[Paper Review] Semi-Device-Independent Quantum Random Number Generator Resistant to General Attacks

Zhenguo Lu, Jundong Wu+5 more|arXiv (Cornell University)|Feb 6, 2026

Chaos-based Image/Signal Encryption0 citations

TL;DR

The paper proposes and experimentally demonstrates a semi-device-independent QRNG that tolerates general (non-i.i.d.) attacks under finite-size effects, using a ternary-input continuous-variable scheme with heterodyne detection to achieve a net rate of 1.165 Mbps at 5.3×10^9 rounds.

ABSTRACT

Quantum random number generators (QRNGs) produce true random numbers based on the inherent randomness of quantum theory, rendering them a foundational segment of quantum cryptography. Distinguished from trusted-device QRNGs whose security depends on characterized devices, semi-device-independent (semi-DI) QRNGs permit partial devices to be defective or even maliciously manipulated, which achieves a good trade-off between generation rate and security. In this paper, we propose a semi-DI QRNG that resists general attacks while accounting for finite-size effects. The protocol requires no rigorous characterization of the source and measurement devices other than limiting the energy of the emitted states, significantly reducing the demands on practical QRNG systems. Leveraging the tight Kato inequality for correlated variables, we show that our protocol generates more randomness than it consumes. Furthermore, we demonstrate the scheme on a continuous-variable system with ternary inputs of states. Heterodyne detection is employed to enable phase compensation through data postprocessing, alleviating the stringent requirement on system stability. The system operates at 100 MHz, achieving a net random number generation rate of 1.165 Mbps at 5.3x10^9 rounds. Our work offers a promising approach to achieve both the robust security and high generation rate with a simple experimental setup.

Motivation & Objective

• Develop a semi-DI QRNG that resists general attacks while accounting for finite-size effects.
• Relax device characterization requirements by constraining only the energy bound of emitted states.
• Enable high randomness generation rates with a simple, practical experimental setup.
• Confirm that the protocol can generate more randomness than it consumes (randomness expansion).

Proposed method

• Use a three-input, multi-output semi-DI QRNG with an untrusted measurement device treated as a black box and a constrained state energy overlap.
• Model general attacks via a SDP-based security analysis that accounts for correlated measurements and finite-size effects using Kato’s inequality.
• Implement in a continuous-variable system with heterodyne detection and three-input ternary states, enabling phase compensation through data postprocessing.
• Compute min-entropy Hmin from the worst-case Eve guessing probability Pg and extract randomness with a strong extractor.
• Demonstrate security and performance for N rounds with a generation-round probability and test-round probability pt, including the asymptotic and finite-size regimes.

Experimental results

Research questions

• RQ1 How much private randomness can be certified in a semi-DI QRNG under general (non-i.i.d.) attacks with finite-size effects?
• RQ2 What is the impact of energy-bound state constraints on randomness generation in a semi-DI protocol without full state characterization?
• RQ3 Can a continuous-variable, heterodyne-detection implementation achieve secure randomness expansion with practical throughput?
• RQ4 How do phase drifts and finite-size statistics alter the net randomness rate and required randomness overhead in the protocol?

Key findings

• The protocol yields a net randomness generation rate that accounts for randomness consumption, with a demonstrated experimental net rate of 0.01165 bits per round.
• In the experiment operating at 100 MHz, the secure net randomness generation rate achieved is 1.165 Mbps from 5.3×10^9 rounds.
• The setup uses three inputs and multiple outputs in a CV system with heterodyne detection, enabling phase compensation through data postprocessing.
• Energy-bound assumptions on the source suffice without full state characterization, and the analysis uses Kato’s inequality for correlated variables to bound Eve’s guessing probability.
• The generated bit sequence passes the NIST Statistical Test Suite, confirming practical randomness quality.