Yandex School of Data Analysis has developed a neural network for cleaning the shores of reservoirs from garbage

571
2
AK&M 13 May 2025 07:58

The Yandex School of Data Analysis, with the support of Yandex B2B Tech and FEFU, has developed and made publicly available a neural network that can determine the volume, mass and types of garbage on the shores of reservoirs. This is stated in the company's message.

The neural network analyzes aerial photographs of the coast and divides the garbage into six types. The classification accuracy exceeds 80%. The model marks the coordinates of the waste location on the map, indicates its composition and weight. This allows you to calculate the size of the required group of people and the number of cleaning equipment.

The new solution was applied in an ecological expedition in the South Kamchatka Federal Reserve, a specially protected natural area under the management of the Federal State Budgetary Budgetary Institution Kronotsky State Reserve in the Far East. Experts have found that 33-39% of coastal pollution comes from plastic containers and packaging, and 28-29% from industrial fishing waste. By using a neural network, it was possible to organize the cleaning of 5 tons of waste 4 times faster.

Currently, the neural network is also being tested in the Arctic and other regions. Environmental control services and volunteers will be able to use the technology for free to collect garbage faster in hard-to-reach places.

ICPAO Yandex is a Russian IT company. Its ecosystem includes more than 90 B2C and B2B services in such areas as information search (Search, Browser, Maps), advertising and business services (Yandex. Direct, Yandex.Metrica, Yandex Cloud, Mail, Disk), taxi ordering, carsharing and scooter rental, food delivery (Shop, Food), e-commerce (Market), education (Workshop, Academy, Textbook).

Yandex's adjusted net profit for 2024 under IFRS increased by 94% to RUB 100.9 billion from RUB 52.1 billion a year earlier. Net profit fell by 79% to 11.5 billion rubles from 55.6 billion rubles. Revenue increased by 37% to RUB 1.09 trillion from RUB 798.1 billion, adjusted EBITDA increased by 56% to RUB 188.6 billion from RUB 120.8 billion.