{"id":"75cc278a-83bc-4915-af6e-c20fe757249f","arxiv_id":"2505.16615","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Continuous feedback control described by the Quantum Fokker-Planck Master Equation is given a first law and a fluctuation theorem in which detector delay produces a new measurement entropy term.","lead":"This paper derives a thermodynamics framework for quantum systems under continuous measurement and feedback, with explicit definitions of work, heat, and measurement energy, plus a fluctuation theorem containing a new quantity called measurement entropy. It applies the framework to a bang-bang controlled two-level system and a measurement-driven qubit engine.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantum coarse-grained FT lacks independent verification: σ_m,cg in Eq. (51) is a signed-quasi-probability integral whose positivity is assumed, and Fig. 6 computes it by inverting the FT.","rationale":"The first-law derivation for the QFPME (Eqs. (9)-(12)) is explicit and convincing: the stochastic calculus in App. A correctly accounts for the Itô correction, and the two examples (bang-bang control and the measurement-driven engine) are consistent with the exact steady-state solutions. The classical detailed and integral FTs, Eqs. (38)-(40), are also carefully derived in App. D; the factorization of the joint path probability and the calculation of the measurement entropy via the Ornstein-Uhlenbeck filter are mathematically sound, and the numerical checks in Figs. 4-5 are non-circular because σ_m is computed directly from Eq. (40). The soft spot is the extension to quantum systems. There, the detailed FT (49) is stated for a Keldysh quasi-probability that may be negative, and the physically meaningful coarse-grained FT (52) involves σ_m,cg defined by Eq. (51) as an average of e^{-σ_m} over a signed measure. The paper's Fig. 6 does not verify Eq. (52) independently: σ_m,cg is obtained by inverting the FT, so the histograms are consistent by construction. The reader's conditional verdict is appropriate: the framework is promising and the derivations are mostly explicit, but the quantum claim needs an independent consistency test before it can be regarded as established. Our proposed test—computing σ_m,cg from its definition—directly addresses this gap. Agreement with the reader is partial: the reader identified the self-referential numerical check and the signed quasi-probability machinery, but framed the weakest assumption as the classical backward experiment; we consider the quantum coarse-grained FT the more load-bearing point.","tokens_in":43075,"tokens_out":21253,"duration_ms":168491,"concrete_test":"For the continuous-measurement-driven engine of Sec. III F with the parameters of Fig. 6 (γ=10κ=5g=5λ=ω, ω=k_B T), compute σ_m,cg[D] directly from the defining integral in Eq. (51) by sampling the Keldysh quasi-probability P[Γ,a_c,D] (or using the steady-state solution of App. C.3), and compare with the values inferred from the detailed FT, Eq. (52). Check that the integral is positive for all sampled trajectories and that ⟨e^{-σ-σ_m,cg}⟩=1 holds with the independently computed values. Release the simulation code for reproducibility.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantum coarse-grained fluctuation theorem, Eq. (52), is the basis for the claimed quantum second law, Eq. (53). Its central object, σ_m,cg[D], is defined in Eq. (51) as the average of e^{-σ_m[a_c,D]} over the conditional quasi-probability P[a_c|Γ,D]=P[Γ,a_c,D]/P[Γ,D]. Because P[Γ,a_c,D] is a real signed measure that can take negative values (Sec. IV C 1), this integral is not guaranteed to be positive; if it were negative, σ_m,cg would be complex and Eq. (53) would be ill-defined. The derivation of Eq. (52) from Eq. (49) implies positivity, but the only numerical evidence, Fig. 6, computes σ_m,cg by inverting Eq. (52) itself, so the agreement is circular and cannot detect a failure of the integral to match. The analytical derivation in App. F is plausible but does not independently establish positivity of the integral or verify Eq. (52) without assuming it. This is load-bearing because the quantum-regime claim that a finite-bandwidth detector introduces a delay-induced entropy term in fluctuation theorems rests entirely on Eq. (52).","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a thermodynamics framework for open quantum systems under continuous measurement and feedback described by the Quantum Fokker-Planck Master Equation (QFPME). It derives a first-law decomposition of the average energy change into power, heat current, and a measurement-induced energy rate, Eqs. (9)-(12), and illustrates it on a bang-bang two-level system and a continuous-measurement-driven engine. It then derives a classical fluctuation theorem, Eq. (38), in which a new stochastic term called measurement entropy, Eq. (40), accounts for the detector's finite bandwidth and delay, together with the resulting second law, Eq. (42). The framework is extended to quantum systems using Keldysh quasi-probabilities, leading to the coarse-grained fluctuation theorem Eq. (52) and the quantum second law Eq. (53).","tokens_in":43324,"tokens_out":8361,"duration_ms":78519,"significance":"If correct, the paper provides a consistent thermodynamic description for continuous feedback with a finite-bandwidth detector, and it identifies a delay-induced measurement entropy that is distinct from the information terms previously studied in feedback fluctuation theorems. The main strengths are the detailed derivations in Apps. A, D, and F, the consistency check of the fast-detector limit against Eq. (47) from Ref. [50], and the two worked examples with Monte-Carlo simulations. The quantum extension, however, rests on signed quasi-probabilities, and the paper does not explicitly prove the positivity that the quantum second law requires; the numerical demonstration in Fig. 6 also computes the coarse-grained measurement entropy by inverting the fluctuation theorem rather than by evaluating its defining integral. These issues are local and repairable, but they affect the quantum-regime claim.","major_comments":[{"comment":"Eq. (51) defines e^{-sigma_m,cg[D]} as an integral of e^{-sigma_m[a_c,D]} over the conditional quasi-probability P[a_c|Gamma,D]=P[Gamma,a_c,D]/P[Gamma,D]. Since P[Gamma,a_c,D] is a signed measure that can take negative values, this integral is not manifestly positive, and the second law in Eq. (53) requires sigma_m,cg to be real. Positivity does follow if one integrates Eq. (49) over a_c, because then e^{-sigma_m,cg[D]} equals P[barGamma,barD]e^{sigma[Gamma,D]}/P[Gamma,D], which is positive, but this argument is not given. Moreover, the numerical evidence in Fig. 6 obtains sigma_m,cg by inverting Eq. (52), not by evaluating the integral in Eq. (51), so it cannot detect a failure of the defining integral to match the fluctuation theorem. Please add an explicit proof that the integral in Eq. (51) is positive and that P[Gamma,D] is a well-defined nonnegative probability, or provide an independent numerical evaluation of Eq. (51).","section":"§IV.C-D and Fig. 6"}],"minor_comments":[{"comment":"The text near Eq. (21) writes 'Heavyside step function'; this should be 'Heaviside step function'.","section":"§III.E"},{"comment":"In the paragraph after Eq. (38), 'time-revered versions' should read 'time-reversed versions'.","section":"§IV.A.1"},{"comment":"In the discussion following Eq. (14), 'the terms on the right-hand site' should read 'the terms on the right-hand side'.","section":"§III.C"},{"comment":"The sentence 'Here, each D_n≥1 is contingent upon ...' is confusing; presumably the intended statement is about indices n≥1, not about the values of D_n.","section":"App. D.1"},{"comment":"The parameter line 'gamma=10kappa=5g=5lambda=omega' is ambiguous; please list the five parameters separately, and report statistical uncertainties for the Monte-Carlo histograms.","section":"Fig. 6 caption"},{"comment":"In App. E.1, 'classical countertpart' should read 'classical counterpart'.","section":"App. E"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely correct in its classical parts, where the derivations in Apps. A, D, and E are detailed and the fast-detector limits match Ref. [50]. The quantum section is the weakest point: the coarse-grained measurement entropy is defined through a signed quasi-probability, and the manuscript does not explicitly resolve the positivity issue, while the only numerical check inverts the fluctuation theorem itself. I would ask the authors to provide a direct proof of positivity of Eq. (51) or an independent numerical evaluation before publication. The extensive reliance on the authors' own QFPME is not improper, since it is independently published work, but the novelty of the quantum claims should be stated carefully."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things you should know. The classical part is a real contribution: a clean first law for QFPME-based feedback with a separate measurement-energy term, and a detailed fluctuation theorem where the detector's finite bandwidth enters as a measurement entropy. The derivations in the appendices are careful, and the Monte Carlo checks in Figs. 4 and 5 verify the integral FT and the second law. The fast-detector limit reproducing Ref. [50]'s Eq. (47) is a good sanity check.\n\nThe soft spot is the quantum extension. Eq. (52) is the basis for the quantum second law, but its central object sigma_m,cg is an average over a signed quasi-probability P[ac|Gamma,D]. The paper never proves that integral is positive, and if it could be negative the coarse-grained entropy is not well-defined. The only numerical evidence, Fig. 6, computes sigma_m,cg by inverting the detailed FT rather than evaluating Eq. (51) directly, so the agreement is circular. The appendix derivation gives plausible structure but does not close this gap. On top of that there is no simulation code, so reproducing the quantum figures would take real effort. These are load-bearing because the quantum claim that a delay-induced entropy must appear in fluctuation theorems rests entirely on Eq. (52).\n\nThe paper positions the measurement entropy as distinct from the information terms in the earlier feedback-fluctuation-theorem literature, and that distinction is argued clearly. The work builds on the authors' own QFPME, which is independent published work, so the self-citation is not a problem.\n\nWho is this for? Anyone working on thermodynamics of continuous measurement and feedback, especially people using the QFPME. The classical first law and FT are the parts I would trust and likely use. The quantum part should be treated as a conjecture until the coarse-grained measurement entropy is computed directly and the positivity issue is addressed.\n\nMy recommendation: send it to peer review, but with a strong request that the authors either prove positivity of sigma_m,cg or restrict the quantum second law to a regime where it holds, and that they release code so the quantum FT can be checked without inverting the relation. This deserves referee time because the framework is useful and the classical results are solid.","headline":"Classical first law and fluctuation theorem are solid and carefully derived; the quantum extension rests on an unverified coarse-grained measurement entropy whose only numerical check is circular.","tokens_in":43830,"tokens_out":2177,"would_cite":true,"duration_ms":20533,"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":"The paper derives exact thermodynamic first and second laws for open quantum systems under continuous measurement and feedback, introducing a new stochastic quantity, measurement entropy, that captures the irreversibility of a…","keywords":["continuous feedback control","quantum thermodynamics","fluctuation theorems","measurement entropy","Quantum Fokker-Planck Master Equation","measurement backaction","Szilard engine","Maxwell's demon"],"falsifier":"In a two-level system under threshold (bang-bang) feedback, vary the detector bandwidth $\\gamma$ and measurement strength $\\lambda$, reconstruct the joint trajectory $(a_t,D_t)$ from the measurement record, compute $\\sigma_m$ via Eq. (40), and test the integral fluctuation theorem $\\langle e^{-\\sigma-\\sigma_m}\\rangle = 1$ and the bound $\\langle\\sigma\\rangle \\ge -\\langle\\sigma_m\\rangle$; the theorem should fail if the actual detector memory deviates from the Ornstein-Uhlenbeck low-pass filter.","tokens_in":42893,"feed_emoji":"⚛️","tokens_out":4844,"duration_ms":40760,"temperature":0.7,"pith_summary":"The paper aims to put the thermodynamics of continuous measurement-and-feedback control on the same footing as ordinary open quantum systems, including the experimentally realistic case of a detector with finite bandwidth. It derives an exact first law that splits energy change into power, heat, and a measurement-energy term, and it derives fluctuation theorems containing a new stochastic quantity, measurement entropy, which quantifies irreversibility caused by the detector's delay. If the paper is right, second-law bounds for feedback-controlled engines must be corrected by this entropy, and measurement backaction can be a legitimate source of work.","feed_headline":"Detector delay adds a new entropy to fluctuation theorems","feed_subtitle":"Finite-bandwidth continuous feedback obeys a corrected first and second law, with measurement backaction as a fuel.","key_machinery":"The central object is the Quantum Fokker-Planck Master Equation (QFPME), which describes the joint state of the system and a detector outcome $D$ evolving as a low-pass-filtered (Ornstein-Uhlenbeck) measurement with bandwidth $\\gamma$ and measurement strength $\\lambda$. The load-bearing new identity is the measurement entropy $\\sigma_m$, defined as the log-ratio of forward and backward detector-trajectory probabilities, which equals $\\frac{8\\lambda}{\\gamma}\\int_0^\\tau (a_t - D_t)\\, dD_t - \\gamma\\tau - \\ln\\frac{P_{\\rm ini}[D_\\tau|a_\\tau]}{P_{\\rm ini}[D_0|a_0]}$ in the continuous-time limit. In the quantum extension, Keldysh quasi-probability trajectories carry the argument, with only the classical branch entering $\\sigma_m$, so that integrating out the quantum branch yields a coarse-grained fluctuation theorem for the experimentally accessible detector trajectory.","core_discovery":"For any system described by the Quantum Fokker-Planck Master Equation, the rate of energy change separates exactly as $\\partial_t U = P + J + \\dot{E}_M$, with power, heat current, and measurement-energy rate given by explicit formulas involving the detector outcome $D$, the measurement strength $\\lambda$, and the detector bandwidth $\\gamma$. The paper further derives a detailed fluctuation theorem $P[\\bar{a}, \\bar{D}]/P[a,D] = e^{-\\sigma - \\sigma_m}$ whose new term, the measurement entropy $\\sigma_m$, is an Itô integral along the detector trajectory and measures the asymmetry between measurement outcomes that lag behind the system and those that anticipate it. The same structure is extended to open quantum systems using Keldysh quasi-probabilities, yielding a coarse-grained fluctuation theorem that depends only on the detector trajectory and a generalized second law $\\langle\\sigma\\rangle \\ge -\\langle\\sigma_m\\rangle$.","pith_inferences":["A testable prediction not spelled out by the paper: the rate of measurement entropy equals $8\\lambda$ times the excess mean squared detector error relative to the static-detector variance, so directly measuring $\\langle(D_t-a_t)^2\\rangle$ should expose the second-law correction without trajectory reconstruction.","The framework suggests that continuous feedback with a finite-bandwidth detector carries an information cost encoded in $\\sigma_m$; optimizing work extraction may amount to trading output power against this entropy cost.","If the detector filter is not a first-order low-pass (for example, a high-pass or non-Markovian filter), the measurement-entropy expression would need replacement and the fluctuation theorem would likely change; the paper explicitly leaves such extensions open.","For quantum systems, $\\sigma_m$ depends only on the classical Keldysh branch, hinting that the measurement-entropy correction is essentially classical even when work extraction exploits coherence; a coherent-work experiment could isolate that part."],"forward_implications":["For QFPME dynamics the first law holds exactly with three terms, so energy-balance analyses of continuously measured engines must include the measurement-energy rate $\\lambda\\langle\\mathcal{D}[\\hat{A}]\\hat{U}\\rangle$.","Fluctuation theorems for feedback with finite-bandwidth detection must include $\\sigma_m$; the standard fluctuation theorem without it is violated, as the paper's Monte Carlo simulations show.","The average entropy production can be negative under feedback, but it is bounded below by $-\\langle\\sigma_m\\rangle$, giving a generalized second law of thermodynamics.","Measurement backaction can act as a fuel: the continuous-measurement-driven engine converts $\\dot{E}_M$ into work and can even absorb heat from the environment for some parameters.","Coarse-graining yields a fluctuation theorem involving only the detector trajectory, so the second-law bound is testable without knowing the hidden system trajectory."],"supporting_citations":[{"why":"Introduces the Quantum Fokker-Planck Master Equation on which the entire formalism is built.","marker":"[50]"},{"why":"Provides the feedback-control fluctuation theorem whose information term the new measurement entropy replaces.","marker":"[98]"},{"why":"Gives a detailed fluctuation relation for arbitrary measurement and feedback schemes and supplies the comparison for time-reversal assumptions.","marker":"[99]"},{"why":"Offers the repeated discrete-feedback fluctuation theorem as a classical baseline.","marker":"[104]"},{"why":"The measurement-driven engine whose continuous version is used as the quantum example.","marker":"[69]"},{"why":"Defines quantum heat / measurement energy that motivates the $\\dot{E}_M$ term.","marker":"[106]"},{"why":"Provides the Keldysh quasi-probability formalism used for the quantum fluctuation theorem.","marker":"[124–128]"},{"why":"Supplies the quantum-jump trajectory entropy production that defines $\\sigma[\\Gamma,D]$.","marker":"[36–39]"}],"fun_headline_variants":["Measurement entropy emerges in continuous feedback thermodynamics","Feedback loops add a new entropy to quantum thermodynamics","Corrected second law from continuous measurement and feedback","Measurement backaction supplies fuel in quantum feedback engines"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The backward experiment is taken to be the exact time reversal of the forward one, with feedback depending only on the current detector outcome; combined with the QFPME's Gaussian, low-pass-filtered noise model, this assumption is what lets detector lag be reinterpreted as a new entropy rather than as an uncontrolled error.","fun_headline_variants_meta":{"raw":{"variants":["Measurement entropy emerges in continuous feedback thermodynamics","Feedback loops add a new entropy to quantum thermodynamics","Corrected second law from continuous measurement and feedback","Measurement backaction supplies fuel in quantum feedback engines"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000176,"raw_usage":{"total_tokens":1242,"prompt_tokens":848,"completion_tokens":394,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":464,"completion_tokens_details":{"reasoning_tokens":337}},"tokens_in":464,"tokens_out":394,"duration_ms":3589,"temperature":1.0,"reasoning_tokens":337,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:57:53.892595+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In a two-level system under threshold (bang-bang) feedback, vary the detector bandwidth $\\gamma$ and measurement strength $\\lambda$, reconstruct the joint trajectory $(a_t,D_t)$ from the measurement record, compute $\\sigma_m$ via Eq. (40), and test the integral fluctuation theorem $\\langle e^{-\\sigma-\\sigma_m}\\rangle = 1$ and the bound $\\langle\\sigma\\rangle \\ge -\\langle\\sigma_m\\rangle$; the theorem should fail if the actual detector memory deviates from the Ornstein-Uhlenbeck low-pass filter.","supporting_citations":[{"cited_title":"Ito and T","cited_arxiv_id":null,"evidence_quote":"Gives a detailed fluctuation relation for arbitrary measurement and feedback schemes and supplies the comparison for time-reversal assumptions."},{"cited_title":"Koski, V","cited_arxiv_id":null,"evidence_quote":"Offers the repeated discrete-feedback fluctuation theorem as a classical baseline."},{"cited_title":"Barker, M","cited_arxiv_id":null,"evidence_quote":"Defines quantum heat / measurement energy that motivates the $\\dot{E}_M$ term."}],"review_version":1}