalization and imputation, we applied correlation analysis, parallel coordinates visualization, t-distributed stochastic neighbor embedding (t-SNE) with k-means clustering, principal component analysis (PCA), and Random Forest modeling with SHAP (SHapley Additive exPlanations) interpretation. Bootstrap resampling was used to estimate 95% confidence intervals for mean absolute SHAP values, assessing feature stability. Results: Consistent patterns across outcomes revealed impaired renal function, reduced physical activity, and high hypertension prevalence in IHD and CVA. t-SNE clustering achieved complete separation of a high-risk group (100% CVD-positive) from a predominantly low-risk group (7.8% CVD rate), demonstrating unsupervised validation of biomarker discriminative power. PCA confirmed multidimensional structure, while Random Forest identified renal function, hypertension status, and physical activity as dominant predictors, achieving robust performance (Accuracy 0.818; AUC-ROC 0.854). SHAP analysis identified arterial hypertension, BMI, and physical inactivity as dominant predictors, complemented by renal biomarkers (GFR, cystatin) and NT-proBNP. Conclusions: This study pioneers the integration of multidimensional visualization and AI-driven analysis for CVD risk profiling, enabling interpretable, data-driven identification of high- and low-risk clusters. Despite the limited single-center cohort (n = 106) and cross-sectional design, the findings highlight the potential of interpretable models for precision prevention and transparent decision support in cardiovascular aging research.