Students to Develop Algorithms for the Computers of the Future

Lev Shchur, Head of the Programme 'Supercomputer Simulations in Science and Engineering'. Photo: HSE/Mikhail Dmitriev

This year, the new Master's programme ‘Supercomputer Simulations in Science and Engineering’ will be launched at MIEM HSE. Graduates will be experts in the interdisciplinary field of computer technology, natural sciences and engineering sciences.

The programme is designed for those who wish to develop, apply and explore modern methods of supercomputer simulations and processing of big data in science and engineering.

‘A typical modern computer has 4 to 8 kernels, but it surpasses the speed of a supercomputer from 30 years ago,’ explains the Head of the new master's programme, Lev Shchur. ‘Computers of the future will consist of millions of kernels, and hence the fundamental task of modern scientists is to develop algorithms for computers that will become a reality in the next 30, 20 or even 10 years. This is what we will be teaching the students in our Master's programme.’

On the Shoulders of the Titans

The programme ‘Supercomputer Simulations in Science and Engineering’ was developed by MIEM with the help of colleagues from leading computer and scientific centers in Russia. This includes the Keldysh Institute of Applied Mathematics (Russian Academy of Sciences), the Landau Institute for Theoretical Physics in Chernogolovka and the Kurchatov Institute, where scientists apply methods of deep machine learning for the processing of experimental results.

This is the first programme of its kind in Russia. In Europe, there is a joint Master's Degree offered by the technical universities of Stockholm, Delft and Berlin. While the idea to create a programme in Russia came about from talks with foreign colleagues, it is by no means a copy of the Western model.

‘We stand on the shoulders of Titans,’ Professor Shchur said. ‘The programme is founded on the achievements of fundamental Russian science: in mathematics - the school of Professor Nikolai Luzin at Moscow State University (Kolmogorov's teacher) and in physics - the school of Abram Ioffe, the founder of St Petersburg's PhysTech, (he taught Landau and Kapitsa). We are adding to the mix cutting-edge research in the field of modern information technologies and new materials.’

Professor Lev Shchur is a leading researcher at the Landau Institute for Theoretical Physics in Chernogolovka and his research interests are at the intersection of theoretical physics and computer software. In 2017, he became the only Russian scientist to be recognised as a distinguished member of the American Physical Society.

The Professor’s young colleagues will be teaching in the Master’s programme, many of whom have either worked or studied abroad and are active members of the international scientific community. Among them is Yury Budkov, a specialist in materials science and an employee at the Institute of Solution Chemistry (Russian Academy of Sciences). He is joined by specialists in statistical and computer physics, Lev Barash, and Evgeni Burovski, who has a PhD from the University of Massachusetts.

There is a particular focus on statistical physics as it now plays a key role in all technical sciences. Therefore, a modern view of machine learning from the viewpoint of statistical physics is extremely important. Although the curriculum of the Master's programme has already been drawn up, new courses can be added to it depending on where modern science will head. Foreign scientists may come to HSE to teach a course, if it is not possible to find the right specialist in Russia.

Students will be able to carry out practical work at partner institutions of the Russian Academy of Sciences, and at international research centres, as most graduates wish to pursue a career in academia. Those who do not go on to work in science will also find work very easily, according to Professor Shchur. Specialists in supercomputer modelling who can create a new task and put it in front of a computer are in demand by high-tech companies both in Russia and abroad. Regardless of whether they end up at a research institute, IT company or IT department, graduates will use the same techniques and algorithms that they studied at the university.

Who Is the Programme Intended for

The Master's programme is designed for graduates of technical universities in areas such as applied mathematics, computer science, physics and software engineering.

Applicants should ideally meet three requirements, Lev Shchur believes, and this is what the entry interview is for.

The first requirement motivation. ‘A person should be motivated to seek new knowledge, that is, ask questions, quickly complete assignments and communicate with colleagues’, he explains.

The second requirement is the ability to reason logically and to identify new ideas. ‘It's enough to talk with a person for five minutes to find out if he or she knows how to do this,’ Professor Shchur believes. ‘We need students who are capable of productive thinking. Of course, we also check basic knowledge, but this is not as important. If a person has not studied mathematics in depth, we will provide this training’.

The third requirement is knowledge of the English language, as the programme ‘Supercomputer Simulations in Science and Engineering’ is bilingual. All teachers are fluent in English, both spoken and written, and the same is required of students.

Due to the fact that students come to the programme with very different backgrounds, experience and training, bridging courses in mathematics, science and engineering will be offered. Each student will be assigned tasks based on their capabilities. A special feature of the programme is constant communication between teachers and students. Teachers treat science as a collective undertaking, and not a task for one person alone. Hence, there will be regular scientific seminars where teachers and students can exchange opinions. All students will participate in research projects, financed, in particular, by the RNF, RFBR and the Higher School of Economics.