‘Borders Between Countries Are Becoming Blurred Thanks to Online Communication’
Professor Oleg Melnikov is among the international professors invited to work remotely with HSE University’s students this academic year. He lives in California, runs the Data Science department at a company in Palo Alto, and teaches at Stanford and other universities in the United States. At HSE University he teaches a course on machine learning for the students of the Faculty of Computer Science and the International College of Economics and Finance (ICEF), as well as a university-wide optional course, ‘Machine Learning in Python’. He spoke about his work in an interview with the HSE News Service.
Oleg, where did you receive your education? What are you doing now? What are your research interests?
I got my degree in mathematics, finance, and computer science in Los Angeles and my PhD in statistics in Texas. Now I live in Silicon Valley, California.
Oleg Melnikov
In my main job, I run the Data Science department at ShareThis Inc, based in Palo Alto. We build machine learning models to classify Internet users that the advertising industry is interested in. This is basically working with hundreds of terabytes of natural language data.
For ten years now, in my spare time, I have been developing and teaching courses in Machine Learning, Statistics, Neural Networks, Data Science, NLP and Python at Stanford, Berkeley, Cornell, the University of Chicago, the University of Washington, Johns Hopkins, and other American universities. For me it is not only a hobby, but also self-development, an opportunity to share knowledge and experience with students and assistants.
This year I started teaching at one of Russia’s best universities — HSE University. Here I am also getting acquainted with an educational system that is new to me. I consider myself very lucky.
How did your collaboration with HSE University begin?
I had considered the possibility of teaching in Russia before, communicating with colleagues from Russian universities. Now it has become much easier to realise this idea thanks to the development of online education during the COVID-19 pandemic.
I applied to the Faculty of Computer Science at HSE University and Vladimir Podolskii, Evgeny Sokolov, and Tamara Voznesenskaya immediately responded to my proposal. We organised several video meetings, where they were interested in my motivation, qualifications, and opportunities, and discussed possible options for collaboration. I was able to get through this stage surprisingly quickly. Now I work under an employment contract, which was easy to arrange.
Evgeny Sokolov, Deputy Head of Big Data and Information Retrieval School, Faculty of Computer Science, HSE University
Professor Oleg Melnikov approached us about a year ago, asking if we had a need for visiting scholars. He already had a lot of experience teaching at universities in the U.S., but now he wanted to see ‘from the inside’ how education works in Russia. Of course, we were interested, and we even discussed the possibility of Oleg coming to Moscow for a semester. But then, once the pandemic began, we decided to arrange remote teaching.
It is certainly an interesting experience for us. For example, Oleg builds his course in a blended format, uses online materials, and prepares a large amount of questions for automatic student testing. The way in which he works with students in his classes is very organized and well-structured. We have professors who assist with Oleg's courses, and I think they have learned a lot from him.
What classes do you teach at HSE University? In your opinion, how are your courses unique and how are they useful to students?
I'm just now finishing up a first-semester English language course in machine learning for students in the double degree bachelor’s programme in ‘Data Science and Business Analytics’ in the Faculty of Computer Science and the double degree programme in Economics at ICEF.
I taught this course last year at Stanford. It is based on the very famous textbook Introduction to Statistical Learning (ISLR) and Elements of Statistical Learning (ESL) by Robert Tibshirani and Trevor Hastie from Stanford, with whom I was fortunate to collaborate. Students are exposed not only to new material, but also to the LMS systems that I use in my work as a teacher in the United States — Canvas LMS, PIazza LMS, Google Colab, and others.
Another course I teach is a university-wide elective. It is devoted to machine learning in Python and is taught in Russian.
How unusual did HSE University’s students find your approach to organising classes?
Of course, many students were new to some of the methods that are often used in the U.S. educational system. These are grading on a curve (an unusual method of assessment for Russia), weekly, individual and automated tests through Canvas, and, of course, the Canvas and Piazza systems themselves, which I have been using for many years. I am pleased that I am not only sharing my knowledge with students, but also introducing them to other teaching approaches.
Many of my Russian assistants are also learning new systems and teaching methods that will help them create their own courses in the future. By the way, I meet with my wonderful team of fellow professors and teaching assistants on a weekly basis through Zoom, where we discuss current work and plan tasks for the coming weeks and months. As they say, one head is good, but fourteen is even better. They are all great guys and a pleasure to work with.
I am also very inspired by the support from the HSE University administration and my colleagues, who are responsible for organising my work. I have the opportunity to interact with outstanding Russian professors. It seems that at no other university have I worked with so many people at the same time. Absolutely everyone is very friendly and always ready to help with advice and effort.
What are your plans for further cooperation with HSE University?
We are witnessing amazing changes in education, where borders between countries are becoming blurred thanks to online communication between students and teachers. So there are a lot of different ideas for collaboration and hopefully it will only get stronger every year.
For example, based on the material created with automated and individual forms of student assessment, we can significantly increase the scale of my course without increasing the number of assistants. We can also involve students in interesting projects in machine learning with companies in the U.S., where they will gain additional experience, including research and practise their English. We also plan to invite prominent researchers from the U.S. to conferences and presentations at HSE University.
Tamara Voznesenskaya, First Deputy Dean, Faculty of Computer Science, HSE University
Professor Oleg Melnikov had the difficult task of combining machine learning course materials from the London School of Economics and the materials that are traditionally used in the machine learning course offered by the Faculty of Computer Science. We were lucky that not only does he know the London course well, but he also worked with the authors of the textbook and has a good sense of what is written between the lines. We are also happy to have the opportunity to explore Stanford's teaching traditions in action and to take a detached view of HSE University and the Faculty from the outside. We are learning a lot from each other. In the second semester, Oleg is continuing to teach his course in English. I hope our cooperation will only strengthen and expand in the coming academic year.