A Simulation Framework for Noise and Error-Correction Analysis in Device-Independent Quantum Key Distribution

arXiv:2604.22232 · quant-ph · Submitted 2026-04-24 · Read on arXiv

quant-ph

Submitted: 2026-04-24

Updated: 2026-09-19

Comments: 13 pages. This paper is submitted to SOICT2026

Code: https://github.com/duyndh98/DI

License: http://creativecommons.org/licenses/by/4.0/

The gist: Device-Independent Quantum Key Distribution (DIQKD) certifies security from the violation of a Bell inequality alone, but its practical margin against noise is narrow and is usually quoted only

Terminology

Abstract

Device-Independent Quantum Key Distribution (DIQKD) certifies security from the violation of a Bell inequality alone, but its practical margin against noise is narrow and is usually quoted only through the Bell violation itself. We build a simulator of the event-ready DIQKD experiment and use it to trace readout noise through the Bell test into classical post-processing. For a symmetric outcome-flip channel of strength e, we reproduce the simulated violation window e<5.86%, to within 0.11 percentage points. Combining these with a multi-pass Cascade reconciliation and the Devetak-Winter bound gives r<=0.162 bits per round, and shows that a positive key rate requires e<0.89 - roughly one-sixth of the window in which a Bell violation is still observed. The gap quantifies how far certified nonlocality overstates the usable noise budget of a DIQKD link.

Sources

Related papers