Ivan Arzhantsev on the Future of Computer Science Education in Russia
Interviewed by Peter Birgir, Slon
Jun 26, 2014
The Higher School of Economics (HSE) and Yandex have joined forces to open a new Faculty of Computer Science. Led by Ivan Arzhantsev, the new faculty combines software engineering, applied mathematics and computer science. Arzhantsev, who also directs a group of online academic programs for Yandex, spoke to Slon about the present and future of computer science in Russia.
Yandex has said that they want the HSE Faculty of Computer Science — the first in Russia — to be a world-class campus. Are you breaking ground on new territory or are there already Russian centers of competence in this area?
Currently, there is no obvious leader in the field and it is unlikely that we will be able to create a single center of competence as the scientific community will never be so rigidly structured. However, the Yandex School of Data Analysis (SDA) has, in recent years, begun to produce the world’s best computer science experts alongside a number of other educational institutions.
About 300 students have graduated from the Yandex program so far and another 100 graduate every year — almost as many as from Stanford. The employment of new graduates at Yandex was never an end in itself; developers from SDA go on to work for many well-known organizations. The new department at HSE appears to meet the demand for such specialists on a larger scale. We expect 180 scholarship students will enter the bachelor’s program at HSE and 60 more will enter the master’s program.
Which other schools focus on computer science?
First of all there is the Faculty of Computational Mathematics and Cybernetics at Moscow State University; there are several faculties at the Moscow Institute of Physics and Technology — in particular their faculty for innovation and high tech — and at the St. Petersburg State University. The St. Petersburg University of Information Technologies Mechanics and Optics and the Computer Science Center in St. Petersburg, one of the organizers of which was SDA, also have strong computer science faculties. For all of them computer science is no longer an experimental direction but rather part of a regular and well-defined strategy.
The only problem is that the number of graduates is still too small to satisfy the market and the community cannot cope. There are even fewer people who are willing to simultaneously run their own projects, conduct research and teach at the same time. Such people tend to work in many places at once and are in great demand. The task of the new HSE faculty is to attract more such people and cultivate a new generation of computer scientists.
What does computer science mean in the Russian context? Do the scientific and corporate communities share a common understanding of computer science?
In the West, the term is more or less understood. We have a bit more complicated definition due to the way things have developed here but it includes related and applied mathematics, programming and information technology.
Ideally, a graduate should posses a complete set of professional programming skills and a deep knowledge of applied mathematics. Then it does not matter which field of knowledge is covered by the term but rather the areas in which they will be used. This includes the analysis and processing of large data, information retrieval, machine learning, computer vision, computational linguistics and machine translation.
These are fields in which Yandex is well-versed. Does that mean that there won’t be a comprehensive computer-science program at HSE?
We came up with 20 subjects before looking at how many of them Yandex could confidently cover. It became clear that Yandex could only oversee four or five of them. Then we started looking for partners who could join us. As a result, we have covered about ten of the original subjects in concert with our partners. Not all of them are interesting for Yandex, but Yandex is just one of several potential employers. Our goal was not to establish a corporate department.
It is understandable why machine learning is one of the subjects in which Yandex can provide competitive education that can compete on a global level. Computer vision technology is also used by Yandex services and SDA has a deep background in this.
Several years ago, a team that deals exclusively with research in the areas of information retrieval was founded at Yandex and has participated in leading conferences and has had its results widely published. This is our strong point.
In the field of academic computer science, timely publication is very important. In other scientific fields peer-reviewed journals are a crucial element of the system. A measure of success for mathematicians, for example, is publication in these journals. There are also journals for computer science, of course, but the pace of publication is very different. Results need to be published immediately; otherwise they would be out of date. Therefore, conferences play an important role in the selection, review and publication of new information.
Will the new faculty focus on training specialists in applied fields rather than researchers?
The faculty will prepare developers and researchers. In their first two years, students follow core courses before choosing one of several paths: Software engineering, data analysis, big data, complex systems modeling techniques, analysis and decision-making, intelligent systems or theoretical computer science.
Only the last is designed for students who want to conduct scientific research. We will provide both alternatives. The majority, which is understandable, will choose a career in business and will go on to become developers, analysts and managers. But if at least 10 percent of graduates decide to conduct research, either as part of the HSE faculty or not, it will be very good.
Beyond fact that Yandex is behind the faculty and that HSE has extensive experience delivering top-quality education, what else will distinguish the new faculty from the rest of the education market in computer science?
There are several features that I think will help us stand out. HSE is now conducting bold experiments to reduce the proportion of theory-based class time. This is a complex question and there has been constant debate about this in academic circles. It was finally decided that it was time to make a bold move. Students at the faculty will do a lot of independent work, which will account for about 40 percent of their time. Of course, it will be rigidly structured by setting tasks and monitoring their completion.
The second important feature is a focus on design and practical work; projects will be offered by faculty departments, research institutes and commercial enterprises. Students, in turn, will be able to choose those in which they are most interested and gain valuable experience working in teams. This will cover not only programming because our task is to offer students interesting projects across the board with plenty of choice.
We are trying to combine three separate traditions. First, there is the tradition inherited from Russian mathematical education, which is justly celebrated. The second is the use of corporate experience from the larger IT firms such as Yandex, which are all well aware of what subjects should be taught. Finally, we will use the tradition of programming competitions. Russian programmers regularly win international championships and for many the experience of participating in such tournaments is comparable to a second higher education. The new faculty boasts a number of winners from such programming competitions.
So, first of all there will be a reliance on the traditions of Russian technical education?
Not at all. We are situating ourselves within a global context. Stanford University, MIT and EPFL are all clear benchmarks for us. The HSE master’s and undergraduate programs are largely the result of following European and American educational practices. But it is important to combine the Western approaches to the formation of educational programs using the expertise of our leading universities and mathematics schools. Our graduates have always been known for their wide-ranging knowledge and we plan to maintain this tradition.
Yandex has said that they want the HSE Faculty of Computer Science — the first in Russia — to be a world-class campus. Are you breaking ground on new territory or are there already Russian centers of competence in this area?
Currently, there is no obvious leader in the field and it is unlikely that we will be able to create a single center of competence as the scientific community will never be so rigidly structured. However, the Yandex School of Data Analysis (SDA) has, in recent years, begun to produce the world’s best computer science experts alongside a number of other educational institutions.
About 300 students have graduated from the Yandex program so far and another 100 graduate every year — almost as many as from Stanford. The employment of new graduates at Yandex was never an end in itself; developers from SDA go on to work for many well-known organizations. The new department at HSE appears to meet the demand for such specialists on a larger scale. We expect 180 scholarship students will enter the bachelor’s program at HSE and 60 more will enter the master’s program.
Which other schools focus on computer science?
First of all there is the Faculty of Computational Mathematics and Cybernetics at Moscow State University; there are several faculties at the Moscow Institute of Physics and Technology — in particular their faculty for innovation and high tech — and at the St. Petersburg State University. The St. Petersburg University of Information Technologies Mechanics and Optics and the Computer Science Center in St. Petersburg, one of the organizers of which was SDA, also have strong computer science faculties. For all of them computer science is no longer an experimental direction but rather part of a regular and well-defined strategy.
The only problem is that the number of graduates is still too small to satisfy the market and the community cannot cope. There are even fewer people who are willing to simultaneously run their own projects, conduct research and teach at the same time. Such people tend to work in many places at once and are in great demand. The task of the new HSE faculty is to attract more such people and cultivate a new generation of computer scientists.
What does computer science mean in the Russian context? Do the scientific and corporate communities share a common understanding of computer science?
In the West, the term is more or less understood. We have a bit more complicated definition due to the way things have developed here but it includes related and applied mathematics, programming and information technology.
Ideally, a graduate should posses a complete set of professional programming skills and a deep knowledge of applied mathematics. Then it does not matter which field of knowledge is covered by the term but rather the areas in which they will be used. This includes the analysis and processing of large data, information retrieval, machine learning, computer vision, computational linguistics and machine translation.
These are fields in which Yandex is well-versed. Does that mean that there won’t be a comprehensive computer-science program at HSE?
We came up with 20 subjects before looking at how many of them Yandex could confidently cover. It became clear that Yandex could only oversee four or five of them. Then we started looking for partners who could join us. As a result, we have covered about ten of the original subjects in concert with our partners. Not all of them are interesting for Yandex, but Yandex is just one of several potential employers. Our goal was not to establish a corporate department.
It is understandable why machine learning is one of the subjects in which Yandex can provide competitive education that can compete on a global level. Computer vision technology is also used by Yandex services and SDA has a deep background in this.
Several years ago, a team that deals exclusively with research in the areas of information retrieval was founded at Yandex and has participated in leading conferences and has had its results widely published. This is our strong point.
In the field of academic computer science, timely publication is very important. In other scientific fields peer-reviewed journals are a crucial element of the system. A measure of success for mathematicians, for example, is publication in these journals. There are also journals for computer science, of course, but the pace of publication is very different. Results need to be published immediately; otherwise they would be out of date. Therefore, conferences play an important role in the selection, review and publication of new information.
Will the new faculty focus on training specialists in applied fields rather than researchers?
The faculty will prepare developers and researchers. In their first two years, students follow core courses before choosing one of several paths: Software engineering, data analysis, big data, complex systems modeling techniques, analysis and decision-making, intelligent systems or theoretical computer science.
Only the last is designed for students who want to conduct scientific research. We will provide both alternatives. The majority, which is understandable, will choose a career in business and will go on to become developers, analysts and managers. But if at least 10 percent of graduates decide to conduct research, either as part of the HSE faculty or not, it will be very good.
Beyond fact that Yandex is behind the faculty and that HSE has extensive experience delivering top-quality education, what else will distinguish the new faculty from the rest of the education market in computer science?
There are several features that I think will help us stand out. HSE is now conducting bold experiments to reduce the proportion of theory-based class time. This is a complex question and there has been constant debate about this in academic circles. It was finally decided that it was time to make a bold move. Students at the faculty will do a lot of independent work, which will account for about 40 percent of their time. Of course, it will be rigidly structured by setting tasks and monitoring their completion.
The second important feature is a focus on design and practical work; projects will be offered by faculty departments, research institutes and commercial enterprises. Students, in turn, will be able to choose those in which they are most interested and gain valuable experience working in teams. This will cover not only programming because our task is to offer students interesting projects across the board with plenty of choice.
We are trying to combine three separate traditions. First, there is the tradition inherited from Russian mathematical education, which is justly celebrated. The second is the use of corporate experience from the larger IT firms such as Yandex, which are all well aware of what subjects should be taught. Finally, we will use the tradition of programming competitions. Russian programmers regularly win international championships and for many the experience of participating in such tournaments is comparable to a second higher education. The new faculty boasts a number of winners from such programming competitions.
So, first of all there will be a reliance on the traditions of Russian technical education?
Not at all. We are situating ourselves within a global context. Stanford University, MIT and EPFL are all clear benchmarks for us. The HSE master’s and undergraduate programs are largely the result of following European and American educational practices. But it is important to combine the Western approaches to the formation of educational programs using the expertise of our leading universities and mathematics schools. Our graduates have always been known for their wide-ranging knowledge and we plan to maintain this tradition.