Machine learning and deep learning methods for wireless network applications

Wireless networks have been widely adopted and introduced in the areas of engineering, manufacturing, weather monitoring, transportation, etc., to collect data to improve the quality of decision making, but issues arise, such as large volumes of data, incomplete and incompatible data sets, and noise data that prevent from realizing the true value and exploiting their full potentials. Machine learning and deep learning methods have been used as powerful tools to perform feature detection/extraction and trend estimation/forecasting in wireless networks applications. Supervised machine learning methods, for example, neural network (NN) [1, 2], convolutional neural network (CNN) [3], and recurrent neural network (RNN) Furthermore, the reinforcement learning methods, including generative adversarial networks (GANs) and deep Q-networks (DQNs) These well-developed methods can contribute substantially, to better improve predictions and classifications in the relevant applications, but there are some issues and limitations that require further attention from the research communities.