5G technology can transmit information at high speed and at the same time with low delay, while artificial intelligence minimizes operational complexity by using efficient algorithms. This means that the devices that use them will be faster, more productive and more profitable. In addition, the combination of 5G and UI will improve the cost-effectiveness of a wide range of processes through the use of automated systems. This automation can reduce the effort and cost required to perform industrial functions while increasing accuracy. We can say that UI and 5G are pushing each other forward. 5G together with IoT provides a large amount of data on which the AI can be trained and subsequently the trained algorithms can be used to optimize 5G networks.[O43]

UI and its sub-categories such as ML and DL are evolving to such an extent that currently this mechanism enables fifth generation (5G) wireless networks to be proactively anticipated, which is critical to the realization of the 5G vision. The vision is to develop self-determining intelligent base stations, mobile devices forming dynamically adaptable clusters based on learned data rather than pre-set rules. Thanks to the use of these mechanisms, the efficiency and reliability of current network applications, including real-time applications, are increased. This improves latency and overall quality. Many problems in mobile and wireless communications are not linear or polynomial and therefore need to be approximated. Artificial Neural Networks (ANNs) are a technique that was designed to model the objective function of a nonlinear problem that requires optimization. [O30]

The goal of ML is not to find the absolute best learning algorithm. Instead, we need to understand what kind of partitioning is relevant for a particular 5G/B5G application and which ML algorithm performs best on that particular data. The main challenge in 5G traffic design is the recognition and classification of heterogeneous massive traffic. Instead, new reliable UI-based methods should be introduced. The UI has proven to be a powerful tool for solving problems that require a lot of manual debugging or complex lists of rules. In addition, UI is suitable for solving complex problems where traditional approaches cannot provide an adequate solution. The unpredictability of the operation of 5G networks is a challenge mainly for network operators[O40].

In wireless networks, it is crucial to ensure efficient control and digital processing of signals, so it is essential to obtain channel state information (CSI). A channel state indicator (CSI) is used in the context of wireless communication, and an NN-based approximation has been developed to determine channel conditions. This method allows inferring unobservable channel information from known observable channel information. When estimating propagation losses in dynamic environments, input parameters can be selected from information about the transmitter, receiver, buildings, frequency, etc. The network learns from this data to estimate a function that best approximates the propagation loss for next-generation wireless networks. Among other things, UI applications can be extended to cell selection in multi-level networks, device discovery for device-to-device D2D communication, or user pooling to balance the overall network load. UI, especially DNN (Deep Neural Network), has been applied to various functional blocks in the physical layer, e.g. modulation recognition, polar code decoder and MIMO detection [O44].