{"id":"e363b42b-7d2f-42c2-9941-40f82ef5ee34","arxiv_id":"2412.09744","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For an even-length bosonic Kitaev chain, arbitrarily large loss on every other site leaves the zero-frequency topological amplification exactly unchanged, while odd-sized unit cells undergo a topological transition at a closed-form critical loss rate.","lead":"This paper studies how on-site losses affect the directional exponential amplification in a bosonic Kitaev chain, a model for nonreciprocal quantum amplifiers. It proves that loss on one sublattice leaves the steady-state amplification unchanged, while odd-sized unit cells lose amplification above a critical rate, with implications for loss-resilient amplifier design.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The arbitrary-loss robustness requires exactly zero loss on all even sites; since site N is even, any readout coupling ε>0 restores a critical odd-site loss ~1/ε, so the theorem does not support the sensor claim.","rationale":"The paper's algebraic core is strong: Appendix E gives a parameter-free Dyson induction, and the zero-mode argument in Section V A correctly shows no PBC zero eigenstate for even length with arbitrary odd-site losses. The reader's conditional verdict is appropriate because the exact cancellation depends on the complementary sublattice being completely lossless. I sharpen that concern: the even sublattice includes the output site N, so a realistic readout with any nonzero coupling at N breaks the immunity. Moreover, the L=2 exact condition sqrt(γ1γ2)/2=2Δ shows that for fixed small ε on even sites, sufficiently large odd-site loss always triggers a topological transition. The concrete test with ε=0.1Δ would directly demonstrate this failure of extrapolating to readout-capable settings. I agree with the reader's weakest assumption and recommend keeping the verdict unchanged: the mathematics is correct, but the framing as robust quantum sensing needs qualification.","tokens_in":120054,"tokens_out":18792,"duration_ms":215611,"concrete_test":"Run the L=2 calculation with γ2=ε>0 and γ1=Γ. For ε=0.1Δ, Eq. (27) predicts the PBC winding number ν(E) changes from -1 to 0 when Γ=(4Δ)^2/ε=160Δ. Compute the smallest singular value of M_{x,obc} for a 50-site chain and |χxx[50,1;0]| for Γ just below and above this threshold. If the exponentially small singular value disappears at the predicted Γ, then any finite even-site loss destroys the 'arbitrarily large' odd-site robustness, confirming the concern.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central theorem of Appendix E is algebraically sound: the Dyson induction proves χ_{N-1}[n,1;0] = χ[n,1;0] when γ2=γ4=...=γN=0. However, the physically advertised conclusion is narrower than the abstract and conclusion suggest. The invariance is destroyed by any nonzero even-site loss, and this is not a negligible technicality. For the two-site unit cell, Eq. (27) gives a topological transition when sqrt(γ1γ2)/2 > 2Δ. Thus, for fixed ε>0 on the even sites, arbitrarily large odd-site loss γ1 always drives the system out of the amplifying phase once γ1 > (4Δ)^2/ε. In particular, the last site N is even, so a readout port at site N with γN=ε>0 — required for an output field, since the input-output relation (12) carries a factor sqrt(γN) that vanishes when γN=0 — immediately reintroduces a critical odd-site loss and removes the claimed 'arbitrarily large' robustness. The paper gives no estimate of how small ε must be and no SNR or added-noise analysis, so the statement that the system 'will always exhibit exponential amplification' for arbitrary odd-site losses is true only in the exactly lossless-even-sublattice, unreadable limit. This is a scope limitation in the physical claims, not an error in the algebraic proof.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the robustness of topological amplification in the bosonic Kitaev chain subject to non-uniform on-site Markovian losses. After reviewing uniform loss (γ_c = 4Δ), it analyzes unit cells of length L. For L = 2 it derives the PBC spectrum and shows that the winding number is nonzero when √(γ1γ2)/2 < 2Δ, so an arbitrarily large loss on one sublattice can be compensated by a small loss on the other; when γ2 = 0 the system remains amplifying for all γ1. For arbitrary even-length chains with losses on all odd sites and exactly zero loss on even sites, Appendix E proves by Dyson induction that the zero-frequency susceptibility from site 1 to any even site is exactly independent of the odd-site loss rates. For odd unit-cell lengths with loss on the first site, a topological transition occurs at the critical value given in Eq. (33). The paper concludes that the BKC is a promising quantum sensor with strong robustness to disorder.","tokens_in":120320,"tokens_out":3680,"duration_ms":44785,"significance":"If the advertised claim is taken in its precise form, the paper contains a striking and elegant exact result: the steady-state susceptibility from site 1 to the even sublattice is exactly invariant under arbitrary odd-site losses when the even sites are lossless. The proof in Appendix E is a parameter-free Dyson induction and is a genuine contribution to the understanding of non-Hermitian topological amplification beyond the usual gap-size criterion. The L = 2 critical condition √(γ1γ2)/2 = 2Δ is also clean and useful. The paper is honest about the exactness condition in Section V A, but the abstract and conclusion state the robustness claim without that condition, which is not a negligible presentation detail: it directly affects whether the claimed sensor scenario is physically realizable. The availability of code on Zenodo is a strength.","major_comments":[{"comment":"The headline claim is stated more broadly than the theorem that supports it. Appendix E proves χ_{N-1}[n,1;0] = χ[n,1;0] only under the condition γ2 = γ4 = ... = γN = 0, which is used in the Dyson induction and is stated in Section V A. The abstract and the sentence 'no matter the on-site losses that are introduced on the odd sites, the system whose length is even will always exhibit exponential amplification' omit this condition. Since N is even, the last site is even, so a readout at site N requires γN > 0; the input-output relation in Eq. (12) contains the factor √γN, which vanishes at γN = 0. For the two-site unit cell, Eq. (27) gives a topological transition at √(γ1γ2)/2 = 2Δ, so with a fixed even-site loss ε > 0, arbitrarily large odd-site loss γ1 destroys amplification once γ1 > (4Δ)^2/ε. The abstract and conclusions should therefore state the exact lossless-even-sublattice condition, and the paper should analyze the consequences of a finite readout loss.","section":"Abstract, §V A, Appendix E"},{"comment":"The paper gives no quantitative account of how the robustness degrades when the even sites have a small but nonzero loss, which is the physically relevant case for the sensor claim. The exact cancellation in Appendix E relies on the absence of bath coupling on every even site; it provides no bound on the deviation of the susceptibility for γ_even > 0. For the two-site unit cell Eq. (27) already implies the critical scaling γ1^crit ∝ 1/ε for small even-site loss ε, but the disordered N-site case with finite even-site loss is not treated, and no signal-to-noise or added-noise analysis including the output port is provided. Without such an analysis, the concluding statement that the BKC is robust in 'realistic lossy scenarios' is not supported by the presented results. This is a scope limitation in the physical claims, not an error in the algebraic proof.","section":"§V A, Eq. (27); Conclusion"}],"minor_comments":[{"comment":"The phrase 'very large loss rates which exceed the system's non-Hermitian gap' should be qualified with the condition that the complementary sublattice is exactly lossless; otherwise the reader can infer robustness for small even-site losses, which Eq. (27) shows is not the case.","section":"Abstract"},{"comment":"The ansatz φ3 = 0 is introduced as 'numerically-informed'; since Eq. (B2) and the localization transition analysis rest on this assumption, the derivation is not fully analytic. Please state this limitation explicitly and, if possible, verify Eq. (B2) by direct numerical diagonalization.","section":"Appendix B"},{"comment":"There are typographical errors: 'Tthe retarded susceptibility' should be 'The retarded susceptibility', and 'functon' in the caption of Fig. 8 should be 'function'.","section":"Appendix C and Fig. 8 caption"},{"comment":"The notation for the intra-unit-cell eigenstate is ambiguous: the expression |ϕ⟩ = (1/N)(1, t+∆/ε, 0)^T should specify the normalization constant and the components more clearly, because ε is also used for energy in that appendix.","section":"Appendix B, Eq. (B5)"},{"comment":"The name 'Atland-Zirnbauer' should be 'Altland-Zirnbauer'.","section":"Section II"}],"recommendation":"major_revision","confidential_remarks":"The algebraic core is sound and the exact invariance theorem in Appendix E is a genuinely nice result. The reason I recommend major revision rather than reject is that the advertised robustness claim, as stated in the abstract and conclusion, goes beyond the proven statement: a readout port on the last (even) site necessarily introduces γ_N > 0, and Eq. (27) then restores a finite critical odd-site loss. This is fixable within the manuscript's scope by restating the theorem's condition precisely and adding an analysis of finite even-site loss and readout. The availability of the code on Zenodo is a positive point."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper's real result is the exact invariance: for an even-length bosonic Kitaev chain with zero loss on every even site, the zero-frequency susceptibility from site 1 to any even site is exactly independent of all odd-site loss rates. The proof in Appendix E is a clean Dyson induction, and Appendix C gives a closed-form susceptibility for the two-site unit cell with the critical condition sqrt(gamma1*gamma2)/2 = 2 Delta. These are parameter-free derivations from the stated model, not fits, and the code is on Zenodo. That is a solid piece of work.\n\nWhat is genuinely new is that it goes beyond the uniform-loss and gap-heuristic results in the literature. The paper shows that topological amplification can survive losses far exceeding the non-Hermitian gap in a specific sublattice configuration and gives exact critical values for odd unit cells. The L=1 check recovering the uniform case is a good consistency check.\n\nThe soft spot is in the physical framing, not the math. The invariance requires exactly gamma2=gamma4=...=gammaN=0. Since site N is even, any readout port at site N with gammaN>0 immediately breaks the condition. The stress-test note gets this right: for the two-site unit cell the topological phase survives only while sqrt(gamma1*gamma2)/2 < 2 Delta, so for fixed small even-site loss epsilon, an odd-site loss above (4 Delta)^2/epsilon drives the system out of the amplifying phase. The paper acknowledges the zero-loss-even-sublattice condition in Section V A, but the abstract and conclusion say \"the system whose length is even will always exhibit exponential amplification\" and promise a \"multimode quantum sensor in realistic lossy scenarios\" without this caveat, and without any SNR or added-noise analysis. That is an overstatement. The robustness is exact but lives in an unreadable limit: you cannot couple to the output without restoring a finite critical odd-site loss.\n\nI do not think these flaws sink the paper. The central theorems are correct and the design rule—put all unavoidable loss on one sublattice, keep the other lossless—is useful and non-obvious. But the authors should be asked to either provide a perturbation analysis for small even-site loss or temper the sensor claims.\n\nThis is a paper for researchers working on non-Hermitian photonics and topological amplifiers. It deserves a serious referee. I would recommend sending it out, with a request to revise the abstract and conclusion and ideally add a short section on finite even-site loss.","headline":"The exact sublattice-loss invariance is real and cleanly proven, but the 'arbitrary loss' claim depends on a strictly lossless complementary sublattice that any physical readout would break.","tokens_in":120850,"tokens_out":2099,"would_cite":true,"duration_ms":24456,"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":"Even-length bosonic Kitaev chains keep exponential amplification under arbitrarily large loss on every other site, provided the remaining sites are lossless.","keywords":["bosonic Kitaev chain","topological amplification","non-Hermitian skin effect","directional amplification","dissipation robustness","winding number","quantum sensing","non-Hermitian dynamics"],"falsifier":"Fix an even-length open chain with $t>\\Delta>0$, set $\\gamma_2=\\gamma_4=\\cdots=0$ and $\\gamma_1,\\gamma_3,\\ldots$ to a large common value, then compute the smallest singular value of $M_x^{\\mathrm{OBC}}$ and the normalized steady-state particle number on the last site as $N$ grows; the claim predicts both remain exponentially small and localized for every such loss, so observing the singular value to plateau with $N$ or the particle distribution to spread would falsify it.","tokens_in":119832,"feed_emoji":"⚛️","tokens_out":8472,"duration_ms":89269,"temperature":0.7,"pith_summary":"The paper asks how on-site dissipation affects topological amplification in the bosonic Kitaev chain, whose quadrature dynamics are non-Hermitian and directional. It establishes a sharp asymmetry: a chain of even total length with loss placed only on odd sites keeps its exponential steady-state amplification, and the response from site 1 to any even site is exactly independent of the odd-site loss rates, however large. By contrast, dividing the chain into odd-length unit cells and putting loss on the first site of each cell drives a topological phase transition at an explicit critical loss value. These exact statements matter because they set the real loss budget for using the bosonic Kitaev chain as a multimode quantum amplifier or sensor.","feed_headline":"Arbitrarily large loss on odd sites leaves amplification intact","feed_subtitle":"Even-length Kitaev chain keeps steady-state gain from site 1 exactly independent of odd-site loss rates.","key_machinery":"The carrying object is the non-Hermitian dynamical matrix $M_x$ in the quadrature basis, with on-site loss entering as diagonal imaginary terms $-i\\gamma_j/2$; its periodic-boundary eigenvalue curve winds around the origin, and the associated winding number $\\nu$ is the topological invariant. Topological amplification is diagnosed through the singular value decomposition $M_x=U\\Sigma V^\\dagger$: a nonzero winding number gives zero singular modes localizing at opposite edges, and the steady-state response $\\chi_{xx}[0]=iM_x^{-1}=\\sum_j (i/\\sigma_j)|u_j\\rangle\\langle v_j|$ then grows exponentially along the chain. The robustness result is carried by a Dyson-equation induction in Appendix E: adding loss on an odd site only changes zero-frequency matrix elements that already vanish in the lossless chain, leaving the response from site 1 to even sites exactly unchanged.","core_discovery":"In the dynamically stable regime $t>\\Delta>0$, the paper proves that an even-length bosonic Kitaev chain with bath couplings $\\gamma_{2j+1}\\ge0$ on odd sites and $\\gamma_{2j}=0$ on even sites always remains in the exponentially amplifying topological phase. Appendix E shows by induction with Dyson's equation that the zero-frequency susceptibility obeys $\\tilde{\\chi}_{N-1}[n,1;0]=\\tilde{\\chi}[n,1;0]$ (equal to $0$ for odd $n$ and to $\\tilde{t}^{-1}$ for even $n$), so the steady-state average particle number is exponentially localized on the last site no matter how large the odd-site losses are. The paper also derives exact loss values at which amplification is lost in other configurations: for a two-site unit cell with both sites lossy, the phase transition occurs at $\\sqrt{\\gamma_1\\gamma_2}/2=2\\Delta$, while for an odd-length unit cell $L$ with loss on its first site it occurs at $\\gamma_c/2=(t+\\Delta)e^{r(L-1)}-(t-\\Delta)e^{-r(L-1)}$ with $e^{2r}=(t+\\Delta)/(t-\\Delta)$. The asymmetry is traced to non-reciprocal dynamics: the periodic-boundary spectrum cannot develop a zero mode for arbitrary odd-site loss because the open-boundary eigenstates are exponentially localized on one edge.","pith_inferences":[],"forward_implications":["An even-length BKC driven from site 1 delivers the same exponential steady-state response at even sites no matter how the odd-site loss rates are chosen, so a sensor design can place strong loss on those sites without sacrificing gain.","Topological amplification is not bounded by the non-Hermitian point gap in this configuration; the usual robustness criterion that loss stay below the gap is sufficient but not necessary.","For odd-length unit cells, loss on the first site provokes a topological phase transition at the explicit value $\\gamma_c/2=(t+\\Delta)e^{r(L-1)}-(t-\\Delta)e^{-r(L-1)}$, so odd periodicity sets a hard loss budget.","For a two-site cell with loss on both sites, exponential amplification survives exactly while $\\sqrt{\\gamma_1\\gamma_2}/2<2\\Delta$, and the appearance of a line gap in the periodic spectrum does not by itself destroy it.","These exact thresholds give testable design rules for optomechanical or superconducting implementations of the bosonic Kitaev chain, since the paper expects the results to transfer to other non-reciprocal bosonic systems.","Because the proof relies on the directional, non-reciprocal structure of the quadrature dynamics rather than on the specific bosonic realization, the same sublattice-loss invariance should appear in any one-dimensional directional bosonic amplifier with strictly one-way transport.","A natural next calculation the paper leaves open is the effect of a small baseline loss on even sites; one would expect the amplification exponent to acquire a correction, and quantifying that decay rate would tell whether arbitrarily large odd-site loss remains useful in realistic platforms.","The exact invariance suggests a design rule: place unavoidable dissipation on the sublattice that carries signal away from the output, and keep the output-side sublattice lossless, to preserve zero-frequency gain."],"supporting_citations":[{"why":"Introduces the bosonic Kitaev chain model, its stable regime $t>\\Delta>0$, and the phase-dependent chiral transport that the paper builds on.","marker":"[1]"},{"why":"Provides the steady-state susceptibility setting, exponential signal-to-noise argument, and the drive-from-site-1 protocol used throughout.","marker":"[4]"},{"why":"Introduces topological amplification in photonic lattices, the phenomenon the paper studies.","marker":"[2]"},{"why":"Supplies the topological framework for directional amplification in driven-dissipative cavity arrays that links winding numbers to amplification channels.","marker":"[3]"},{"why":"Gives the singular-value-decomposition restoration of the non-Hermitian bulk-boundary correspondence used to identify zero singular modes as amplification channels.","marker":"[56]"},{"why":"States the disorder-robustness criterion (loss below the non-Hermitian gap) that this paper shows is sufficient but not necessary.","marker":"[55]"},{"why":"Provides the input-output and Heisenberg-Langevin formalism used to define the susceptibilities and loss rates.","marker":"[69]"},{"why":"Supplies the non-Hermitian symmetry classification used to assign the winding-number invariant.","marker":"[71]"}],"fun_headline_variants":["Odd-site loss can't break bosonic Kitaev amplification","Even chain amplifies despite arbitrary odd-site loss","Topological gain survives large odd-site dissipation","Odd losses leave Kitaev steady-state gain unaffected","Non-uniform loss: odd sites immune to amplification loss"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The exact cancellation requires the even sites to have exactly zero bath coupling ($\\gamma_2=\\gamma_4=\\cdots=\\gamma_N=0$); if every even site has even a small baseline loss, the proof of zero-frequency invariance no longer applies, and the paper does not analyze how the robustness degrades.","fun_headline_variants_meta":{"raw":{"variants":["Odd-site loss can't break bosonic Kitaev amplification","Even chain amplifies despite arbitrary odd-site loss","Topological gain survives large odd-site dissipation","Odd losses leave Kitaev steady-state gain unaffected","Non-uniform loss: odd sites immune to amplification loss"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000182,"raw_usage":{"total_tokens":1394,"prompt_tokens":1113,"completion_tokens":281,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":729,"completion_tokens_details":{"reasoning_tokens":208}},"tokens_in":729,"tokens_out":281,"duration_ms":3393,"temperature":1.0,"reasoning_tokens":208,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T16:47:39.592998+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fix an even-length open chain with $t>\\Delta>0$, set $\\gamma_2=\\gamma_4=\\cdots=0$ and $\\gamma_1,\\gamma_3,\\ldots$ to a large common value, then compute the smallest singular value of $M_x^{\\mathrm{OBC}}$ and the normalized steady-state particle number on the last site as $N$ grows; the claim predicts both remain exponentially small and localized for every such loss, so observing the singular value to plateau with $N$ or the particle distribution to spread would falsify it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"States the disorder-robustness criterion (loss below the non-Hermitian gap) that this paper shows is sufficient but not necessary."},{"cited_title":"Fortin, Topological amplification of the bosonic kitaev chain with non-uniform loss,https://doi.org/10.5281/ zenodo.16764370 (Zenodo) (2025)","cited_arxiv_id":null,"evidence_quote":"Supplies the non-Hermitian symmetry classification used to assign the winding-number invariant."}],"review_version":1}