{"id":"be142df5-a7df-4cc8-822d-7c3747e00db2","arxiv_id":"1908.04432","paper_version":1,"verdict":"REJECT","confidence":"HIGH","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The central formula for a coarse-grained effective state is a direct restatement of a known quantum state pooling theorem, with no new proof or construction.","lead":"The paper reframes quantum coarse-graining as a problem of two agents pooling their beliefs about a system, and states that a pooled effective state can be written as a product of the agents' assignments and the inverse prior. It mainly reuses an existing quantum pooling theorem from the conditional states literature and adds no derivation of its own.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central map (17) is never shown to arise from the coarse-graining scenario; without a construction of X1, X2, and sufficient statistics from (U, Λ), Result 7 does not address the claimed problem.","rationale":"The reader's weakest_assumption identifies the same bottleneck that I find decisive: the paper never connects the coarse-graining ingredients (U and Λ) to the pooling scenario, and without that connection Result 7 is simply a restatement of a known theorem. I read the manuscript in good faith and acknowledge that its explicit goal is to reinterpret the coarse-graining problem in decision-theoretic terms, and that such a reinterpretation could in principle be a legitimate contribution. But the abstract promises \"necessary and sufficient conditions\" for the existence of a well-defined coarse-grained state, while the actual result is only a conditional statement about a pooled state. The text after Eq. (17) acknowledges that the assignments must be traceable back to ρ_B, but no such traceability is demonstrated. The uniqueness assumption for σ1 and σ2 is also asserted, not proven; without it, Eq. (17) may not even define a map. The erroneous Proposition 2 is a separate mathematical issue that further weakens confidence, but the missing translation is the main structural gap. Since this gap prevents the central claim from being a theorem about the original coarse-graining problem, I agree with the REJECT verdict.","tokens_in":11336,"tokens_out":4581,"duration_ms":48129,"concrete_test":"Take the simplest nontrivial coarse-graining example from Ref. [14] (e.g., a qubit system, U a rotation, Λ a dephasing or partial trace). Write down explicitly the two agents' data X1, X2, the prior ρ_B, and minimal sufficient statistics s1, s2, following the paper's intended translation. Then compute the pooled state from Eq. (16) and compare it with the exact coarse-grained state Λ(U ρ U†). If the two differ, or if the translation cannot be defined, the central claim fails. Ideally also check one case where the original diagram commutes but Eq. (14) fails, which would refute the claimed equivalence.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the unstated translation from the coarse-graining scenario of Ref. [14] into the two-agent pooling model. Result 7 in Sec. III is a restatement of Theorem 6 (already quoted from [19]); it gives a sufficient condition for the formula σ_pooled = c σ1 ρ^{-1} σ2. For this to be a result about coarse-graining, the paper must specify, from the original data (unitary U, CPTP map Λ, and prior ρ_B), what X1, X2, s1, s2 and the assignments σ1, σ2 are. The paragraph after Eq. (17) asserts that these can be traced back to ρ_B and \"could have originated out of a channel, from Bayesian condition or by any other means\", but no construction or proof is supplied. Moreover, the abstract's \"necessary and sufficient conditions\" is not supported by Result 7, which is only conditional and has no converse; no equivalence with the original diagrammatic compatibility is established. The unproven uniqueness claim (\"For each initial prior ρ_B there is only one σ1_B...\") is also needed for the map in Eq. (17) to be well-defined. If the translation cannot be made, the paper's central claim does not apply to its motivating problem.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript proposes to reinterpret the coarse-graining problem of Ref. [14]—where a microscopic unitary evolution U and a lossy CPTP map Λ are to be replaced by an effective description—as a quantum state pooling problem between two Bayesian agents, Wanda and Theo. It reviews the conditional quantum states formalism, defines objective and subjective compatibility, and quotes a theorem from Leifer and Spekkens (Theorem 6) asserting that if minimal sufficient statistics for the agents' data are conditionally independent given the system B, then the pooled state is sigma_pooled = c sigma1 rho^{-1} sigma2. The paper states this same statement as Result 7, defines a non-linear map Gamma_tilde(rho_B) = c sigma1 rho^{-1} sigma2 in Eq. (17), and concludes that it has provided necessary and sufficient conditions for the existence of an effective coarse-grained state.","tokens_in":11642,"tokens_out":7125,"duration_ms":66736,"significance":"If the proposed translation were valid, the decision-theoretic reformulation could offer a new way to characterize effective coarse-grained states. The paper is transparent in attributing the pooling formula to Ref. [19], and it does not hide the fact that Result 7 is a restatement. However, the central claim of the paper—that Result 7 gives necessary and sufficient conditions for the original coarse-graining problem—is not established: the connection between the physical data (U, Lambda) and the pooling variables (X1, X2, s1, s2) is never constructed, no necessity proof is given, and an illustrative proposition in the preliminaries is false. These gaps are load-bearing, so the paper as it stands does not provide the advertised result.","major_comments":[{"comment":"Result 7 is identical to Theorem 6 quoted from Ref. [19] and is stated without proof. The abstract and conclusion claim 'necessary and sufficient conditions' for the existence of a well-defined coarse-grained state, but Result 7 is a conditional sufficiency statement: if the independence condition in Eq. (14) holds, then Eq. (16) holds. No converse is proved, and no equivalence with the commuting-diagram compatibility of Ref. [14] is established. The paper therefore does not support its advertised necessary-and-sufficient claim.","section":"Section III, Result 7"},{"comment":"The paper never constructs, from the original coarse-graining data (unitary U, CPTP map Lambda, and prior rho_B), the random variables X1 and X2, their minimal sufficient statistics s1 and s2, or the assignments sigma1_B and sigma2_B that appear in Result 7. Without this translation, Result 7 is a statement about an abstract pooling scenario, not about the coarse-graining diagram of Fig. 1; the sentence that the assignments 'could have originated out of a channel, from Bayesian condition or by any other means' does not supply the missing construction or a proof that the translation is faithful.","section":"Section III, paragraph after Eq. (17)"},{"comment":"The map Gamma_tilde is only well-defined if each prior rho_B determines a unique pair of agent assignments sigma1_B and sigma2_B. The paper asserts this uniqueness in the paragraph after Eq. (17) but gives no proof and specifies no model of the agents' interactions or Bayesian updates. If the same prior can lead to different assignments (for example, through different data-sets or likelihoods), the map is not a function of rho_B alone.","section":"Section III, Eq. (17)"},{"comment":"Proposition 2 is false as stated. For binary Y, the assignments Q1=(1,0) and Q2=(1/2,1/2) have overlapping support, so Theorem 1 implies they are compatible, yet Proposition 2 declares them incompatible because p is not in (0,1). The proof's case analysis incorrectly assumes that if p=1 then q must be 0 or 1. Although this example is not used in the derivation of Result 7, it is a stated theorem in the paper and must be corrected.","section":"Section II C, Proposition 2"}],"minor_comments":[{"comment":"The first condition of Definition 4 reads 'Tr_B(rho_{X=x|B} sigma^i_B) for all x and for all i' with no predicate; presumably it should state that this quantity is positive (or nonzero).","section":"Section II C, Definition 4"},{"comment":"The proof of Proposition 2 refers to 'Eq. (12)' but the relevant condition is Eq. (9) from Theorem 1; the equation numbering should be corrected.","section":"Section II C, Proposition 2 proof"},{"comment":"There are several typographical and formatting issues, including 'Schroedinger' for 'Schrödinger', missing accents, and some incomplete sentences; a careful proofreading pass is needed.","section":"Throughout"}],"recommendation":"reject","confidential_remarks":"The paper's main result is a direct restatement of a known theorem, the promised application to coarse-graining is not carried out, and Proposition 2 is false. I do not see a way to fix these issues within the scope of a revision because the missing translation and the necessary-and-sufficient claim require substantial new technical work. If the claims were scaled back to an expository note about state pooling, the manuscript might be suitable for a different venue."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the short version: the paper's central result is Theorem 6 from Leifer and Spekkens, restated as Result 7, and the advertised bridge to the coarse-graining problem is never built. There is also a false proposition in the preliminaries. I don't think it is salvageable as is.\n\nWhat's actually new is thin but not zero. Framing the effective coarse-grained state as a pooled state shared by two Bayesian agents is a real perspective shift, and it points toward a connection between coarse-graining and causal inference that could be worth exploring. The exposition of conditional states and pooling is clear and mostly correct, and the paper is upfront that Eq. (16) comes from an existing theorem. The observation that Eq. (17) defines a non-linear map is small but legitimate.\n\nThe soft spots are load-bearing. First, Result 7 is a restatement, not a derivation: no proof is given, and the abstract's claim of \"necessary and sufficient conditions\" goes beyond what a conditional sufficient condition supports. Second, the paper never constructs the data X1, X2, or the minimal sufficient statistics s1, s2 from the original coarse-graining data (U, Λ, ρB). Without that construction, Eq. (17) does not actually address the problem in Ref. [14]; it just relabels the pooling formula as an effective state. The uniqueness claim used to define the map — that each prior ρB determines exactly one assignment per agent — is asserted, not proved, and is not obviously true. Third, Proposition 2 is simply false: with Q1=(1,0) and Q2=(0.5,0.5) the supports overlap, so by the paper's own Theorem 1 the assignments are compatible, but Proposition 2 excludes this case. That is a concrete error, not a matter of taste.\n\nWho gets value from this? Someone wanting an accessible recap of conditional states and state pooling might find the early sections useful. As a research contribution to the coarse-graining problem, it does not land. The math does not support the main claims, and the central formula is prior literature.\n\nRecommendation: desk reject. To pursue this idea, the author would need to actually reduce the coarse-graining scenario to a pooling problem, prove the necessity direction, and fix the binary compatibility claim.","headline":"The main result is a restatement of a known theorem from Leifer and Spekkens, the advertised bridge to coarse-graining is never built, and a preliminary proposition is false; this does not deserve referee time.","tokens_in":12116,"tokens_out":4157,"would_cite":false,"duration_ms":39421,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper claims that an effective coarse-grained state exists when two agents' data summaries are conditionally independent, and then the state is $c\\,\\sigma_1\\rho^{-1}\\sigma_2$.","keywords":["coarse-graining","quantum state pooling","conditional quantum states","Bayesian inference","compatibility","sufficient statistics","effective dynamics","decision theory"],"falsifier":"Take an explicit coarse-graining instance with a concrete unitary $U$, lossy map $\\Lambda$, and prior $\\rho_B$, then construct the agents' data sets from that instance. If Eq. (14) holds but $c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$ fails to be a positive trace-one state, or if a compatible pooled state can be shown to exist while no choice of minimal sufficient statistics satisfies Eq. (14), then the claimed necessary and sufficient condition is false.","tokens_in":11109,"feed_emoji":"⚛️","tokens_out":6660,"duration_ms":67008,"temperature":0.7,"pith_summary":"The paper claims that the existence of a meaningful coarse-grained quantum state can be decided by a single compatibility condition between two Bayesian agents, rather than by algebraic commutativity of a diagram. Starting from a shared prior $\\rho_B$, two agents assign states $\\sigma^1_B$ and $\\sigma^2_B$ after seeing different data. If the minimal sufficient statistics of their data are conditionally independent given the system, then the pooled, effective state is forced to be $\\sigma_{\\mathrm{pooled}} = c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$. This reframes the coarse-graining problem as a state-pooling task and yields a generally non-linear effective map on state space. The contribution matters because it replaces an algebraic test with a tangible independence condition.","feed_headline":"One independence condition fixes the coarse-grained quantum state","feed_subtitle":"Two Bayesian agents merge assignments into one state when their data summaries are conditionally independent.","key_machinery":"The machinery is the quantum conditional-state formalism, in which probabilities become trace-class operators and conditioning is written as $\\sigma_{B|A} = \\sigma_{AB} \\star \\sigma_A^{-1}$ using a non-commutative $\\star$-product. Hybrid classical-quantum states keep classical variables classical, and compatibility of two assignments is characterized by overlapping supports. The state-pooling theorem, stated as Theorem 6, gives the pooled state $c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$ when the minimal sufficient statistics satisfy Eq. (14). Result 7 applies that theorem to the coarse-graining scenario by treating the two agents' states as posterior assignments and the prior as $\\rho_B$. The independence condition in Eq. (14) is the load-bearing step that converts the existence of an effective state into a factorization statement.","core_discovery":"On the paper's own terms, the discovery is that quantum state pooling supplies necessary and sufficient conditions for a coarse-grained effective state. Two assignment maps are compatible when the supports of their assigned states overlap, and when the data behind them have minimal sufficient statistics $s_1$ and $s_2$ satisfying the conditional-independence relation $\\rho_{s_1(X_1)s_2(X_2)|B} = \\rho_{s_1(X_1)|B}\\rho_{s_2(X_2)|B}$, the unique pooled state is $\\sigma_{\\mathrm{pooled}} = c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$. The same formula defines a map $\\rho_B \\mapsto c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$ on the state space, which is generally non-linear. The paper presents this as a concrete alternative to demanding that the coarse-graining diagram commute or to solving an infinite family of semidefinite programs.","pith_inferences":["Not stated in the paper: the concrete construction of $X_1$, $X_2$, $s_1$, and $s_2$ from a given unitary $U$ and lossy map $\\Lambda$ is left implicit; spelling out that construction would turn Result 7 into an applicable algorithm.","A testable extension is to take a simple two-qubit model with two measurement devices, compute the minimal sufficient statistics explicitly, and check whether Eq. (14) holds and whether the resulting operator is positive and trace-one.","If the non-linearity of the effective map is physically meaningful, then coarse-grained quantum dynamics may not be describable by completely positive trace-preserving maps at all, which would connect this result to broader discussions of state merging and non-linear Bayesian updating.","The support-overlap compatibility condition suggests an experimental route: prepare ensembles whose assigned states have overlapping supports and verify that measurements match the predicted pooled state $c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$."],"forward_implications":["Whenever two agents' data have conditionally independent minimal sufficient statistics, their combined assignment is fixed by $c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$, so no additional rule for merging beliefs is needed.","The effective map $\\rho_B \\mapsto c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$ need not be a quantum channel, so coarse-graining can produce effective states that channel-based approaches would miss.","The existence of a meaningful coarse-grained state can be certified by a single independence condition instead of an infinite family of semidefinite programs.","The pooled state depends explicitly on the shared prior $\\rho_B$, so the prior is not washed out by pooling.","The compatibility criterion based on overlapping supports gives a direct way to check whether two agents' assignments can be reconciled at all."],"supporting_citations":[{"why":"Defines the original coarse-graining diagram and the algebraic commutativity requirement that this paper replaces with decision-theoretic compatibility.","marker":"[14]"},{"why":"Supplies the quantum conditional-state formalism, the compatibility definitions, and the classical and quantum state-pooling theorems that Result 7 directly applies.","marker":"[19]"},{"why":"Provides the hybrid-state and conditioning machinery used for quantum Bayes updating and for expressing conditional independence in Eq. (14).","marker":"[20]"},{"why":"Supports the use of Bayesian conditioning and causal-inference language in the quantum setting, which motivates the independence condition.","marker":"[21]"}],"fun_headline_variants":["Conditional independence makes agents' states compatible","Merge quantum states via conditional independence","Quantum pooling decoded by one independence relation","Compatibility condition yields unique effective state","When data summaries are independent, quantum states combine"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole argument rests on assuming that the original lossy quantum process can be exactly recast as two agents sharing a prior and updating on two data sets whose key summaries are independent given the system, with each prior producing a unique pair of assignments.","fun_headline_variants_meta":{"raw":{"variants":["Conditional independence makes agents' states compatible","Merge quantum states via conditional independence","Quantum pooling decoded by one independence relation","Compatibility condition yields unique effective state","When data summaries are independent, quantum states combine"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000727,"raw_usage":{"total_tokens":3192,"prompt_tokens":818,"completion_tokens":2374,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":434,"completion_tokens_details":{"reasoning_tokens":2311}},"tokens_in":434,"tokens_out":2374,"duration_ms":19777,"temperature":1.0,"reasoning_tokens":2311,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:42:44.230021+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take an explicit coarse-graining instance with a concrete unitary $U$, lossy map $\\Lambda$, and prior $\\rho_B$, then construct the agents' data sets from that instance. If Eq. (14) holds but $c\\,\\sigma^1_B \\rho_B^{-1} \\sigma^2_B$ fails to be a positive trace-one state, or if a compatible pooled state can be shown to exist while no choice of minimal sufficient statistics satisfies Eq. (14), then the claimed necessary and sufficient condition is false.","supporting_citations":[{"cited_title":"Wolfram, Reviews of Modern Physics 55, 601644 (1983)","cited_arxiv_id":null,"evidence_quote":"Defines the original coarse-graining diagram and the algebraic commutativity requirement that this paper replaces with decision-theoretic compatibility."},{"cited_title":"Levitt and A","cited_arxiv_id":null,"evidence_quote":"Supplies the quantum conditional-state formalism, the compatibility definitions, and the classical and quantum state-pooling theorems that Result 7 directly applies."},{"cited_title":"Naghibi Beidokhti, D","cited_arxiv_id":null,"evidence_quote":"Provides the hybrid-state and conditioning machinery used for quantum Bayes updating and for expressing conditional independence in Eq. (14)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the use of Bayesian conditioning and causal-inference language in the quantum setting, which motivates the independence condition."}],"review_version":1}