Quantum Sparse Coding and Decoding Based on Quantum Network

TL;DR

This paper introduces a symmetric quantum neural network for sparse coding and decoding, achieving high accuracy in image and quantum state reconstruction, and demonstrating advantages over classical models.

Contribution

It proposes a novel quantum neural network architecture for sparse coding and decoding, suitable for optical circuits, with efficient training and high-performance results.

Findings

98.77% accuracy in image reconstruction

97.68% fidelity in quantum state revivification

Improved robustness over classical models

Abstract

Sparse coding provides a versatile framework for efficiently capturing and representing crucial data (information) concisely, which plays an essential role in various computer science fields, including data compression, feature extraction, and general signal processing. In this study, we propose a symmetric quantum neural network for realizing sparse coding and decoding algorithms. Our networks consist of multi-layer, two-level unitary transformations that are naturally suited for optical circuits. Each gate is described by two real parameters, corresponding to reflectivity and phase shift. Specifically, the two networks can be efficiently trained together or separately using a quantum natural gradient descent algorithm, either simultaneously or independently. Utilizing the trained model, we achieve sparse coding and decoding of binary and grayscale images in classical problems, as well as that of complex quantum states in quantum problems separately. The results demonstrate an accuracy of 98.77\% for image reconstruction and a fidelity of 97.68\% for quantum state revivification. Our quantum sparse coding and decoding model offers improved generalization and robustness compared to the classical model, laying the groundwork for widespread practical applications in the emerging quantum era.