{"id":"733d729a-1cb7-4a90-b8ee-caa820b204db","arxiv_id":"2507.01934","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A general deterministic feedback master equation is derived, unifying existing schemes and enabling time-dependent feedback based on the last quantum jump and the time since it occurred.","lead":"This paper derives a single deterministic master equation that covers many types of feedback control in open quantum systems, including time-dependent protocols based on quantum-jump detections. The framework unifies previously known feedback equations and yields analytical steady-state results for population inversion and transition reversal.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The continuous-measurement limit in Result 2 is only formal: ϱ_t(k,0) is an implicitly rescaled rate, and Eq. (8) uses a nonstandard half-boundary delta convention; unless prefactor and convention are matched, the renewal equation and Eq. (12) develop a factor-2 initial-condition error.","rationale":"The reader's weakest assumption already targets the δt→0 limit and the nonstandard boundary delta convention in Eq. (8). My reading confirms that this is the most load-bearing technical step: Result 1 is a straightforward discrete identity with a one-line proof, but Result 2 is the genuinely new claim, and every downstream analytical result—Eq. (12), the steady-state eigenvalue equation (13), and both worked examples—inherits the integrity of the limit. The concern is not that the final equations are known to be wrong; the special no-feedback case reduces correctly to the Lindblad master equation, and the renewal structure is plausible. The issue is that the paper's notation hides a change of interpretation: the discrete ϱ_n(k,0), which is a vanishing probability mass, becomes a finite rate in the continuum, and the initial-condition delta is given a boundary convention whose factor 2 is introduced without a derivation from the limiting process. Because this is a formal step with no independent numerical or machine-checked verification, it is appropriate to keep the verdict conditional rather than unconditional. The concrete test proposed above would settle the concern by determining whether the factor-2/half-mass convention is exactly the one selected by the discrete stroboscopic limit, and by checking Lindblad recovery in the simplest case. Until that check is performed, the central claim of Result 2 rests on an unproven convention, so the reader's CONDITIONAL verdict should be maintained.","tokens_in":32149,"tokens_out":16524,"duration_ms":198756,"concrete_test":"Run the discrete recursion (A35)-(A38) for a single monitored jump channel with no feedback (spontaneous emission at rate γ), using δt = 10^-1, 10^-2, and 10^-3 of 1/γ. Define the rescaled boundary value a_n(k)=ϱ_n(k,0)/δt and the age density ϱ_n(k,hδt)/δt. Compare the δt→0 limit with two continuous descriptions: (i) the renewal equation a_t(k)=δ_{k,bar k}δ(t)ρbar0 + ∫_0^t J(τ)G(τ)a_{t−τ}(k)dτ using the full-mass endpoint convention ∫_0^t δ=1 and no factor 2; and (ii) the paper's Eq. (8) with factor 2 and the half-mass convention. Whichever matches the discrete-to-continuum limit is the correct bookkeeping. Independently, verify that the unconditional state obtained from Eq. (10) under the paper's convention satisfies ∂_t ρbar_t = L ρbar_t exactly in this no-feedback case; failure of this Lindblad recovery would indicate that the continuous limit underlying Result 2 is not sound.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Result 2 is the paper's main technical contribution, yet its derivation from the discrete recursion (A35)-(A38) requires a limit that is only handled formally. In the discrete setting, ϱ_n(k,0) is a probability mass at τ=0 and is O(δt); as δt→0 it cannot converge to the finite object appearing in Eqs. (8)-(9) unless one implicitly rescales by 1/δt and reinterprets ϱ_t(k,0) as a rate rather than a probability mass. The paper does not state this rescaling, so the left-hand side of Eq. (8) has different physical units from the discrete quantity it replaces. Compounding this, Eq. (8) injects the initial condition as 2δ(t)δ_{k,bar k}ρbar0 together with the nonstandard convention ∫_0^t dτ δ(t−τ)=1/2. The factor 1/2 is not forced by the discrete-to-continuous limit; it is a convention about the location of a boundary atom. The equations are internally consistent only if the factor 2 and this convention are used together: inserting Eq. (8) into Eq. (10) under the stated convention yields Eq. (12) with the correct prefactor 1 on G(bar k,t)ρbar0, whereas the more common endpoint convention ∫_0^t δ=1 would produce a spurious factor 2. Because Eq. (12) is the basis of the steady-state equation (13) and of both examples, any ambiguity in this limit propagates directly into the paper's analytical predictions. The paper states the convention but does not derive it from the limit or prove convergence of the stroboscopic protocol, leaving Result 2's mathematical grounding conditional.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a general framework for feedback control of open quantum systems described by sequential instruments. Result 1 gives a deterministic equation for the memory-resolved state for any causal memory. Specializing to quantum-jump monitoring with memories recording the last jump channel and the time since the last jump, Result 2 provides deterministic integral equations (8)-(9) for the resolved state in the continuous-measurement limit. From these, the authors derive a master equation for jump-only feedback, a closed renewal-type equation for the unconditional state, and a steady-state eigenvector equation. Two examples are treated analytically: population inversion of a qubit against a thermal bath and real-time reversal of quantum transitions. The Supplemental Material contains detailed derivations and shows that several previous feedback master equations are special cases of Result 1.","tokens_in":32516,"tokens_out":10322,"duration_ms":114936,"significance":"If the central derivation is valid, the paper makes a strong contribution: it provides a unified starting point from which known feedback master equations (Wiseman-Milburn, Annby-Andersson et al., Kewming et al.) are recovered, and it extends deterministic feedback equations to time-dependent strategies depending on the last jump and the elapsed time. The two worked examples yield analytical steady-state solutions and threshold predictions (e.g., the drive-strength ratio gamma/lambda < 1.145 for population inversion, and the maximum feedback delay) that are directly testable and would be difficult to obtain from stochastic simulations. The recovery of existing results in Supplemental Sec. S.V is an important independent validation of the framework. The paper also gives explicit, experimentally motivated protocols and shows how jump statistics become accessible in the stationary regime.","major_comments":[{"comment":"The passage from the discrete recursion (A35) to the continuous equations (8)-(9) is formal and not fully specified. In the discrete setting, ϱ_n(k,0) is a probability mass concentrated at τ=0, whereas in Eq. (8) the object ϱ_t(k,0) has the character of a density/rate; the paper does not state the implicit rescaling by 1/δt that makes the continuous limit well-defined. Furthermore, the initial-condition term 2δ(t)δ_{k,¯k}ρ̄0 and the boundary convention ∫_0^t dτ δ(t−τ)=1/2 are asserted rather than derived from the δt→0 limit of the discrete recursion. This matters because Eq. (12) - and therefore the steady-state equation (13) and both examples - depends on the factor 2 being exactly compensated by the half-boundary convention. While the convention is explicit and internally consistent, the absence of a derivation leaves a load-bearing gap: a reader cannot verify from the supplemental material that the discrete limit indeed produces this boundary convention, rather than the more common endpoint convention ∫_0^t δ=1, which would introduce a spurious factor of 2. I recommend adding to the Supplemental Material a subsection that derives the δt→0 limit from Eqs. (A35)-(A38) with a precise statement of the rescaling of ϱ_n(k,0) and the origin of the half-boundary convention.","section":"Supplemental S.II.B.1, Eqs. (A35)-(A40); main text Eqs. (8)-(12)"}],"minor_comments":[{"comment":"There is a typo: \"previouslylly known\" should read \"previously known\".","section":"Introduction, second paragraph"},{"comment":"The phrase \"The term2δ(t)δ_{k,¯k}ρ̄0\" is missing a space after \"term\"; it should read \"The term 2δ(t)\".","section":"Main text after Eq. (8)"},{"comment":"In the sentence \"In the limitδt→0this evolves according to...\", there is a missing space between \"limit\" and \"δt\", and later \"time-ordering operator .\" has an extra space before the period.","section":"Result 2 statement"},{"comment":"In \"Fig. 3(b) showsP e as a function of ¯N\", a space is missing after \"shows\".","section":"Fig. 3 caption"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is well within the scope of a quantum physics letters journal and the Supplemental Material is thorough. The central issue is the lack of a rigorous derivation of the continuous-measurement limit and the boundary delta convention in Result 2; this is a load-bearing point because the steady-state equation and the examples rely on it. The authors should be given the opportunity to add a derivation or at least a precise statement of the limit in the Supplemental Material."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Result 2 is the heart of this paper, and it's a good one: a deterministic renewal-type integral equation for feedback protocols that condition on both the channel of the last detected jump and the time since it occurred. I checked the discrete derivation in the supplement, and the structure is sound; the continuous equations (8)-(9) follow from the stroboscopic recursion, and the recovery of earlier feedback master equations as special cases is a genuinely useful service.\n\nThe paper does what it promises. Result 1 is a clean umbrella statement, even if the memory-resolved/hybrid-state starting point is familiar; the unified derivation and the supplemental table are valuable. The two examples are well chosen, with analytical steady states via an eigenvector problem, and the population-inversion threshold and the reversal protocol give concrete results beyond what stochastic simulation would provide easily.\n\nWhere are the soft spots? The continuous-measurement limit is handled formally. The object ϱ_t(k,0) goes from a probability mass in the discrete recursion to a density/rate in the limit, and the paper never says this explicitly. The half-boundary delta convention ∫_0^t dτ δ(t−τ)=1/2 is stated, and the factor 2 in Eq. (8) is matched to it, so the equations are internally consistent, but the convention is not derived from the limit. A rigorous referee could reasonably demand a short argument that the limit is well-defined. I don't think this breaks the central claim; it's a fixable rigor gap, not a factor-of-two error. The abstract's 'all possible feedback schemes' is a bit sweeping, and the γ/λ<1.145 threshold in Example 1 is stated without proof, though that's acceptable in a letter.\n\nThe citation pattern is honest; the authors appropriately credit prior work on hybrid states and feedback master equations. This belongs in the quantum control/thermodynamics conversation and deserves a serious referee. I'd support sending it to review, with a request to clarify the rescaling and the delta convention.","headline":"Solid derivation of a new renewal-type feedback master equation; the continuous-limit convention needs tightening but the central physics and the recovery of prior results are credible.","tokens_in":33017,"tokens_out":4425,"would_cite":true,"duration_ms":49338,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper derives one deterministic equation describing all feedback schemes for sequentially measured open quantum systems, then specializes it to quantum-jump feedback with memory of the last jump and elapsed time, obtaining integral…","keywords":["open quantum systems","feedback control","quantum jumps","memory-resolved state","deterministic integral equations","continuous measurement","feedback master equations","steady state"],"falsifier":"In a photon-counting experiment on a single driven emitter under the Example-2 feedback protocol (pulse applied $\\tau_c$ after each detected jump), record the histogram of waiting times between jumps and compare it with the stationary prediction $\\mathrm{Tr}[\\varrho_{ss}(\\tau)] = \\gamma_B(\\bar{N}_B+1)\\langle B|\\bar{\\rho}_{ss}|B\\rangle \\mathrm{Tr}[G(\\tau)(|G\\rangle\\langle G|)]$ from Eqs. (S6)-(S7); systematic deviation beyond detector resolution would falsify the deterministic equations.","tokens_in":31937,"feed_emoji":"⚛️","tokens_out":7619,"duration_ms":76905,"temperature":0.7,"pith_summary":"The paper claims that feedback control of open quantum systems, where a system is repeatedly measured and the outcome record or a compressed memory of it selects the next control action, can be described by a single deterministic equation for the memory-resolved state. From that equation, every previously known feedback master equation follows as a special case, and new schemes become tractable. The main technical result is a specialization to quantum-jump detection in which the memory stores the channel of the last detected jump and the time elapsed since it occurred; in the continuous-measurement limit, the memory-resolved state obeys closed integral equations rather than a Markovian differential equation. These equations make it possible to find steady states from an eigenvalue problem and to compute jump statistics, which stochastic trajectory simulations handle only slowly and noisily. The value of the paper, if right, is that a broad class of feedback protocols, including time-dependent drives, delays, and pulses, now has a deterministic, analytical description.","feed_headline":"One equation unifies feedback control of open quantum systems","feed_subtitle":"It makes time-dependent jump feedback analytically solvable, giving steady states without noisy simulations.","key_machinery":"The machinery is the memory-resolved state $\\varrho_n(y)=\\mathbb{E}[\\rho_{x_{1:n}}\\delta_{y,y_n}]$, a deterministic density matrix that averages over trajectories conditioned on a causal memory $y_n=f_n(x_n,y_{n-1})$; Result 1 gives its one-step update $\\varrho_{n+1}(y)=\\sum_{x',y'}\\delta_{y,f_{n+1}(x',y')}M_{x'}(y')\\varrho_n(y')$. For quantum jumps, the special memories are the jump memory $k_n=x_n+k_{n-1}\\delta_{x_n,0}$ and the counting memory $\\tau_n=\\delta_{x_n,0}(\\tau_{n-1}+\\delta t)$. Taking $\\delta t\\to 0$ turns the discrete update into the integral equations of Result 2, where the no-jump propagator $G(k,\\tau)=\\mathcal{T}[e^{\\int_0^\\tau ds\\, \\mathcal{L}_0(k,s)}]$ carries the time-dependent feedback applied between jumps.","core_discovery":"The central claim is Result 2: for a system monitored by quantum-jump detection with instruments $M_0\\rho=(1+\\delta t \\mathcal{L}_0)\\rho$ and $M_k\\rho=\\delta t \\mathcal{J}_k\\rho$, and with memory recording the last jump channel $k$ and the elapsed time $\\tau$ since it occurred, the memory-resolved state in the continuous limit $\\delta t\\to 0$ satisfies $\\varrho_t(k,0)=2\\delta(t)\\delta_{k,\\bar{k}}\\bar{\\rho}_0+\\sum_{q\\in\\Sigma}\\int_0^t d\\tau\\, \\mathcal{J}_k(q,\\tau)\\varrho_t(q,\\tau)$ and $\\varrho_t(k,\\tau)=G(k,\\tau)\\varrho_{t-\\tau}(k,0)$, where $G(k,\\tau)$ is the time-ordered no-jump propagator. These equations are deterministic and non-Markovian in form, yet amenable to analytical treatment: the steady state, when it exists, is the eigenvector of the super-operator $\\Omega=\\sum_k\\int_0^\\infty d\\tau\\, G(k,\\tau)\\mathcal{J}_k$ with eigenvalue 1, and the trace $\\mathrm{Tr}[\\varrho_t(k,\\tau)]$ gives the joint distribution of the last jump and the waiting time. The paper further shows that when feedback depends only on the last jump channel, the equation reduces to a set of coupled Lindblad-like equations, and that Result 1, the generic deterministic equation, recovers all earlier feedback master equations as particular choices of memory function and instrument.","pith_inferences":["Because any causal function $f_n$ defines a valid feedback protocol in Result 1, the framework effectively parameterizes the space of feedback strategies by data-processing functions; one could invert it to design a memory that steers the system toward a target steady state.","The integral-equation form suggests that the feedback dynamics are non-Markovian in the memory variable but can be solved by discretizing $\\tau$; this may make large parameter sweeps numerically cheaper than trajectory ensembles.","The renewal structure of quantum jumps means the elapsed-time memory $\\varrho_t(k,\\tau)$ is a direct probe of the feedback's effect on jump statistics; comparing its prediction for waiting-time distributions with photon-counting data would be a sharp laboratory test of the whole framework.","Extending the same memory construction to more general counting variables, such as accumulated charge or time spent in a given state, could yield deterministic equations for an even broader class of feedback protocols, including those with non-Markovian filters."],"forward_implications":["All previously derived feedback master equations, for weak Gaussian measurements with low-pass memory, homodyne diffusion feedback, single-jump feedback, and charge-based feedback, reduce to Result 1, giving a unified basis for the field.","Time-dependent jump feedback protocols that were previously accessible only through stochastic simulations now have deterministic integral equations; optimal drive durations, delay thresholds, and steady-state populations can be computed analytically.","The steady state of a jump-feedback protocol is obtained as the unit eigenvector of $\\Omega$, and the memory-resolved state yields waiting-time distributions and mean jump intervals directly.","For a qubit coupled to a thermal bath, population inversion $P_e>1/2$ is achievable for $\\gamma/\\lambda\\le 1.145$ and bath occupation below a critical value, with maximum feedback delay of order $1/\\gamma$ in the strong-drive regime.","The three-level 'reverting quantum transitions' experiment, previously modeled by stochastic sampling of homodyne trajectories, is described analytically in the photon-counting limit, including the state immediately after the feedback pulse."],"supporting_citations":[{"why":"Supplies the memory-resolved state formalism and the weak-Gaussian-measurement Fokker-Planck equation that Result 1 recovers as a special case.","marker":"[29]"},{"why":"The homodyne diffusion feedback master equation that Result 1 recovers as a particular case.","marker":"[30]"},{"why":"The single-jump feedback master equation that Result 1 extends and that Result 2 generalizes from instantaneous to time-dependent feedback actions.","marker":"[31]"},{"why":"The charge-based feedback master equation that follows directly from Result 1.","marker":"[32]"},{"why":"The experiment of quantum transition reversal that Example 2 models analytically; its stochastic-simulation results are reproduced by the steady-state eigenvector solution.","marker":"[43]"},{"why":"Supplies the quantum-jump instruments ($M_0$, $M_k$) and the renewal-process/jump-statistics toolbox used in Result 2 and the examples.","marker":"[56]"}],"fun_headline_variants":["One equation unifies quantum feedback control","Deterministic equations tame quantum feedback","Quantum jump feedback: single deterministic equation","Feedback control of open quantum systems unified"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The continuous-measurement limit $\\delta t\\to 0$ of the stroboscopic jump instruments, with the boundary convention $\\int_0^t d\\tau\\, \\delta(t-\\tau)=1/2$ for the initial-condition delta, is well-defined and produces the integral equations of Result 2; if that limit or the delta convention is not legitimate, the mathematical grounding of the central result collapses.","fun_headline_variants_meta":{"raw":{"variants":["One equation unifies quantum feedback control","Deterministic equations tame quantum feedback","Quantum jump feedback: single deterministic equation","Feedback control of open quantum systems unified"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000248,"raw_usage":{"total_tokens":1594,"prompt_tokens":1038,"completion_tokens":556,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":654,"completion_tokens_details":{"reasoning_tokens":505}},"tokens_in":654,"tokens_out":556,"duration_ms":6787,"temperature":1.0,"reasoning_tokens":505,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T20:39:25.315362+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In a photon-counting experiment on a single driven emitter under the Example-2 feedback protocol (pulse applied $\\tau_c$ after each detected jump), record the histogram of waiting times between jumps and compare it with the stationary prediction $\\mathrm{Tr}[\\varrho_{ss}(\\tau)] = \\gamma_B(\\bar{N}_B+1)\\langle B|\\bar{\\rho}_{ss}|B\\rangle \\mathrm{Tr}[G(\\tau)(|G\\rangle\\langle G|)]$ from Eqs. (S6)-(S7); systematic deviation beyond detector resolution would falsify the deterministic equations.","supporting_citations":[{"cited_title":"Annby-Andersson, F","cited_arxiv_id":null,"evidence_quote":"Supplies the memory-resolved state formalism and the weak-Gaussian-measurement Fokker-Planck equation that Result 1 recovers as a special case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The homodyne diffusion feedback master equation that Result 1 recovers as a particular case."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The single-jump feedback master equation that Result 1 extends and that Result 2 generalizes from instantaneous to time-dependent feedback actions."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The charge-based feedback master equation that follows directly from Result 1."},{"cited_title":"Catching and Reversing a Quantum Jump Mid-Flight","cited_arxiv_id":"1902.10355","evidence_quote":"The experiment of quantum transition reversal that Example 2 models analytically; its stochastic-simulation results are reproduced by the steady-state eigenvector solution."}],"review_version":1}