In memory of Alexei Yakovlevich Chervonenkis

    This text was written in memory of Alexei Yakovlevich by Arkady Volozh, co-founder and CEO of Yandex.

    On September 22, Alexei Yakovlevich Chervonenkis , a leading employee of the Institute for Management Problems of the Russian Academy of Sciences, professor at the Moscow Institute of Physics and Technology and University of London, a teacher at the School of Data Analysis, a man who made a huge contribution to the theory of machine learning , was tragically killed .

    So far, the science of machine learning has had three periods: the pre-computer, computer, and modern big data periods.

    The first great work of Chervonenkis and Vapnik was this article, published in 1971. The theory of convergence of frequencies to their probabilities determined the development of this part of science for several decades to come.

    This was a period of "theoretical" development of machine learning. Then it was possible to count only on some M-200s, or, in a good case, on BESMs, so there was no talk of “widespread use in the national economy”. But to distinguish targets in the air, for example, or to seek out noise on echo-cardiograms, this already helped.

    Then came the second stage of the science of machine learning, computer. In the 1990s, people learned, for example, to pretty well recognize and digitize texts (including handwritten ones) or to clear mail from spam. Half of these methods worked on the famous SVM (Support Vector Machine, support vector method ), invented in the early 1990s by Chervonenkis and Vapnik ( VC-Dimension= Vapnik-Chervonenkis dimension). In the mid-2000s, they worked for SVMs in any well-known office - and here, and in Yakhu, and in Google, and in Amazon, and in Netflix. SVM is described in any textbook on our topic.

    And then came the third era in the development of machine learning: big data and methods of working with them appeared. Now it seems that everything that surrounds us, all objects and services will become a little smarter and learn to help us in every little thing, a little anticipating our desires. It’s just like before that different mechanical and chemical inventions changed our life a little - only now in a slightly new area.

    In the third era, Chervonenkis taught at SHAD-e, and last year at our conference he spoke with the development of his fundamental work in 1971.

    Alexey Yakovlevich loved to walk. He walked 20 kilometers a day - in Moscow, in London, in the forest - so he thought. In the summer he had an operation, and he could not walk for three weeks. Then one day he walked again - a kilometer, then two, three. And last week, he went his 20 kilometers along a familiar road in Elk Island.

    Also popular now: