The article is devoted to the study of the impact of digital transformation and artificial intelligence on the formation of sustainable ecosystems in higher education. The paper discusses theoretical and practical aspects of the implementation of advanced digital technologies such as cloud services, big data, artificial intelligence and machine learning, which provide new opportunities for optimizing the educational process and management of educational institutions.

Artificial intelligence (Al) is becoming an integral part of everyday life, actively influencing many aspects of human activity. Al technologies are used to automate routine tasks, improve the quality of service and enhance convenience in various fields, such as medicine, education, transportation, finance and entertainment. For example, voice assistants, recommendation systems, smart homes and chatbots significantly simplify the performance of daily tasks. However, along with the benefits, Al raises questions related to ethics, privacy and data security. The impact of artificial intelligence on the labor market raises concerns about replacing human labor, while the rapid development of technology generates the need to adapt to new conditions. This topic emphasizes the importance of studying the benefits and risks associated with the implementation of Al, as well as developing strategics for its effective use to improve the quality of life and minimize possible threats. Keywords: Artificial intelligence (Al), machine learning, deep learning, automation, smart devices, robotics, data analysis, computer vision, chatbots, digital transformation, social networks and algorithms.

This article examines the role of algorithms in solving artificial intelligence (Al) problems and ways to increase their effectiveness. The main algorithmic approaches used in systems (Al) will be analyzed, as well as their advantages and limitations will be highlighted. The article uses advanced literature that will help to study the theoretical and practical aspects of artificial intelligence algorithms.

This study comprehensively analyzes palmprint databases as a fundamental resource for biometric identification systems. It provides a detailed exploration of the database creation process, their technical specifications, and application areas. Existing databases such as CASIA Palmprint, NEC Palm Database, and PolyU, Multispectral have also been examined.During the research, issues related to quality, privacy concerns, standardization challenges, and technical limitations were identified, with recommendations proposed to address them.

This work is dedicated to exploring the application of machine learning (ML) in transforming business processes within the digital economy. The study examines the potential of ML algorithms for automating management, forecasting key performance indicators (KPIs), and optimizing resource allocation. The article provides a detailed overview of theoretical foundations, the methodology for developing software solutions, and the results of experiments conducted on real-world data from logistics and e-commerce. Examples of using linear regression, random forest, gradient boosting, and neural networks are presented, demonstrating their effectiveness in enhancing productivity and reducing costs. The work emphasizes the strategic role of ML as a tool for achieving competitive advantages and suggests directions for further research in adapting these technologies to various industries.

The article explores approaches based on machine learning for the prevention of endocrine diseases in medicine, highlighting their advantages and prospects for application. The possibility of predicting early stages of diseases using data collection, analysis, and machine learning algorithms is examined.

The paper presents a comparative analysis of machine learning algorithms applied for the early diagnosis of oncological diseases. The paper examines algorithms such as Random Forest, XGBoost, AdaBoost, and others, tested on various clinical tasks including cervical, lung, and skin cancer. Special attention is given to ensemble methods, which demonstrated the highest accuracy, particularly the Random Forest algorithm. The study emphasizes the versatility of the methods, their adaptability to heterogeneous medical data, and their potential for developing intelligent clinical decision support systems.

This paper examines modem algorithms used for the intelligent analysis of speech signals and their scientific and practical significance. The development of artificial intelligence and machine learning technologies has expanded the capabilities of automatic speech signal processing, feature extraction, and recognition. The study analyzes modeling processes based on advanced methods such as MFCC, CNN, and RNN. It also explores algorithms used for spccch-to-tcxt conversion, speaker identification, and context understanding. The results may be applied in intelligent voice interfaces, security systems, and linguistic applications.

This article proposes a deep learning-based model for extracting key entities from texts and creating a knowledge base. The Long Short-Term Memory (LSTM) model is used for the Named Entity Recognition (NER) task. The data is preproccssed and converted into a digital format using tokenization and one-hot encoding. The model is trained and evaluated to extract various types of entities (e.g., person names, dates, and location names). Experimental results demonstrate the model’s effectiveness, and the impact of different parameters is analyzed.

This article analyzes the processes of analyzing and classifying textual data, considers the types of textual data, namely structured, unstructured and semi-structured data, and presents their characteristics. In addition, special attention is paid to the existing opportunities and problems in processing textual data in the Uzbek language. In particular, the achievements and shortcomings of analyzing textual data in the Uzbek language using the example of the "Tahrirchi" system are presented.

Nowadays, systems ensuring natural interaction between humans and machines are rapidly evolving. Among them, the task of identifying the user’s language holds particular importance. This article analyzes the problem of language identification (LID) based on speech signals, its application areas, challenges, and modem approaches. It compares traditional machine learning methods (GMM, SVM, i-vcctor) with deep neural network-based approaches (CNN, RNN, Transformer) for language recognition. Additionally, the paper discusses key evaluation metrics such as Accuracy, Precision, Fl-score, and Equal Error Rate (EER) for assessing system performance. Advanced methods for handling complex scenarios like code-switching and openset LID are reviewed, with a focus on practical perspectives for under-resourced languages like Uzbek. The results of the study provide a solid theoretical and practical foundation for developing multilingual interactive voice systems.

This paper focuses on the training data preparation stage of the CRISP-DM methodology in data mining. Because this stage takes up to 80% of the model development time, the steps of the training data preparation stage are explained with illustrations, approaches to training data in the model are presented, and considerations for machine learning engineers and data analysts are mentioned.