A Channel Modeling Method Based on Quantum Generative Adversarial Networks

Abstract

The invention discloses a channel modeling method based on a quantum generation confrontation network, comprising: constructing a quantum generation confrontation network model, the generation model is composed of multi-layer quantum generation circuits, and each layer of the quantum generation circuit is constructed by adjustable parameter quantum gates The rotation layer and the entanglement layer constructed by the controlled NOT gate, the discriminant model is composed of a deep neural network; the random channel is sampled to obtain the data sample; the mutual information between the channel sample data sets is calculated as the weight and the Chow-Liu Tree is constructed to Select entangled qubit pairs; according to the confrontation training algorithm, optimize and update the parameters of the corresponding network model through the batch gradient descent optimization of the quantum generation model and the discriminant network cost function; the present invention uses the natural advantages of quantum in fitting probability distribution to perform channel modeling , is an important application of the combination of quantum computing and machine learning in communication scenarios.

Description

technical field [0001] The invention relates to quantum machine learning and channel modeling, in particular to a channel modeling method based on quantum generation confrontation network, which belongs to quantum machine learning. Background technique [0002] In the field of communication, machine learning techniques have been applied to the physical layer processing of communication systems. With the development of this technology in the field of communication, many studies are devoted to applying machine learning algorithms to different communication scenarios. In traditional communication systems, it always consists of different modules such as source coding, channel coding, modulation, demodulation, estimation, equalization, etc. For example, an end-to-end communication system under an AWGN channel based on a fully connected neural network like an autoencoder. It can achieve performance similar to conventional systems of (7,4) Hamming codes and BPSK modulation. This...

AI Technical Summary

Problems solved by technology

[0004] From the theory and experiment, the above studies are either the combination of the field of machine learning and communication, or the combination of quantum computing and machine learning. There are few applications of quantum machine learning in the field of communication.

Moreover, the current research on quantum generation networks is insufficient, generally focusing on the field of quantum neural networks, and less research on generative confrontation models

Images