Yandex Introduces AI-Based Copying Network For Waste Cleaning

JAKARTA - Yandex B2B Tech, Yandex School of Data Analysis, and Far Eastern Federal University (FEFU) introduced a neural network, an AI solution to accelerate the process of cleaning up waste on the beach.

Developed by Yandex in the form of open source, this open source AI-based solution can help environmental institutions and volunteer communities, in accelerating the process of transporting solid waste, including plastic, in ecological zones that are considered sensitive around the world.

With machine learning or machine learning features, detection and analysis can be done automatically by neural networks. Even the pollution assessment process was simplified.

This technology has been successfully used in remote areas of the South Kamchatka Federal Nature Reserve area, East Russia and is currently being tested in the Arctic region and a number of other areas.

During the trial process in the Kamchatka nature reserve, the neural network found that 33-39% of the waste on the coast consisted of containers and plastic packaging and 27-29% came from the fishery industry.

Neural network technology can be developed and implemented further by local volunteer teams and government institutions, both in Indonesia and other countries.

In addition, with a complete open source code source base, customization of the neural network can be done to detect new types of waste, monitor threatened species, and support other environmental efforts.

Neutral networks by Yandex and FEFU can also be integrated with various mapping tools, such as QGIS spatial software which is also open source.

The AI solution developed takes advantage of computer visualization, to ensure that solid waste detection can be done automatically. This method can divide certain types of waste such as fishing nets, iron, rubber, large pieces of plastic, concrete, and wood, reaching an accuracy level of more than 80%.

Neutral network then maps the waste location, estimates its volume and weight, calculates the workforce and equipment needed - including garbage trucks and vehicles for all terrain.