Neural network will predict wildfires with 87% accuracy

Skoltech employees have created a system based on artificial intelligence that predicts the development of forest fires in Russian regions. Similar programs form such forecasts based on one kind of information. Unlike them, the new development makes predictions by analyzing a variety of data. Thanks to this approach, the accuracy of calculations can reach 87%. It depends on the quality of information that the machine takes into account. The system is now being tested in regions, such as Sakhalin.

"We used machine learning technology. It was conducted on the basis of archived data for 10 years about whether or not there was a fire in a particular place at a particular time. The algorithm analyzed them and itself identified patterns, by which we can already make predictions. Since each region of the country has its own peculiarities, the AI for it needs to be trained separately. To begin with, we chose several pilot regions. It took about one week to train the neural network for each of them," Svetlana Illarionova, head of the research group at the Skoltech Artificial Intelligence Center, told Izvestia.

The AI makes a forecast for five days ahead, as it needs to have reliable weather data. The creators obtained prediction accuracy from 70 to 87% depending on the region. This quality of results is enough for practical use of the system, so that it is possible to introduce timely firefighting measures.

Read more in Izvestia's exclusive article:

AI with fire: neural network will predict forest fires with 87% accuracy