MOSCOW, June 27Beeline and Brand Analytics presented a rating of companies involved in artificial intelligence in Russia, which was compiled by a neural network — in the main categories, VKontakte, Yandex and Sberbank were in the lead.
Beeline, together with Brand Analytics, presented a rating of companies and institutions involved in artificial intelligence in Russia, which was compiled by a neural network,» the press release says. Thus, Yandex became the leader in the categories «Import substitution», «Cloud solutions», » Significant decisions», «Contribution to the development of AI». VKontakte was the leader in the nominations «Contribution to personnel training», «Social AI» and «Presence in the information field». Sberbank became the first in the nomination «Investments in AI».
A total of 248 companies were included in the rating. The neural network analyzed about 500 thousand materials — these are publications in the media and telegram channels dedicated to the topic of artificial intelligence. In addition to the eight main nominations, the winners were determined in four special ones: «Retail and artificial intelligence» (in the first place — «Magnit»), «Telecom and AI» (leader — MTS), «Banks and AI» (in the first place — Sberbank) , «Universities and AI» (Moscow State University named after M.V. Lomonosov was in the lead).
“We are well aware that it is important for market participants not only to keep track of news in the field of AI, but also to have before their eyes a navigator of companies and solutions in this industry, which is important for our future with you … The project turned out to be interesting. The rating also takes into account the media activity of players , and attention to the topic and companies from the media, and «popular» discussion in social media, in this case — in telegram channels,» said Natalia Sokolova, managing partner of Brand Analytics.
She clarified that the analysis of materials through a neural network has a number of advantages: «Firstly, it became possible to build a final rating based on individual nominations with a large number of indicators. Secondly, we have access to raw data,» having «fallen» into which we «We can make sure that the calculation is correct. Of course, there will be a lot of controversy about the rating. You won't pass by such a project, and we want it.»
The full list of companies and the AI algorithm training methodology are posted on a special AI rating page.