Microsoft azure-da süni intellektTopics

In the subject, the integration of artificial intelligence models, which require a lot of computing power and resources, into cloud environments (especially the microsoft azure platform), the role of cloud technologies in the development of ai, as well as real application scenarios of artificial intelligence in various industries such as manufacturing, finance, medicine, retail and logistics are studied.

In the subject, the mechanisms of integration of artificial intelligence models into the microsoft azure cloud environment, cloud structures used for storing large volumes of data (azure blob storage, azure data lake), as well as data management, cleaning and security standards for effective training of AI models are studied.

The subject studies the history of artificial intelligence, its evolutionary stages, narrow and general artificial intelligence concepts, as well as the fundamental differences between artificial intelligence, machine learning and deep learning, and the role of AI in modern cloud technologies.

The subject studies the main components that make up the architecture of artificial intelligence systems (data collection modules, computing infrastructure, machine learning models and application interfaces), as well as the general working mechanism through which these systems go from processing input data to making decisions and forming output results.

The subject explores the life cycle of a machine learning project from inception to deployment—problem definition, data collection, cleaning and preprocessing, feature engineering, model training, evaluation, and production.

Methods of raw data cleaning and structuring, including missing values ​​filling, outlier analysis, data normalization, and coding of categorical variables are studied in the subject to improve the accuracy of machine learning models.

In order to increase the predictive ability and effectiveness of machine learning models, methods of creating new meaningful features from existing data are studied, including feature selection, dimensionality reduction, and the principles of extracting new features from text and time series.

In the subject, the main types of machine learning models divided according to the principles of knowledge acquisition from data - tutored learning, untutored learning, semi-tutored learning and reinforced learning methods, as well as the application areas of each method (regression, classification, clustering) are studied.

The subject studies traditional algorithms that form the basis of machine learning – linear regression for predicting dependencies, logistic regression for classification problems, and decision trees. also, the concept of error matrix and the main evaluation metrics used to measure the quality of the built models are analyzed.

The subject studies the working principles of deep learning algorithms, which are complex architectures of artificial neural networks. students are introduced to the basic structure of convolutional neural networks for visual data processing and long-short-term memory networks for working with time series.

Optimization techniques applied to improve the accuracy and efficiency of machine learning models are studied in the topic. it includes hyperparameter tuning mechanisms to determine the model's best-performing structural adjustments, as well as tuning methods to prevent model overlearning.

The subject studies the principles of making trained machine learning models available to users and other services in real time (model serving). students get acquainted with the mechanisms of building an application programming interface (api endpoint) and the basics of microservice architecture to ensure the integration of the model with external systems.

The subject studies the principles of protecting artificial intelligence systems and especially large databases stored in the cloud environment in accordance with cyber security standards. students get familiar with ensuring the privacy of sensitive data, regulating role-based access rights of users and services accessing the system, and security mechanisms of cloud infrastructure.

In this topic, the essence of nature and its interaction with man are studied from a philosophical point of view. environmental problems and principles of sustainable development are discussed. philosophical foundations of a responsible attitude to nature are explained.

The subject provides a general review and synthesis of all artificial intelligence, machine learning, deep learning and cloud integration concepts covered during the semester. students orally defend the architecture, model choices and results of the applied projects they have prepared in groups in front of the audience.