Terminal and base station in wireless communication system
The present disclosure provides a terminal and a base station in a wireless communication system. The terminal may include a control unit configured to input a downlink channel to a neural network of the terminal; and the control unit further configured to control the neural network of the terminal to process the input and output feedback information.
1. A terminal comprising:
a controller configured to:
input a signal related to a downlink channel to a neural network of the terminal, and
control the neural network of the terminal to process feedback information; and
a transmitter configured to:
transmit the feedback information to a base station,
encode and modulate the feedback information or analog modulate the feedback information, and
transmit the encoded and modulated or analog modulated feedback information to the base station,
wherein an encoding and modulation scheme used for the feedback information is related to number of nodes in intermediate output layers of the neural network of the terminal, and
wherein the transmitter is configured to encode and modulate the feedback information or to analog modulate the feedback information comprises: the transmitter quantizes the feedback information by using a neural network-based codebook, and transmits quantized feedback information to the base station, the neural network-based codebook is determined according to an activation function used in the output layers of the neural network of the terminal.
2. A base station, comprising:
a receiver configured to receive feedback information from a terminal, wherein the feedback information is encoded and modulated feedback information which is encoded and modulated by the terminal, or analog modulated feedback information which is analog modulated by the terminal, wherein an encoding and modulation scheme used for the feedback information is related to number of nodes in intermediate output layers of the neural network of the terminal; and
a controller configured to determine an input of a specific layer in a neural network of the base station according to the feedback information,
wherein the feedback information is quantized feedback information which is quantized by using a neural network-based codebook, the neural network-based codebook is determined according to an activation function used in the output layers of the neural network of the terminal.
3. A terminal comprising:
a receiver configured to receive indication information on codebooks from a base station, wherein the indication information indicates whether the terminal uses neural network-based codebooks, the indication information is generated according to the base station's support capability for a neural network and the terminal's support capability for a neural network; and
a controller configured to determine whether to use the neural network-based codebooks according to the indication information,
wherein the neural network-based codebooks are determined according to an activation function used in the output layers of the neural network of the terminal.
4. The terminal according to claim 3 , further including:
a transmitter configured to transmit information on characteristics of the neural network of the terminal to the base station.
5. The terminal according to claim 4 , wherein the information on characteristics of the neural network of the terminal includes information on at least one of a network type and a network size of the neural network of the terminal.
6. The terminal according to claim 3 , wherein the neural network-based codebooks are related to the characteristics of the neural network.
7. The terminal according to claim 3 , wherein the neural network-based codebooks are determined according to output characteristics of the neural network.