SWSN SCIENTIFIC REVIEW SERIES SWSN-SCI/ORP-2035-04-R3
2. a correlated gravimetric or orbital residual with the predicted sign and lag structure;
3. an astrophysical propagation anomaly whose detector-to-detector timing is consistent
with the same sky region but inconsistent with the best-fit general-relativistic source
waveform.
Gravitational-wave signals encode compact-object masses, spins, sky location and distance
when compared with waveform models [7]. We therefore propose residual analysis after
standard parameter estimation, rather than treating a black hole as a continuously broad-
casting beacon.
A minimal delayed-path phenomenology is
˜hobs(f) = ˜hGR(f)
[
1 + αW (f)e2πif ∆t
]
+ ˜n(f), (4)
where α, ∆t and W (f) are to be constrained by data. A statistically insignificantα across the
predicted band counts against H2.
5 From a quantum arrow of time to inverse control
For a Markovian open quantum system, a standard Lindblad description is
˙ρ = − i
¯h[H(t), ρ] +
∑
j
(
CjρC†
j − 1
2{C†
j Cj, ρ}
)
, (5)
with density operatorρ, HamiltonianH and collapse operatorsCj. Numerical toolkits such
as QuTiP implement these equations for open-system simulation [5, 6].
The 2026 result of García-Pintos et al. is trajectory-level and measurement-conditioned [1].
Our extrapolation preserves that conceptual restriction. MCSR is formulated as an inverse-
control objective
J[u] = 1 − F
(
ρu(T ), ρtarget
)
+ λ
∫ T
0
∥u(t)∥2dt + µ R[u], (6)
where F is state fidelity,u(t) is a control field andR penalizes unsafe or nonrobust control.
This is not a global inverse of entropy production. It is a search for a reachable state under
constrained dynamics.
5.1 Conditional phase-noise toy model
We tested the simplest transparent case: an initial|+x⟩ qubit accumulates a stochastic phase
ϕ. Without access to the noise record, ensemble averaging dephases the state. With a con-
ditional estimate of the phase, feedback removes part of the accumulated rotation. Across
60,000 trajectories, the uncontrolled mean fidelity is 0.550. At effective record efficiency
η = 0 .85 and unit feedback gain, mean fidelity is 0.854. Perfect record access in this toy
model returns the state exactly. Figure3 shows the full gain-efficiency surface.
The operational lesson is modest but important: trajectory information changes what control
is possible. If a macroscopic restoration protocol exists at all, it must preserve a sufficiently
rich measurement history rather than infer the past from a final state alone.
CONTROLLED SCIENTIFIC DOCUMENT Internal review copy Page 5 of9