{"id":"7dabc926-ad0b-46a3-b1f9-347e1977d042","arxiv_id":"1908.04912","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A resource theory for the coherence of Gaussian quantum channels is proposed, with structural characterizations of incoherent channels and superchannels and a computable relative-entropy measure.","lead":"This paper defines what it means for a quantum operation on light modes to lack quantum coherence, and it builds a way to measure how much coherence such operations have. It gives explicit formulas for basic examples such as shifting a light field's amplitude.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 2's Gaussian-superchannel representation depends on an unproved lifting to finite-squeezing Choi states; Theorems 4 and 5 inherit this gap.","rationale":"The reader's weakest assumption identifies exactly the same load-bearing gap: the proof of Theorem 2 assumes, rather than demonstrates, that a Gaussian superchannel induces a Gaussian channel on finite-squeezing Choi states. This is not a minor technicality because the structural characterization of incoherent Gaussian superchannels in Theorem 4, and hence the monotonicity axiom C2 used by Theorem 5, rests on that representation. I checked the surrounding argument for alternative concerns: Theorem 1's characterization of incoherent Gaussian channels appears internally consistent, Theorem 5's lifting argument is valid provided Theorem 4 holds, and the relative-entropy measure satisfies the required monotonicity because it is a distance to the free set contracted by any channel that preserves thermal states. The proof of Lemma 2 is terse and its stated quantification is stronger than what is used, but that is secondary compared with the missing Choi-state lifting. The paper's defined notation, explicit examples, and the naturalness of the proposed measure all count in its favor, and I found no internal contradiction. However, without a rigorous derivation of the Choi-state representation (e.g., via a proper infinite-squeezing limit or an explicit Stinespring construction), the central claim is not fully established. A conditional verdict is therefore appropriate, and the proposed concrete test would settle whether the gap is merely a proof deficiency or a genuine obstruction.","tokens_in":9002,"tokens_out":31100,"duration_ms":329840,"concrete_test":"Re-derive Theorem 2 from the Stinespring representation of a Gaussian superchannel, e.g. Phi(phi)=Tr_E[V(phi otimes I_E) U rho_sq U^dagger V^dagger] with Gaussian U,V and squeezed vacuum rho_sq, and compute the effective 2n-mode channel (X,Y0,d~) in terms of the beam-splitter and squeezing parameters. Then verify explicitly that X V(r) X^t + Y0 equals the covariance matrix of rho_{Phi(phi)}(r) for r=0.5, 1, 2 and for several phi(T,N,d). If equality holds for all finite r, the finite-r lifting is sound for that class; if it fails or requires r to go to infinity, the missing limit in the proof is load-bearing.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing step is in Section II / Appendix B: after defining the finite-squeezing Choi state rho_phi (Eq. 16), the proof of Theorem 2 asserts 'we will find a Gaussian channel' that maps rho_phi to rho_{Phi(phi)}. No argument is given that a Gaussian superchannel Phi actually induces such a 2n-mode Gaussian channel, nor that the matrices X, Y0 can be chosen independently of the squeezing parameter r and of the input channel phi. The coefficient-matching argument in Eqs. (A17)-(A25) only derives necessary conditions once that channel is assumed to exist. The Choi state at finite r is not the standard Choi-Jamiolkowski operator of the channel, and the paper never passes to an infinite-squeezing limit that would make the correspondence faithful. Since Theorem 3 builds on Theorem 2, Theorem 4 (the left-right form of incoherent Gaussian superchannels) inherits this gap, and Theorem 5's monotonicity axiom (C2) for C(phi)=sup_{rho in IGS} C[phi(rho)] depends on Theorem 4. If the lifting fails for a legitimate Gaussian superchannel, the claimed resource theory of Gaussian channel coherence is not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a resource theory for coherence of Gaussian channels in continuous-variable systems. It defines incoherent Gaussian channels (IGC) as Gaussian channels that map every thermal Gaussian state to a thermal Gaussian state, and characterizes them in Theorem 1 as channels with zero displacement, a block-permuted orthogonal matrix T, and diagonal noise satisfying an inequality. It then defines incoherent Gaussian superchannels (IGSC) as superchannels that preserve IGC, and claims in Theorem 4 that a Gaussian superchannel is incoherent iff it acts as Phi(phi)=chi2 o phi o chi1 with fixed chi1,chi2 in IGC. Theorem 5 constructs a coherence measure for Gaussian channels from any coherence monotone for Gaussian states via C(phi)=sup_{rho_th in IGS_n} C[phi(rho_th)], with the relative-entropy measure giving an explicit formula; two examples (constant channels and displacement channels) are computed.","tokens_in":9228,"tokens_out":41488,"duration_ms":403448,"significance":"If the structural results were correct, the paper would provide a clean framework for dynamic coherence in the Gaussian setting: an explicit characterization of free operations and a generic lifting of state coherence measures to channel coherence measures. The displacement-channel formula Cr[D(lambda)] = sum_j f(|lambda_j|^2) is concrete and testable, and the idea of defining channel coherence by maximizing a state coherence monotone over thermal inputs is natural. The main weakness is that the central structural theorem for Gaussian superchannels is not established as stated and appears to be false for a natural class of superchannels. This affects the derivation of the free-operation set and hence the monotonicity axiom (C2) for the proposed resource theory.","major_comments":[{"comment":"The representation in Theorem 2 is not valid for all Gaussian superchannels. Let U_S be the Gaussian unitary associated with a real symplectic matrix S that is not orthogonal (e.g., single-mode squeezing), and consider the superchannel Phi(phi)=U_S o phi o U_S^dagger. This is a completely positive linear map sending Gaussian channels to Gaussian channels, hence it belongs to GSC_n by the paper's definition. It acts on the channel parameters as T'=S T S^{-1}, N'=S N S^t, d'=S d. If Theorem 2 held, then for the identity channel (T=I,N=0,d=0) the equations T'=A T R with R=Sigma O^t Sigma and R R^t=I would give A R=I and Y=0,d=0. Hence T'=A T A^{-1} for all T. This forces A^{-1}S to commute with all matrices, so A=lambda S and R=lambda^{-1} S^{-1}. Since R must be orthogonal, S must be orthogonal up to scale, hence S orthogonal. This contradicts the assumption that S is a non-orthogonal symplectic matrix. Thus Theorem 2 needs either a symplectic (not necessarily orthogonal) O or an explicit restriction of the class of Gaussian superchannels; the proof in Appendix B does not address this issue.","section":"Section II, Theorem 2"},{"comment":"Independently of the previous comment, the proof of Theorem 2 does not rigorously justify the Choi-state argument. The state rho_phi in Eq. (16) is defined with a finite squeezing parameter r, and it is not the standard Choi-Jamiolkowski operator, which requires an infinite-squeezing limit or a proper normalization. The proof merely says 'we will find a Gaussian channel' that transforms rho_phi into rho_{Phi(phi)}, and then matches coefficients as r varies in Eqs. (A17)-(A25). No limit r -> infinity is taken, and no argument is given that the matrices A,O,Y can be chosen independently of r and of the input channel phi. Since Theorems 3 and 4 inherit this representation, and Theorem 5's monotonicity axiom (C2) depends on Theorem 4, this is a load-bearing gap that must be fixed with a rigorous Choi-isomorphism argument or an alternative derivation.","section":"Section II, Eq. (16) and Appendix B"},{"comment":"The proof of (1) implies (2) in Theorem 4 is too compressed. The key step showing that O belongs to T_n is summarized in the sentence 'Varying t_j, T_j, for all j, and using the facts of Lemma 1 and lemma 2, we can get O' in T_n.' The proofs of Lemma 1 and especially Lemma 2 are also only sketched. Because Theorem 4 is the structural characterization of the free operations of the resource theory, this derivation must be presented in full. In addition, if Theorem 2 is reformulated as suggested in the first major comment, the proof of Theorem 4 will need to be reworked.","section":"Appendix D"},{"comment":"The proof of Theorem 5 is omitted ('the proof is simple'). The paper does not explicitly verify conditions (C1) and (C2), and the closing remark acknowledges a possible divergence of the supremum in Eq. (26) without stating whether infinite values are admissible for a coherence measure. Since Theorem 5 is the central construction of the resource-theoretic measure, a complete proof and a clear statement of the admissible range of the measure are needed.","section":"Section III, Theorem 5"}],"minor_comments":[{"comment":"The headings of Examples 1 and 2 read 'Exmple' and should be 'Example'.","section":"Section III"},{"comment":"The Introduction refers to 'Section VI' but the paper has only four sections; this should be 'Section IV'.","section":"Introduction"},{"comment":"The abstract has the inconsistent capitalization 'We establish'; 'We' should be lowercase, or the sentence should be rephrased.","section":"Abstract"},{"comment":"The symbol Phi is used both for the Gaussian superchannel and for the Gaussian channel on Choi states in Appendix B; these should be distinguished to avoid confusion.","section":"Section II, Eq. (16) and Appendix B"},{"comment":"In Example 2, D(lambda) denotes both a displacement operator and a displacement channel; the distinction should be made explicit in the notation.","section":"Example 2"}],"recommendation":"major_revision","confidential_remarks":"The main concern is that Theorem 2 appears to be false for the class of Gaussian superchannels that conjugate by a non-orthogonal symplectic unitary. This is not a small gap in the proof but a structural issue that affects the characterization of incoherent superchannels and the monotonicity of the proposed measure. If the authors can reformulate the theorem (for example, by allowing O to be symplectic rather than orthogonal) and rework the dependent results, the paper's core idea may still be salvageable. I would not recommend acceptance without a complete revision of the central structural theorem."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is a sincere attempt to build a resource theory for coherence of Gaussian channels, and the parts that are fully proven are worth having. The structural characterization of incoherent Gaussian channels (Theorem 1) and the lifting construction (Theorem 5, C(φ)=sup_{ρth} C[φ(ρth)]) are new relative to the finite-dimensional channel-coherence papers it cites, and the displacement-channel example computes cleanly to Σ f(|λ_j|^2). The paper is well organized and the definitions are natural.\n\nThe soft spot is Theorem 2, the representation of Gaussian superchannels. The proof in Appendix B assumes that a Gaussian superchannel induces a Gaussian channel on the finite-squeezing Choi state ρ_φ and then matches coefficients as r varies. That existence assumption is never justified. The finite-r Choi state is not the standard Choi–Jamiolkowski operator of the channel, and the paper never passes to the infinite-squeezing limit that would make the correspondence faithful. Since Theorem 3 and Theorem 4 build on Theorem 2, and the monotonicity axiom (C2) in Theorem 5 needs Theorem 4, the resource theory's axioms are not established as written. This is a real gap, not a cosmetic one. It looks addressable—a proper Choi-isomorphism argument for Gaussian superchannels plus a limiting procedure should fill it—but right now the central structural result is more of a conjecture.\n\nTwo smaller things bother me. Reference [11] is the author's own prior preprint, and that overlap is not disclosed; if it contains the same or overlapping results, that should be said. And the proof of Theorem 4 in Appendix D is compressed: Lemma 1 is stated but not proven, and the step 'varying t_j, T_j' hides a fair amount. These are minor by comparison.\n\nBottom line: the paper is for readers who want a concrete channel-coherence measure in continuous variables and who are comfortable supplying missing technical steps. I would send it to peer review, because the framework and examples are useful and the gap is probably repairable, but the referee would need to enforce a complete proof of Theorem 2. I would not cite it in my own work until that happens. For a reading group, maybe—good for a critical session.","headline":"An interesting and clearly written resource-theory framework for Gaussian channel coherence, but the central superchannel representation is asserted rather than proved.","tokens_in":9716,"tokens_out":14039,"would_cite":false,"duration_ms":138206,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.65.Ud","03.67.Mn","03.65.Aa"],"model":"deepseek-v4-flash","headline":"Coherence measures for Gaussian channels can be built from state coherence measures.","keywords":["Gaussian channels","coherence measures","incoherent Gaussian channels","Gaussian superchannels","resource theory","continuous-variable quantum information","relative entropy of coherence","Choi state"],"falsifier":"A decisive two-mode test is to choose $A=0$ and $O$ a four-by-four orthogonal matrix that mixes the two modes instead of mapping each mode block to a single output block, set $Y = \\eta_1 I_2 \\oplus \\eta_2 I_2$, and ask whether the resulting Gaussian superchannel $\\Phi$ maps every incoherent Gaussian channel to an incoherent Gaussian channel. Since Theorem 4 predicts $O$ must lie in $T_n$, an $O$ that mixes modes while preserving the incoherent set would falsify the classification; the same example can be checked against the finite-r Choi equations to test the missing infinite-squeezing limit.","tokens_in":8789,"feed_emoji":"⚛️","tokens_out":11783,"duration_ms":111377,"temperature":0.7,"pith_summary":"This paper extends coherence from quantum states to Gaussian channels, the continuous-variable operations on modes of light used throughout quantum optics. It defines a Gaussian channel as incoherent when it sends every thermal Gaussian state to an incoherent Gaussian state, and it characterizes both these channels and the superchannels that preserve them. The central result is a recipe: given any coherence measure on Gaussian states that is nonnegative, vanishes only on thermal states, and does not increase under incoherent Gaussian channels, the supremum of that measure over thermal inputs defines a genuine coherence measure for Gaussian channels. Applying the recipe to relative entropy yields an explicit, computable measure with closed formulas for constant and displacement channels. If the result stands, the static theory of state coherence transfers wholesale to the dynamic setting of channels.","feed_headline":"Gaussian channel coherence becomes a full resource theory","feed_subtitle":"Any state coherence measure that is monotone under incoherent Gaussian maps lifts to channels by Eq. (26).","key_machinery":"The load-bearing object is the Choi state of a Gaussian channel, Eq. (16), a $2n$-mode Gaussian state built from the channel's matrices $T,N$ and a finite squeezing parameter $r$. The proof varies $r$ over all real values, which forces the block matrices in the output covariance matrix to satisfy the identities (A.17)-(A.25); these pin down the form of a Gaussian superchannel and, via two matrix lemmas, force incoherence-preserving superchannels to be conjugation by incoherent channels. The identity $\\Sigma_n\\Omega\\Sigma_n = -\\Omega$ is used repeatedly to move between equivalent forms of the superchannel condition. The channel-state correspondence remains at finite $r$ throughout; no infinite-squeezing limit is taken.","core_discovery":"The paper's central claim is that a resource theory for the coherence of Gaussian channels is determined by two free-operation definitions. Theorem 1 says an incoherent Gaussian channel has zero displacement, acts on each mode's covariance matrix through at most one orthogonal $2\\times 2$ block (up to a scale factor), and adds isotropic noise at least as large as the leakage from mode coupling. Theorem 2 represents every Gaussian superchannel by matrices $(A,O,Y,\\bar d)$ acting on the channel parameters as $T' = A T \\Sigma_n O^t \\Sigma_n$, $N' = A N A^t + Y$, $d' = A d + \\bar d$; Theorem 3 rewrites this as $\\varphi_2 \\circ \\varphi \\circ \\varphi_1$ with fixed Gaussian channels. Theorem 4 proves that the incoherent Gaussian superchannels are exactly those of the form $\\chi_2 \\circ \\varphi \\circ \\chi_1$ with fixed incoherent Gaussian channels $\\chi_1,\\chi_2$. Theorem 5 then lifts any Gaussian-state coherence monotone $C$ to a channel coherence measure $C(\\varphi) = \\sup_{\\rho_{\\rm th}\\in IGS_n} C[\\varphi(\\rho_{\\rm th})]$, and the relative-entropy choice gives $C_r(\\varphi)$ with the closed formulas in Eqs. (28)-(31).","pith_inferences":["The same supremum construction should generalize beyond Gaussian channels: the proof of Theorem 5 uses only faithfulness, monotonicity, and closure of the free operations, so any resource theory of states whose free operations form a closed set of channels inherits a channel-level measure by the same formula; the paper does not state this extension.","The divergence caveat the paper flags is likely to matter in practice: for some state measures (Bures or Hellinger based) the supremum over unbounded thermal squeezing may be infinite, so a usable theory may need finite-energy cutoffs or a renormalized version of Eq. (26).","A natural next test is to compute $C_r$ for elementary continuous-variable gates such as beam splitters, squeezers, and amplifiers; additivity and the displacement formula give the base values, and axiom (C2) predicts concrete inequalities among these gate coherences that can be checked numerically.","Theorem 4 implies the Gaussian coherence resource theory has no hidden free superchannels beyond pre- and post-composition with incoherent channels, which if correct makes the monotonicity axiom fully determined by the free operations alone."],"forward_implications":["Any state coherence measure satisfying faithfulness and monotonicity under incoherent Gaussian channels automatically defines a valid Gaussian-channel coherence measure by Eq. (26), so the existing menu of state measures becomes a menu of channel measures.","The free operations are exactly composition with incoherent Gaussian channels before and after the channel: by Theorem 4, monotonicity under the pair of conditions (C3a) and (C3b) is equivalent to the full monotonicity axiom (C2).","The relative-entropy measure $C_r$ is additive under tensor products, and for a multimode displacement channel it evaluates as $C_r[D(\\lambda)] = \\sum_j f(|\\lambda_j|^2)$, giving a closed, mode-by-mode formula.","Constant Gaussian channels have channel coherence equal to the state coherence of their fixed output state, so creating a coherent output from a thermal input costs exactly the coherence of that output state.","Because the definitions are resource-nongenerating, no free operation can increase the coherence of a channel, making the measure operationally meaningful as a monotone."],"supporting_citations":[{"why":"Supplies the Gaussian-state and Choi-state background used to justify that the state in Eq. (16) is a Gaussian state.","marker":"[14]"},{"why":"Provides the analytical expression for the relative entropy of coherence of Gaussian states and the fact that diagonal Gaussian states are exactly thermal states.","marker":"[16]"},{"why":"Gives the $(T,N,d)$ parameterization of Gaussian channels and the entropy formula used in Eqs. (28)-(30).","marker":"[24]"},{"why":"Defines incoherent states and the relative entropy of coherence for states, the static resource the paper lifts to channels.","marker":"[12]"},{"why":"Supplies the resource-theory formalism in which free operations are resource-nongenerating, the principle the paper's definitions follow.","marker":"[21]"},{"why":"Provides the review background on Gaussian states, covariance matrices, and symplectic eigenvalues used in the formulas.","marker":"[13]"},{"why":"Defines superchannels as maps from channels to channels, which the paper adapts to the Gaussian setting.","marker":"[25]"}],"fun_headline_variants":["Coherence of Gaussian channels becomes a resource theory","Incoherent Gaussian maps fix channel coherence","Channel coherence lifted from Gaussian states","Defining free ops for Gaussian channel coherence","Two theorems pin down incoherent Gaussian channels"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof treats the finite-squeezing Choi state as a faithful stand-in for the channel and draws conclusions from matrix equations that hold as the squeezing parameter $r$ varies, but it never passes to the infinite-squeezing limit; if some Gaussian superchannel acts differently on that limit, the structural characterization in Theorem 4 is not established.","fun_headline_variants_meta":{"raw":{"variants":["Coherence of Gaussian channels becomes a resource theory","Incoherent Gaussian maps fix channel coherence","Channel coherence lifted from Gaussian states","Defining free ops for Gaussian channel coherence","Two theorems pin down incoherent Gaussian channels"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000178,"raw_usage":{"total_tokens":1246,"prompt_tokens":845,"completion_tokens":401,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":461,"completion_tokens_details":{"reasoning_tokens":336}},"tokens_in":461,"tokens_out":401,"duration_ms":4281,"temperature":1.0,"reasoning_tokens":336,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:29:42.836576+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive two-mode test is to choose $A=0$ and $O$ a four-by-four orthogonal matrix that mixes the two modes instead of mapping each mode block to a single output block, set $Y = \\eta_1 I_2 \\oplus \\eta_2 I_2$, and ask whether the resulting Gaussian superchannel $\\Phi$ maps every incoherent Gaussian channel to an incoherent Gaussian channel. Since Theorem 4 predicts $O$ must lie in $T_n$, an $O$ that mixes modes while preserving the incoherent set would falsify the classification; the same example can be checked against the finite-r Choi equations to test the missing infinite-squeezing limit.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the analytical expression for the relative entropy of coherence of Gaussian states and the fact that diagonal Gaussian states are exactly thermal states."},{"cited_title":"Coherence of quantum channels","cited_arxiv_id":"1907.07289","evidence_quote":"Defines incoherent states and the relative entropy of coherence for states, the static resource the paper lifts to channels."},{"cited_title":"Albarelli, M","cited_arxiv_id":null,"evidence_quote":"Supplies the resource-theory formalism in which free operations are resource-nongenerating, the principle the paper's definitions follow."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines superchannels as maps from channels to channels, which the paper adapts to the Gaussian setting."}],"review_version":1}