{"id":"baa6da75-b43a-45d0-829a-8a2d74907a79","arxiv_id":"1908.00502","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For a free Brownian particle in contact with multiple reservoirs, each bath contributes a fraction gamma_k T_k / (gamma T) of the particle's energy change during both energy-gaining (upside) and energy-losing (downside) events.","lead":"This paper derives exact formulas for how much heat a free Brownian particle exchanges with each of several heat baths, separating the particle's energy-gaining and energy-losing events. It shows each bath's share of the particle's energy change during both activation and relaxation is simply proportional to that bath's friction and temperature product.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Per-bath factorization of restricted noise-velocity correlations (Eqs. 49-50) is asserted but not rigorously proved; Appendix B's linear-decomposition argument does not cover the nonlinear conditioning, and the master-equation corroboration contains sign errors.","rationale":"The reader's weakest_assumption identifies the same load-bearing concern: the per-bath factorization of restricted noise-velocity correlations. My analysis confirms that this is the critical step for the energy-partition result, and that Appendix B does not rigorously prove it because the conditioning is nonlinear. However, I found no independent error in the heat-current formulas Q_up_k and Q_down_k: they are internally consistent, and the simulation comparisons in Figs. 2 and 4 support them. The remaining vulnerability is specifically the decomposition of the total restricted correlation into per-bath terms, which is needed for the partition ratio. The master-equation section contains typos and concerns a different discrete-level system, so it does not provide independent support. These findings leave the reader's CONDITIONAL verdict unchanged: the core heat-transfer formulas appear solid, but the energy-partition claim is conditional on a factorization that is likely true yet not rigorously established. The proposed simulation test would settle the issue directly.","tokens_in":21132,"tokens_out":25386,"duration_ms":251333,"concrete_test":"Simulate the N=3 Brownian system with three distinct temperatures and frictions, and compute the restricted average <xi_k(t') v(t')>_up / (gamma_k T_k) for each bath k over the interval [0,t] with the condition E(t) > E(0). If Eqs. (49)-(50) are correct, the three normalized curves must coincide for all t' and equal the integrand of the common D_up(t). If they differ, the factorization fails and the energy-partition ratio is unsupported. This test isolates the load-bearing assumption without relying on the unmeasurable heat-current decomposition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central energy-partition result R_up_DeltaE_k = R_down_DeltaE_k = gamma_k T_k / (gamma T) rests on the per-bath factorization of the restricted noise-velocity correlation: the paper writes ∫ m <xi_k v>_up dt' = k_B gamma_k T_k D_up(t) (Eq. 49) and the analogous downside expression. This is what converts the total restricted correlation (Eq. 45) into per-bath contributions, and it underlies the linear decomposition of <Delta E> in Eqs. (87)-(88) and (94)-(95). Appendix B attempts to justify the decomposition with a linear-function argument, but the conditioning E(t) > E(0) is nonlinear in the Gaussian noises. The argument that g_l -> 0 when T_l -> 0 for l != k does not exclude cross-temperature terms at finite T_l, so the proof of equality f_k = g_k is incomplete. The corroborating master-equation section (Sec. IV.C) does not close this gap: Eqs. (97)-(100) contain sign and bracket errors (the steady-state condition as written would make P(E) negative), and the model treats discrete two-level transitions rather than the continuous Brownian dynamics used in the main derivation. Because the paper itself notes that the energy partition cannot be measured directly, the only support for R is this unproven factorization. The claim is plausibly correct (a Gaussian regression argument would likely establish it), but as written the proof is incomplete.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript considers a free Brownian particle driven by N thermal baths at different temperatures and uses the upside/downside trajectory classification of the companion paper (Ref. 30) to split, at time t, the ensemble into trajectories whose energy is above or below a threshold (the initial energy E(0) or the average energy <E>). It derives restricted heat currents Q^up_k and Q^down_k (Eqs. (51)-(52) for E^‡=E(0), Eqs. (71)-(72) for E^‡=<E>), restricted energy fluxes, flux ratios, heat ratios, and claims the energy-partition ratio R^up_ΔE_k = R^down_ΔE_k = γ_k T_k/(γ T) (Eqs. (86), (93)). The heat-current formulas are validated against Euler-Maruyama simulations for N=2; the paper explicitly notes that the energy-partition ratio itself cannot be measured directly in simulation.","tokens_in":21435,"tokens_out":13040,"duration_ms":135832,"significance":"If the underlying factorization assumptions hold, the paper provides the first closed-form per-bath decomposition of heat into a heat-current part and a system-energy-change part for conditioned Brownian trajectories, together with a strikingly simple, parameter-free partition ratio. The derivation has no fitted parameters, and the unrestricted heat currents reduce to the known Lebowitz result. The simulation agreement for Q_k and the flux ratios in Figs. 2, 4, and 6-9 is a genuine strength. The main limitation is that the central partition ratio rests on the per-bath factorization of the restricted noise-velocity correlation, which is asserted rather than rigorously proved, and the paper itself states that this ratio cannot be measured in simulation. The significance is therefore conditional on closing that proof gap, although the result is plausibly correct and a Gaussian-conditioning argument would likely establish it.","major_comments":[{"comment":"The load-bearing step is the per-bath factorization of the restricted noise-velocity correlation, Eqs. (49)-(50): ∫ m <ξ_k v>_up dt' = k_B γ_k T_k D_up(t) and the analogous downside expression. Appendix B attempts to justify this by a linear-function decomposition, but the argument assumes that the restricted averages are linear in the bath temperatures with no cross-temperature terms. The nonlinear conditioning E(t) > E(0) (or E(t) > <E>) is a function of the Gaussian noises, and the limit T_l → 0 does not exclude cross-temperature terms at finite T_l; the condition g_l → 0 in that limit only fixes the behavior on a low-dimensional boundary. Because Eqs. (51)-(52), (71)-(72), and the energy-partition ratios (86), (93) all follow from this factorization, the proof is incomplete as written. I ask the authors to supply a rigorous derivation (for example, by Gaussian regression/conditioning on v(t) and v(0), which shows the coefficient of T_k is independent of the other bath temperatures) or to state the factorization explicitly as a conjecture and mark the downstream results as conditional.","section":"Sec. III.B, Eqs. (45)-(50), and Appendix B"},{"comment":"The derivation of the central ratio R^up_ΔE_k = R^down_ΔE_k = γ_k T_k/(γ T) proceeds by 'identifying the terms ... proportional to a temperature gradient between baths, i.e., Q^(hc)_k ∝ J_ss_k' and subtracting them to obtain <ΔE_k>. This identification is not independently derived; <ΔE_k> is defined only through the per-bath split of the noise-velocity correlation used in Eqs. (49)-(50). In other words, Eq. (86) is a restatement of the assumed factorization rather than an independent consequence of energy conservation. Since the paper explicitly states that the energy ratio cannot be measured in simulation, the analytical proof is the only support for this claim, and it needs to be made rigorous.","section":"Sec. IV.A-IV.B, Eqs. (84)-(86) and (93)"},{"comment":"The master-equation corroboration contains sign and notation errors. In Eq. (97) the loss term should carry a minus sign, and the bracket in γ_k [(n_k(ΔE)+1] P(E) is unbalanced; as written, combining Eq. (97) with the steady-state condition Eq. (98) gives P(E) < 0 for positive rates. Eq. (100) should be written as (1-P(E))/P(E) = Σ γ_k n_k e^{ΔE/k_B T_k} / Σ γ_k n_k, which follows from the corrected steady-state balance. Please fix these errors and clarify whether this section is intended as a rigorous derivation for general systems or as a heuristic analogy; in its present form it cannot support the statement that the result 'appears to be valid for a robust class of systems.'","section":"Sec. IV.C, Eqs. (97)-(101)"}],"minor_comments":[{"comment":"In the definition of <ΔE>_↓, the second term on the right-hand side conditions on E(t) > E^‡; it should condition on E(t) < E^‡.","section":"Sec. III.B.1, Eq. (33)"},{"comment":"The second displayed ratio repeats R^up_ΔE_k; it should be R^down_ΔE_k = <ΔE_k>_↓ / <ΔE>_↓.","section":"Sec. IV, Eq. (79)"},{"comment":"The bracket in γ_k [(n_k(ΔE)+1] P(E) is missing a closing parenthesis or bracket; please correct the typesetting.","section":"Sec. IV.C, Eq. (97)"},{"comment":"It would help to state explicitly that the restricted second velocity moments are for the case t' < t with the upside/downside constraint imposed at t, since the notation is otherwise easy to misread.","section":"Sec. III.B, Eqs. (40)-(41) and (65)-(66)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid follow-up in the authors' upside/downside series, and the heat-current formulas are well supported by simulation. The decisive issue is the missing rigorous proof of the per-bath factorization of the restricted noise-velocity correlation; this is a correctable gap (a Gaussian conditioning argument should suffice) rather than a fundamental error. The master-equation section also needs correction. I recommend major revision rather than rejection, contingent on the authors supplying a rigorous derivation of Eqs. (49)-(50)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a serious paper that deserves peer review, but the marquee result—the gamma_k T_k / gamma T energy-partition ratio—is not as firmly established as the paper claims. The heat-current formulas (Eqs. 51-52 and 71-72) are new, clean, and backed by simulation; the partition ratio follows from a per-bath factorization of restricted noise-velocity correlations that is plausible but not rigorously proven.\n\nWhat is new: the paper applies the authors' upside/downside formalism to heat transfer between multiple baths, derives restricted heat currents for two thresholds (E(0) and <E>), and extracts the additive partition rule. The D factors are derived from energy conservation rather than fitted, and the unrestricted limits reproduce Lebowitz. Simulation support for the Q_k curves is convincing, including the N>2 check mentioned in the text.\n\nThe soft spot is real. Equations (49)-(50) assert integral m<xi_k v>_up dt' = k_B gamma_k T_k D_up(t) with a bath-independent D. Everything about the partition ratio depends on it. Appendix B tries to justify the decomposition with a linear-function argument, but the conditioning E(t)>E(0) or E(t)><E> is nonlinear in the Gaussian noises, and the T_l->0 limit does not exclude cross-temperature terms at finite T_l. I think the factorization is actually correct—a Gaussian regression argument would likely prove it—but the paper does not supply that proof. The energy ratio is not directly measurable, as the paper honestly notes, so this gap leaves the central claim conditional.\n\nThe master-equation section IV.C has a sign error: Eq. (97) should have a minus sign before the second term, and Eq. (98) as written would make P(E) negative. The final result (101) looks like the correct detailed-balance expression, so it is probably a transcription slip, but as printed the derivation does not make sense. That section is also a discrete two-level model, so it is corroboration by analogy rather than a derivation for the Brownian system.\n\nMinor: no simulation code or error bars; for a paper where the key ratio cannot be measured directly, that leaves the simulation evidence thinner than it could be. The Q_k agreement is good enough to carry the heat-current part.\n\nBottom line: the heat currents are likely solid and publishable; the partition rule is probably right but needs a proper proof or an explicit caveat. I would send this to a referee who knows Gaussian conditioning, and expect a revise rather than an accept.","headline":"Solid upside/downside heat-current results, but the energy-partition ratio rests on a plausible yet unproved per-bath factorization, and the master-equation check has a sign error.","tokens_in":21968,"tokens_out":7669,"would_cite":true,"duration_ms":75206,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-14T15:52:26.101309+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}