The program draws on the expertise of some top international (Stanford University) and Russian (Yandex School of Data Analysis, Faculty of Computer Science at Higher School of Economics) study programs. It delivers a competence level that requires no further training with IT companies.
Key principles of the program:
Academic
Trajectory
Introduction
to Artificial
Intelligence
Machine
Learning
A basic induction course that clarifies the central concepts of all key chapters. The course is based on CS 221 Artificial Intelligence: Principles and Techniques, a required course for Standford students having picked Artificial Intelligence as their major.
This is a special course in the 2021 spring semester. From academic year 2021/2022 onwards, it will be taught as a cross-faculty course. The way the course is structured, lectures on every topic are given by experts in the respective fields.
The course consists of the following chapters:
The Machine Learning course has multiple focus points, namely, the Python programming language, specialist libraries (numpy, pandas, scikit-learn), and programming environments (Jupyternotebook). The course teaches key problems of machine learning (with labeled/unlabeled data, with reinforcement) and the methods of solving them (from the classical linear, metric and Bayesian to very up-to-date ensemble methods).
The course discusses main problems arising from the practice of precedent-based learning (machine learning) methods. A brief overview is given of the existing recognition and regression analysis methods. Students get familiarized with the methods of accuracy assessment on a universal set (generalization ability). The module discusses the different ways of raising the generalization ability of the machine learning methods. On top of the classical subjects, the course dwells on complex network analysis, as well as data and model interpretation methods.
Bachelor's
select 4 courses
Master's
select 6 courses
The course takes a look at the existing practices of using artificial intelligence methods in management theory. The focus is on three key areas: evolutionary optimization, fuzzy logic, and neural nets. The course also dwells on how the methods are realized within the computer mathematics system MATLAB and in open-source software packages.
Anticipated course outputs: you learn to solve optimization and structural synthesis (parsing of genetic algorithms) problems like a pro, you gain hands-on experience using the MATLAB application package, and you get to use the baseline AI methods of management theory in your own research.
An in-depth course for analytical processing of big and small datasets, an important step towards a holistic command of Python and other programming languages.
Anticipated course outputs: you get introduced to the specifics of linear, integer-valued, non-linear, dynamic programming, you work through mathematical analysis problems (maxflow, zero sum games, extreme solutions, and so on), you learn the mathematical take on optimization (including complex NP tasks), become familiar with audio and video data encoding (using the Lagrangian multiplier method), and obtain powerful formulas to use with management theory.
This course delivers solid knowledge about neural nets and artificial intelligence. The course includes a theory-to-practice module where you get the opportunity to generate custom analytical research for your paper.
Anticipated course outputs: you become competent in the use of different types of analysis (descriptive, correlation, regression, component), learn about clusterization and intelligent analysis (data compression, discovery of novelty), study different methods of solving DataTrees, and familiarize yourself with limited search algorithms for neural nets.
The use of software applications will help to consolidate new knowledge about the contemporary programming tools, including C++ based solutions, automatic differentiation, and hierarchical model assessment.
Anticipated course outputs: you develop skills in analysis and planning for statistical research of varying complexity, develop pro-level command of known applications for big and small dataset analysis, you get to practice statistical methods while working with hypotheses (test, development, assessment, and so on), you get the hang of the key data mining processes: prognostic modeling, solution management, etc., you find out what is different about word processors, and you learn to work with AI (voice modules and voice synthesizers).
The course deals with automatic stylizing of images and a range of transformative tasks (neural style transfer) in shaping the new style of one or several images. The methods, which are the subject of this course, are widely used in mobile apps and editors, in animated cartoons and teasers, in films, video games, simulators, augmented reality tools, and generally wherever accurate rendering of graphic special effects is of the essence.
Anticipated course outputs: you become adroit at multi-style transformation of images, you learn to work with middleware and source codes for detailed study of neural nets, and develop practical skills in working with neural net architectures, classification and stylizing of graphic images.
The course presents mathematical methods for automatic text analysis, processing, and data mining. The course builds on the fundamentals of Natural Language Processing (NLP) and AI programming (robotic software, virtual assistants).
Anticipated course outputs: you gain familiarity with the engineering aspects (cases with tasks from front-tier IT corporations), learn about the use of text segmenting and summarizing, learn to process texts in the natural language (inclusive of parsing the word sequences), you work through real-life cases (up-to-date specimens of work with neural nets and artificial intelligence), and you get the hang of neural net and combined tagging models.
Reinforcement Learning is a central machine learning paradigm characterized by the prevalence of experimental learning by preset behavioral metrics for more advanced interaction with artificial intelligence. The course involves practical honing of system algorithm skills.
Anticipated course outputs: you learn to fast-track solutions to engineering problems (classic tasks posed to developers by big-name IT corporations), you work through real-life cases (you learn from up-to-date specimens of work with neural nets and artificial intelligence), and you get new, inspiring ideas for your research.
The course addresses the basics of the Bayesian approach and its applications in hierarchical, mixed, linear, and generalized linear models, as well as the different decision-making styles involved.
Anticipated course outputs: you get introduced to the nuances of style modeling with Bayesian hierarchical modeling methods, you learn everything about spam filters and the percent of likelihood that your letter will be read (the log likelihood function logarithm, the Bayes Classifier), you learn to minimize the probability of common error using the Bayesian decision-making theory, and you pick the Bayesian approach best suited to the purposes of your own research.
The course addresses the different ways to formulate management problems for nonlinear dynamic entities. The course expounds the methods of inverse dynamic problems, feedback linearization, high gain factor, variable structure systems, Lyapunov functions, and coordinate-to-operator feedback in order to resolve stabilization issues in nonlinear dynamic systems.
Anticipated course outputs: you master the practical application of the all-important inverse dynamics principles, learn to apply Lyapunov's quadratic functions in systems with latency, and you gain a mathematical toolkit you can put to work in your own projects.
The course leads to conversance with the theoretical and practical architecture of machine learning processes: representation/presentation.
Anticipated course outputs: you receive solid grounding in neural simulation, data transformation, and selection with additional competence in quality analytical decision-making for handling standard deviations; you also enhance your knowledge of genetic algorithms and purpose-specific hyperparameters (optimization).
Entry requirements: basic level Python, scholarly interest in intelligent data mining, analytical thinking (the course includes problems requiring complicated calculations and analysis).
The extended course introduces students to the theory and practice of building the machine learning processes: representation/presentation. The course is centered around a theory-to-practice module expounding the fundamentals of computer vision in different programming applications.
Anticipated course outputs: you receive solid grounding in neural simulation, data transformation, and selection with additional competence in quality analytical decision-making for handling standard deviations; you also enhance your knowledge of genetic algorithms and purpose-specific hyperparameters (optimization).
Entry requirements: basic level Python, scholarly interest in AI.
This course is designed for specialists motivated to discover the intelligent video processing techniques embedded in the works of the popular video hosts and platforms, as well as in the widespread mobile apps. Your learning experience will assist you in the development of a unique science project.
Anticipated course outputs: you develop specialist skills in intelligent video processing and data mining, you get the hang of mathematical and cloud computation, and you perfect your skills in gathering and mining digital data, including the standard set of operations: reading, processing, and subsequent coding of video recordings (DV, AVI, MPEG, MOV, DVD, FLV, a.o.).
The course is designed for students who are prepared to take on the challenges of management and are conversant with the basic terminology of probabilistic models.
This course will supplement your core knowledge of probabilistic models with applied knowledge of the risk accounting methods used in management problem-solving.
Anticipated course outputs: you develop competence in different probabilistic models for accurate forecasting, you learn the designer probabilistic modeling methods for research (for instance, the Monte Carlo method and Markov additive chains), and learn how to cull the suitable probabilistic models for tests, team or individual assignments, and homework.
This course targets specialists with a stake in gaining high-quality practical knowledge of programming, artificial intelligence, and neural network-based digital programs; the course includes a theory-to-practice module where you get to generate a body of research under the tutelage of experienced machine learning practitioners.
Anticipated course outputs: you learn to interpret basic knowledge of binary compounds and neural nets (work with program algorithms), you master special aspects of machine learning, and learn to process missed data (the Bayesian estimation, the Laplace approximation, factorization, etc.).
The use of Bayesian neural methods helps to develop competence in the methods of Bayesian subjective estimation.
Anticipated course outputs: you gain supplemental expertise in machine learning and AI, you will be able to complement your research with your own technical ideas and qualified analysis, and you become familiar with the contemporary trends in deep neural nets/belief nets.
Students are advised to complete a course in Bayesian methods of machine learning before taking this course.
Key requirement for entrants: basic Python.
This course leads to generating highly accurate research and data for the development of mobile apps and other software (which may become highlights in the history of your faculty and Lomonosov Moscow State University).
Anticipated course outputs: you gain new experience, learning to invert negative feedback to positive (by parsing the frequently made programming mistakes), you analyze specimen feedback in a digital environment, you discover some technical life-hacks for the avoidance of classic errors, you complete practical assignments involving cases of stability and transitional processes inside operating systems (the phenomena that engineers and developers in major IT corporations get to deal with) for your own research.
An enhanced course focused on the technical aspects of machine learning involving graphic and textual data mining, and study of the main trends in machine learning. The course includes an overview of algorithms employed in ML, neural nets, and artificial intelligence (AI).
Anticipated course outputs: you get to search for high-quality solutions to complete applied tasks across different levels of complexity, including unique development in forward-looking fields of science (AI) for your own research, you will be enabled to map out the right trajectories towards the acquisition of singular professional experience in deep study, Data Science, and neural nets.
Research
support
The program draws on the expertise of some top international (Stanford University) and Russian (Yandex School of Data Analysis, Faculty of Computer Science at Higher School of Economics) study programs. It delivers a competence level that requires no further training with IT companies.
Course Leaders:
Vasily Fomichev
Vasily is a Doctor of Physics and Mathematics, Professor and Deputy Dean of the Faculty of Computational Mathematics and Cybernetics, and Head of the Department of Nonlinear Dynamic Systems and Control Processes at Lomonosov Moscow State University
Program Instructors:
Sergei Berezin
Education
Mathematical Engineer
Moscow Institute of Physics and Technology
Areas of Expertise
Text mining, data mining, probabilistic topic modeling, transactional data mining, biomedical data mining, electrocardiogram-based diagnostics
Dmitry Vatolin
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Intelligent video processing
Dmitry Vetrov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Bayesian machine learning methods, generalization assessment of image recognition algorithms, statistical relational learning methods, graphic models, Markov chain Monte Carlo methods, image processing, mathematical methods of neuroimaging, statistical analysis of cerebral activity, pattern mining in discrete time series
Konstantin Vorontsov
Education
Mathematical Engineer
Moscow Institute of Physics and Technology
Areas of Expertise
Text mining, data mining, probabilistic topic modeling, transactional data mining, biomedical data mining, electrocardiogram-based diagnostics
Oleg Goncharov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Nonlinear dynamic system control problems
Sergei Dukanov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Sound processing
Alexander Dyakonov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Machine learning and applied data mining tasks, algebraic approach and interpolation theory, discrete mathematics
Vladimir Zakharov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Mathematical logic, computational complexity theory, distributed computing models, formal languages, mathematical fundamentals of cryptography
Viktor Kitov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Machine learning, classification and regression methods, feature conversion, prediction by algorithm patterns
Anton Konushin
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Computer vision
Viktor Korolev
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Probability theory and mathematical statistics
Valery Kostenko
Education
Automation and Computers
Taganrog Institute of Wireless Engineering
Areas of Expertise
Computational synthesis and planning in real-time systems, methods and tools of computational planning in data centers and cloud platforms, combinatory optimization, machine learning, federated learning
Dmitry Kropotov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Bayesian machine learning, learning and rendering methods in graphic models, optimization methods in machine learning, applied intelligent data mining
Natalia Lukashevich
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Automatic processing of text, ontology
Viktor Malyshko
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Artificial intelligence, object-oriented methods of analysis and design, software engineering, knowledge engineering
Mikhail Petrovsky
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
System programming, intelligent data mining, machine learning methods, decision-making support systems, cybersecurity
Alexander Rogovsky
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Artificial intelligence methods in management theory
Alexander Ryzhov
Education
Aeromechanics and Aircraft
Moscow Institute of Physics and Technology
Areas of Expertise
Fuzzy mathematics
Andrei Fursov
Education
Mathematics
Lomonosov Moscow State University
Areas of Expertise
Stability theory, management theory, stabilization theory, changeover systems
Mikhail Tselishchev
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Bayesian methods, Markov models
Oleg Shestakov
Education
Applied Mathematics and Information Science
Lomonosov Moscow State University
Areas of Expertise
Theories of stochastic tomography and wavelet analysis
The Introduction to Artificial Intelligence course is currently available to all students of the Faculty of Computational Mathematics and Cybernetics of Lomonosov Moscow State University, from the second year onwards. This subject area attractsa great deal of interest, so anyonewishing to sign up for the course is welcome to do so.
The first track started class online (viaZOOM) on February 8, 2021. There is a dedicated platform where all lecture material is conveniently available to students.
To sign up for the course:
The Introduction to Artificial Intelligence course is currently available to all students of the Faculty of Computational Mathematics and Cybernetics of Lomonosov Moscow State University, from the second year onwards. This subject area attractsa great deal of interest, so anyonewishing to sign up for the course is welcome to do so.
The first track started class online (viaZOOM) on February 8, 2021. There is a dedicated platform where all lecture material is conveniently available to students.
Admission dates for the 2022 program will be posted later.
