Sberbank of TagMe

Main articles:

• Marking of data (data labeling)
• Machine learning (Machine Learning)
• Big Data

TagMe is the high-loaded platform of Sberbank for a marking of data. Using TagMe large volumes of the marked data are created ("data sets") at which the artificial intelligence studies. Many AI products of Sberbank (chat-bots, voice assistants, robots lawyers) already use TagMe data sets for improvement of a performance of the ML models.

2020: An opportunity to earn from a marking of data for residents of North Caucasus federal district

On September 11, 2020 Sberbank announced TAdviser providing seven regions of the North Caucasian Federal District (NCFD) to jobless inhabitants an opportunity to earn from the platform of a marking of data TagMe.

Sberbank will pay 400 unemployed of the Caucasus for viewing and a marking of pictures with food. Photo history: fertilizerdaily.ru

Representatives of SberAI (group of the divisions of Sberbank which are engaged in the advanced projects in the field of artificial intelligence) Maxim Eremenko, Andrey Chertok and Marianna Danilina discussed start of a pilot project with representatives of office of the envoy of the Russian President in North Caucasus federal district.

And that who needs a side job it will be offered to jobless inhabitants of regions of the North Caucasus to be engaged in a marking of audio recordings and images of food for a number of projects of SberAI in the field of speech recognition and computer vision. At the first stage about 400 residents of North Caucasus federal district will take part in the project. If "pilot" is successful, he can be scaled including in other federal districts.

That the artificial intelligence learned to recognize the speech or images, it needs to show thousands, tens of thousands of examples (the more — the better) where the same problem was solved by the person. Preparation of such examples is also called a marking of data. At the expense of it the basis for machine learning prepares.

Markers of data earn the monetary reward for correctly performed work. On TagMe for September, 2020 also about 500 employees of Sberbank are engaged in a marking. The platform is already used in eight of its divisions. For all the time of work of TagMe 3.5 million objects were already marked.

Alexander Vedyakhin, the First Deputy Chairman of the board Sberbank:

Thanks to a large number of qualitatively marked data Sberbank creates products and solutions which eventually increase quality of life of our citizens on the basis of artificial intelligence. We are glad that thanks to a pilot project in the North Caucasus we solve this problem, giving to people an opportunity to be engaged in interesting and useful work and to receive for it money.

Yury Chaika, Russian President's Plenipotentiary in the North Caucasus federal district:

One of the main social and economic problems of subjects of the North Caucasus to which solution efforts of federal and regional authorities are directed is still the high level of unemployment. In the period of a complication of an epidemiological situation and, respectively, a situation in labor market other tools which in short terms will allow residents of the district to gain an additional income are required essentially. The platform of a marking of data of Sberbank in this plan — very perspective project which will undergo approbation in the North Caucasus.