Hybrid Quantum-Classical Autoencoders for End-to-End Radio Communication

TL;DR

This paper proposes hybrid quantum-classical autoencoders for radio communication, demonstrating improved robustness against noise and efficient inference through a novel data re-uploading scheme in quantum circuits.

Contribution

It introduces a new hybrid quantum-classical autoencoder architecture tailored for end-to-end radio communication systems, combining quantum decoding with classical encoding.

Findings

Enhanced robustness to noisy channels

Effective latent space representation of radio signals

Feasible inference-time performance with data re-uploading

Abstract

Quantum neural networks are emerging as potential candidates to leverage noisy quantum processing units for applications. Here we introduce hybrid quantum-classical autoencoders for end-to-end radio communication. In the physical layer of classical wireless systems, we study the performance of simulated architectures for standard encoded radio signals over a noisy channel. We implement a hybrid model, where a quantum decoder in the receiver works with a classical encoder in the transmitter part. Besides learning a latent space representation of the input symbols with good robustness against signal degradation, a generalized data re-uploading scheme for the qubit-based circuits allows to meet inference-time constraints of the application.