Machine learning is the process of developing and applying predictive algorithms to predicting future outcomes using available training data. Machine learning is a corner-stone for Data science, Big Data analytics, Robotics, Natural language processing and AI in general. Course introduces students to supervised and unsupervised machine learning concepts and algorithms from basic classification to Deep Learning with artificial neural networks. As computers and GPUs become more powerful, Deep Neural Networks are gradually taking over from simpler Machine Learning methods.

Course should give insight into the principles of design of parallel algorithms and overview of some standard parallel algorithms. The course will teach basic theoretical models of parallel computing and performance and complexity metrics of parallel algorithms. The course will also provide overview on the current industrial parallel computing systems and programming languages.

This course is intended to provide an overview of modern applications of machine learning and develop practical skills in using deep neural networks for common machine learning tasks. The course provides an introduction into artificial neural network based models, as well as an introduction to existing API frameworks for training such models. Previous knowledge regarding machine learning is not expected. The practical assignments will be developed in Python programming language.

The course introduces students to the basic principles that should be acquired to mathematicians, statisticians, economists and financiers and others in order to use software packages R, SPSS and MATLAB for their scientific research and practical work. An overview is given about the appropriate application features and options of the software packages R, SPSS and MATLAB focusing primarily on the students’ needs to work out different practice, bachelor and other scientific works.

The course focuses on numerical methods of statistical physics. Classes includes lectures about theoretical aspects of Monte-Carlo and molecular dynamics methods, and practical work on computer to solve some statistical physics problems.

As a result of this study course the student acquires basic knowledge about multiphysical modeling, basic steps of problem-solving, verification and analysis of results. The acquired competencies allow to  hoose a physical model suitable for the description of the physical  process, to explain physical processes on the basis of obtained results and to give recommendations for optimization of the physical process.

The aim of the course is to acquaint students with creation of mathematical models for description of complex physical processes and methods for solving the corresponding problems of mathematical physics, using different program tools – both commercial and specialised mathematical modelling programs.

The course focuses on two methods designed for calculation of physical fields: finite element method (FEM) and boundary element method (BEM). Students learn basics of both methods.

Theoretical lectures are complemented by laboratory work sessions, where students acquire practical skills in the use of the appropriate software.

In addition to the theoretical background students acquire numerical aspects of realization of these methods in computer codes.
Open source software „freefem++” and "gmsh" are used as basic tools to learn FEM and BEM.

The goal of the course is to introduce the students to the basics of data analysis and machine learning methods as an additional tool for finding patterns in data and issuing predictions, by working with data from various physical systems.

The tasks of the course are to introduce the students to the elements of data analysis - cleaning, analysis and visualization, based on data from real physical systems; to apply the machine learning algorithms by mathematically modelling various physical systems.