AttacksContinuous
Homodyne saturation and local-oscillator calibration against CV-QKD
10/10 passed in 1787ms
| Test | What it does | Result |
|---|---|---|
test_calibration_anchor | Published anchor, Jouguet, Kunz-Jacques and Diamanti, PRA 87, 062313 (2013), arXiv:1304.7024 Eq. (9): 2.1 SNU at T = eta = 0.5 behind a shot-noise ratio of 1.5 reports exactly 1/15 SNU and 61/40 reports zero, both derived from Eqs. (7) and (8) and printed nowhere. | ✅ pass |
test_calibration_break | qkd's key rate is positive on the reported 0.0667 SNU and negative on the true 2.1 SNU, and the Reading reports the shot-noise ratio beside them. | ✅ pass |
test_calibration_raw | calib_xi is unclamped: the reported excess noise equals the true one at ratio 1, falls monotonically with the ratio, and goes negative past 1.525. | ✅ pass |
test_saturation_bias | The estimated excess noise falls through zero as the displacement nears the detector limit, and every target below 2.1 SNU is reproduced by some displacement to 1e-9, Qin's Proposition. | ✅ pass |
test_saturation_break | qkd's asymptotic key rate is positive on the channel the saturated estimator reports and negative on the true one. | ✅ pass |
test_saturation_clip | sat_clip matches the censored-Gaussian second moment by Simpson integration over ten standard deviations and never exceeds the unclipped variance. | ✅ pass |
test_saturation_domain | sat_estimate and sat_gain refuse a negative displacement, a non-positive linear range, a displacement clipping T-hat to zero, and an unreachable level II gain. | ✅ pass |
test_saturation_levelii | Published anchor, partly: holding T-hat by solving Qin Eq. (19) for the gain, the displacement minimising the reported excess noise at 31 km is 19.54 against their 19.5, with V_A from this exam's fixed-SNR rule. | ✅ pass |
test_saturation_null | Published anchor, Qin arXiv:1511.01007 Sec. VII.3.1: at zero displacement a full intercept-resend with 0.1 SNU technical noise estimates xi = 2.1 SNU and an unbiased T-hat at gain G = 2. | ✅ pass |
test_saturation_printed | Eq. (17) of arXiv:1511.01007 as printed returns -3.2955 against the stated 2.1 SNU at zero displacement, and doubling its Var(X_B,lin) term, as Eq. (B14) and Qin's thesis Eq. (7.10) do, matches sat_estimate to 1e-12 at every displacement. | ✅ pass |