The article discusses modern approaches to forecasting and detecting forest fires using machine learning technologies and remote sensing data. Special attention is paid to the use of computer vision algorithms, such as convolutional neural networks and transformers, to detect and segment fires in images from unmanned aerial vehicles. The high efficiency of hybrid architectures and lightweight models for real-time operation is noted.

Keywords: forest fires, forecasting, unmanned aerial vehicles, deep learning, convolutional neural networks, transformers, image segmentation

The article discusses the use of a recurrent neural network in the task of predicting pollutants in the air based on simulated data in the form of a time series. Neural recurrent network models with long Short-Term Memory (LSTM) are used to build the forecast. Unidirectional LSTM (hereinafter simply LSTM), as well as bidirectional LSTM (Bidirectional LSTM, hereinafter Bi-LSTM). Both algorithms were applied for temperature, humidity, pollutant concentration, and other parameters, taking into account both seasonal and short-term changes. The Bi-LSTM network showed the best performance and the least errors.

Keywords: environmental monitoring, data analysis, forecasting, recurrent neural networks, long-term short-term memory, unidirectional, bidirectional