Vladimir Kukushkin is one of the participants of the course on neural networks.

I came here just to get an idea of what it is. Saw an advertisement, got interested. Neural networks were interesting for me, I saw an ad about learning this type of stuff and decided to try. I started from square one, with an "empty head" in this area. I was interested. Interested both from a professional point of view - as an IT specialist, my profile - IT project management, and as a teacher in the past - 10 years of teaching at the university will never be forgotten. In addition, it should be noted that you need to keep up with the IT advancements in the modern realities. It is interesting. It's tight. I would say that it is hard to study extra, i.e. when you study in parallel with your work and personal life. It's a lot, I can't keep up. Even without laziness, without slacking off, you can't make it. But it's interesting. It motivates you, forces you to concentrate. I come from work, do my homework, like in college, like in school. That's awesome. It's interesting. It's great. Honestly, at first I didn't understand the concept of the program, teaching, there was some chaos in my head: "What are we doing, for what?" A lot of assignments, and I don't know why. I contacted Dima, we talked, he explained to me that it's deliberate that that we are building a foundation, and then in the laboratories, coursework, diplomas, we will work in a specific direction. Everything's fine now, except for the lack of time. One of the important advantages of the current course, the current team, in my opinion, is its "anti-commercial" component. Obviously, the whole team works not for money, but for their own interest, and this can be seen when you have classes with the lecturer and all people's eyes are burning with enthusiasm. This is awesome. I mean, people work just because they don't like it, they like it. I'll tell you the truth, I enjoyed even this atmosphere in class, I even like it, not to mention the material that is useful and necessary.

My name is Timofey Skrylnik, I'm a member of the Neural Networks on Python course, February section. It was a very cool course. I had to study, do a lot of stuff, practice a lot to get to the result. The course was cool not only in terms of the information it provided, but also a huge amount of practical work, which helped me to sharpen my skills and pull them up to a level that will be interesting in real projects.

The teachers and curators are great. Many thanks to Gerard Kostin: he was very clear about his homework, gave feedback, always helped with advice. This is the curator of my group.

It was a good project in the lab, where we tried to make a generative network. We didn't finish it - it was very difficult, but we trained very interesting skills that are already in demand. The lectures were good, too. Indeed, topics that are very complicated if you Google them. You can find some information, but quite superficial, for example, how to classify comments about films, but how to work with text, how to extract meaning, how to do some more advanced things - this is almost non-existent, and this course covered it.

My project is the recognition of plant diseases by photo, which includes a network and a mobile application so that it is a complete solution that can be applied. The project is successful within this task. It shows 80% disease recognition on the selected dataset. I now look forward to working with another company where there are already real biologists and the biological base of a laboratory to bring this project to real commercial development.

For those who are just planning, I may wish to work hard, study this topic and be very successful. Even if you have doubts, it will at least be your personal growth, drive, you will learn a lot, you will become an actual specialist today who knows the technologies that are relevant now, not those that are outdated and will soon be replaced by newer ones. So study, attend a course, everything will be great.

My name is Yuri Kobyzev. I'm the head of StackCom's information systems department, a satellite TV company. I was interested in things connected with video processing of information, with remote monitoring and transmitting processed information via satellite to the center. At the same time, since the traffic is very expensive, we pre-processed and transmitted already very little information about the objects that we detect in remote mode. For this purpose, I use single-core computers by Nvidia Jetson Nano. And everything worked out thanks to our teachers, who inspired me to work independently, and thanks to some experience, because I was looking for a solution for my business. Although, of course, and as self-education, it is very useful. I got a good platform, after which I'm learning more complicated things now. NLP, for example, from Samsung. I really like neural networks there. I want to know what the meaning of text is and how to calculate it. I really liked the atmosphere, the people I had to talk to here, and the experience I gained.

I had Dmitry Yermilov as my curator. I really liked him. After that I went to his course "Machine Training". He graduated from MSU Faculty of Computational Mathematics and Cybernetics, PhD. Since I'm from MSU myself, I liked him immediately. He told us some basics in different areas of classical machine learning. I realized that I had to pull up some more mathematical slips, after his stories. The most important thing for me is not even the result, but the ideas that lie behind the different approaches, and how these ideas correspond with my own intellect and consciousness, what kind of mathematics lies in me, how complicated I am a neural network or maybe I am not a neural network. I had to compare the inner constants with the outer ones. Everything was interesting.

My name is Sergei Loktev, I'm a graduate of the University of Artificial Intelligence. I remember when in my student years, which was the very beginning of the 90s, I painfully chose between programming and mathematics. At that time, programming was a means of achieving known results, so it didn't seem so interesting to me at that time, so I went into pure mathematics. After a while I learned that there is machine learning where you can get completely unpredictable results, the result is not completely determined by the algorithm, and you can get something unexpected, unusual. I got interested and took a two-month course first, and then a full six-month course.

It was very interesting, very stimulating environment I liked. And the seminars were very nice, and the assignments at Kaggle, where you could compete with others. It's really very warm, the atmosphere is supportive and the tasks are not easy.

My project was about recognizing emotions. This is the most interesting question for me, because the first one where I saw neural networks do wonders is a project implemented by Microsoft, where you can tell from a photo which of the basic emotions a person is experiencing at the moment. I tried to do the same for my voice. It worked out to a certain extent. Of course, to make it work, you need a much bigger dataset, more marked data, but you can still see that even now the neural network can see a lot.

I am now studying Bayesian methods - these are methods in machine learning that were not long before neural networks. They are, in my opinion, underestimated and after a while will be a synthesis of Bayesian methods and neural networks that can predict a lot when there are many random factors and the prediction will be much more accurate.

First of all, I am glad to see the people I met during the course and those I took the exam with. Thank you all: thanks a lot to the University, to the assistants, to the curators, to Dmitry personally, because in fact, we all already know, he's pushing everyone forward. For those who are studying now, I can advise you not to give up even in difficult moments. There are such deadlocks, I know from personal experience that something just does not go further. Here you need to be a little persistent, and then the curators will help, and everything will move, and everything will be cool, and everything will go. I think a lot of people have been through this.

As for the diploma, in a nutshell, it's bar code detecting. In images, when there's a lot of them, the scanner can't read them. It turned out that it crossed with a commercial project and it was necessary to apply some model that would detect them and then the scanner would already read them. Of course, we had to dig with R-CNN, Faster R-CNN.As a result, everything took off on Faster R-CNN - this is a good thing, I made this offline solution. Anyway, thank you all, I was very glad to study here and plan to continue.

Let me tell you about the course - for me it was a hell of a transformation, because in recent years I have become such a bureaucrat, sat down with bureaucratic management activities. This program took me out of there by my hair, pulled me out into the white light, where you can create some projects, not drown in bureaucratic procedures, the very head.

I want to say the most important thing is that we shouldn't split up, I think, because we have very cool topics and we can team up to further develop some things and then go out to some customers and offer them that. I hope that after the end all the interesting things are just beginning.