FEEDBACK ON NEURAL NETWORK FORMATION CONFIGURATION FOR WIRELESS TRANSMISSIONS Russian patent published in 2020 - IPC H04L5/00 

FIELD: wireless communication systems.

SUBSTANCE: network entity (a base station, a core network server) determines a set of reference signals of a deep neural network for sending on a physical channel to measure the performance of deep neural networks and / or a set of configurations of forming a neural network. Network entity transmits the set of configurations of forming a neural network and a set of reference signals of a deep neural network to user equipment. Network entity initiates transmission of each reference signal of a deep neural network from a set, and instructs user equipment to measure the performance of a set of deep neural networks formed using a set of configurations of forming a neural network, and select one of a set of configurations of forming a neural network.

EFFECT: technical result consists in improvement of overall efficiency of downlink transmissions, reducing coefficient of erroneous bits and improving signal quality.

18 cl, 31 dwg