Quantum Delta Encoding: Optimizing Data Storage on Quantum Computers with Resource Efficiency

Jiale Zhang, Xilong Che, Yuzhe Fan, Juncheng Hu

Abstract: In recent years, numerous studies have explored efficient methods for encoding classical data into quantum systems by leveraging Quantum Random Access Memory (QRAM) to facilitate subsequent data processing tasks. However, current state-of-the-art encoding techniques rely on extensive multi-qubit controlled-NOT gates and require complex quantum gate decompositions to ensure compatibility with existing hardware. Other approaches have attempted to employ quantum neural networks for state preparation–either to enable quantum data compression or to perform quantum Fourier transforms for preserving frequency-domain information–but these methods typically involve significant preprocessing and fail to accurately recover the original classical data. In this paper, inspired by classical Delta Encoding, we propose Quantum Delta Encoding (QDE), which stores the majority of data in a benchmark and encodes only the deviations via entanglement, thereby significantly reducing the need for entangled qubits and quantum gates during storage. Moreover, QDE can seamlessly integrate with QRAM to support subsequent quantum data processing tasks–such as image processing and data encryption–thus mitigating the additional errors and losses associated with repeated classical-to-quantum data exchanges. We evaluate the advantages of QDE over state-of-the-art models using real-world datasets and assess its robustness against quantum noise. Experiments conducted on both the IBM Quantum platform’s simulator and two real superconducting quantum computers confirm the validity and potential of the QDE approach. All codes and data are available at https://github.com/kennyZhangsky/Quantum-Delta-Encoding.

External IDs:dblp:conf/europar/ZhangCFH25