Qcrypt 2022: Poster71

Опубликовано: 21 Февраль 2026
на канале: QCrypt conference
50
0

Simple and Practical Device-Independent Security Analysis

Authors
Xingjian Zhang (Tsinghua University); Pei Zeng (Tsinghua University); Tian Ye (Tsinghua University); Hoi-Kwong Lo (University of Toronto); Xiongfeng Ma (Tsinghua University)

Abstract
Guaranteed by quantum nonlocality, device-independent cryptography offers information-theoretic security without any trust on devices in use. Despite the elegance in security assumptions, the ignorance of devices inner working makes security analysis a challenging task. Current device-independent security proofs are often complex and quite different from those of their more standard device-dependent cousins. The existing proofs also pose extreme challenges to experiments, requiring an extremely large data size, high state-preparation fidelity, and low transmittance and detection loss.
In this work, we provide a simple security analysis for device-independent cryptographic tasks via phase error correction. By recasting the device-independent scheme into a quantum error correction protocol, we show that similar to their device-dependent cousins, device-independent cryptographic tasks share the common security origin of quantum complementarity. Going beyond the identical-and-independent-distribution (i.i.d.) case, we consider the most general attack and achieve a full and tight security analysis. For parameter estimation, we generalise the sample entropy in classical Shannon theory and adopt a martingale-based analysis.
Our method exhibits good finite-size performance and brings the device-independent scheme to the practical finite-size regime. Under experimental parameters achievable by the state-of-the-art cold-atom and nitrogen-vacancy (NV)-centre platforms, it reduces the requirement on data size by two orders of magnitude. Furthermore, the complementarity approach can be naturally extended to advantage key distillation including two-way information reconciliation via the B~step, the noisy pre-processing, and the use of detection loss events, which ease photonic experiments by tolerating higher loss and lower transmittance.

https://2022.qcrypt.net/
#Qcrypt2022