{"id":"12ef2cce-7c53-4ced-a4d0-9d3ad82316b6","arxiv_id":"2512.20133","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For several families of multipartite product mixed states, k-marginals determine the global state exactly only under strong purity/rank conditions, and UDA additivity holds only in restricted pure-product cases.","lead":"This paper gives criteria for when a multipartite mixed quantum state is the only state compatible with its local reduced density matrices, and characterizes when such uniqueness survives a tensor product. It also argues that almost all high-rank mixed states are not uniquely determined by fixed-order marginals, which matters for how much tomography is needed.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 29's proof omits nine of the ten additivity cases, leaving the complete characterization of three-qubit 2-UDA additivity unsupported.","rationale":"The reader's stated weakest assumption is Lemma 6, but my analysis indicates that Lemma 6 is not a load-bearing flaw: the purification step is a standard fact and the use of n-UDA is correct because σ_second indeed has the same n-marginals as ρ_second. The more serious gap is in Theorem 29, a result explicitly listed among the paper's strongest claims. Its proof reduces to a ten-case analysis but solves only one case and asserts the others without details. Since the theorem provides a 'complete characterization' of three-qubit 2-UDA additivity, the missing case analysis leaves a central claim unverified. My proposed test would close this gap by writing out the remaining cases and checking them against the stated conditions. The verdict should remain CONDITIONAL: the paper is plausible but this proof omission needs to be repaired before the result is accepted as established.","tokens_in":34779,"tokens_out":28408,"duration_ms":258679,"concrete_test":"Write out all ten pairs (α_j, β_k) from (18)–(25). For each pair, derive the most general state σ compatible with α_j⊗β_k via the 2-marginals, following the method used for α1⊗β1 (Lemma 2 plus Theorem 22). Check whether σ = α_j⊗β_k exactly when the corresponding condition in Theorem 29 (i)–(iv) holds. If any pair yields a compatibility condition not covered by (i)–(iv), the theorem needs modification; if all match, the proof gap is purely expositional.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central additivity claim (Theorem 29) is not actually proved. The proof enumerates four normal forms for each factor (α1–α4, β1–β4), asserts that this yields ten cases, analyzes only α1⊗β1, and dismisses the remaining nine with 'Similarly.' The theorem's final conditions (i)–(iv) are intricate, depending on which marginals are pure, which two-qubit factors have rank two, and on local unitary equivalences. Nothing in the text demonstrates that the other nine pairs lead exactly to the stated conditions, nor that no additional class of 2-UDA tensor products exists. Without this case analysis the 'complete characterization' of three-qubit 2-UDA additivity is unsupported. This is distinct from the reader's concern about Lemma 6: that lemma's purification step is standard (any state with the same first marginal can be written as I⊗Λ on a purification, and σ_second shares all n-marginals with ρ_second by compatibility), so that lemma is likely repairable. The missing Theorem 29 cases are a larger, and more directly central, gap.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies states uniquely determined among all states (UDA) by their k-partite marginals. It proves structural lemmas about ranges, faces, and compatibility; gives complete characterizations of 2-UDA product states for three qubits (Theorem 15), four qubits (Theorem 22), and five qubits (Theorem 24); establishes a high-rank non-UDA statement (Proposition 26) and a corollary that almost all mixed states are not k-UDA (Corollary 27); and states complete characterizations of additivity for bipartite 1-UDA states (Theorem 28) and three-qubit 2-UDA states (Theorem 29). A systematic procedure for deciding k-UDA is also proposed.","tokens_in":35116,"tokens_out":7597,"duration_ms":79163,"significance":"If all the main claims are correct, the paper would provide a useful toolbox for local-marginal reconstruction and certification: explicit necessary and sufficient conditions for several product-state families, a recursive reduction method, and a sharp high-rank obstruction. Several auxiliary results (Lemma 5, Lemma 8, Proposition 9) are elementary and generally well argued. The advertised complete characterization of additivity in Theorem 29, however, is not actually proved, and the proof of the high-rank obstruction in Proposition 26 contains a substantial gap. These issues affect the central claims of the paper, so the work is not yet ready in its present form.","major_comments":[{"comment":"The proof of the main additivity theorem is incomplete. After listing four normal forms for each factor, the text states that this gives ten cases and that 'the following analysis holds, up to swapping'; it then analyzes only α1⊗β1 and dismisses the remaining nine cases with 'Similarly.' The conditions (i)–(iv) of Theorem 29 are intricate, depending on which marginals are pure, which two-qubit factors have rank two, and on local unitary equivalence. Nothing in the manuscript demonstrates that the other nine pairs lead to exactly these conditions, nor that no additional class of 2-UDA tensor products exists. As written, the claimed complete characterization of three-qubit 2-UDA additivity is unsupported.","section":"Section IV, Theorem 29"},{"comment":"The proof of Lemma 6 is under-justified and is used later in Proposition 20(iii) and Theorem 22(iv). The text says 'For any n-partite channel Λ ... we have σ = I⊗Λ(|ψ><ψ|)'; the quantifier should be existential: every compatible σ can be represented in that form for some Λ. This requires a purification argument that is not supplied. The step 'Since ρ_{An+1...Am} is n-UDA, we have σ_{An+1...Am}=ρ_{An+1...Am}' also needs an explicit justification that compatibility of σ and ρ at the global level implies n-compatibility of σ_{An+1...Am} with ρ_{An+1...Am}. Finally, the conclusion M_j|β_k> ∈ R(ρ_{An+1...Am}) does not follow immediately from σ_E=ρ_E and should be derived. Because the lemma is load-bearing for the four-qubit characterization, this gap must be repaired.","section":"Appendix A, Lemma 6"},{"comment":"The proof of the rank D−1 case is too sketchy to be verifiable. For n qubits, it invokes an unstated algebraic-geometry fact from [29] about the kernel of a seven-dimensional subspace, asserts the existence of a range vector LU-equivalent to a quasi-GHZ state, and then says 'Using the proof for n=3' for n>3 and for unequal local dimensions. The exact statement needed from [29] is not given, and the reduction to general dimensions is not spelled out. The full-rank part also asserts without proof the existence of a nonzero Hermitian χ with vanishing k-partite marginals. Since Proposition 26 supports the high-rank non-UDA claim and the measure-zero discussion leading to Corollary 27, this needs either a rigorous proof or a precise reference with the exact lemma stated.","section":"Section III E, Proposition 26"}],"minor_comments":[{"comment":"The statement has indexing errors: condition (ii) refers to ρ_A5 in a four-qubit system, and condition (iii) uses undefined indices n, m1, m2. The labeling of H and D in equations (8), (15), and in the theorem statements is inconsistent and should be unified.","section":"Theorem 22"},{"comment":"The text 'if rank(ρ_A1A2) = and rank(ρ_A3A4) = 4' is missing the value of the first rank; it should presumably be 2. Please correct this typo.","section":"Appendix C, Proposition 21"},{"comment":"The operator α = D⊗ρ_A3A4 defined in equation (9) is not normalized. The argument can be applied after normalization, but this should be stated explicitly.","section":"Lemma 18"},{"comment":"The proof would be easier to follow if the channel Λ were defined as 'there exists a CPTP map Λ' rather than 'for any n-partite channel Λ'. Also, the notation R(ρ_{An+1...Am}) should be defined as the support of the reduced density operator on that subsystem, consistent with the rest of the paper.","section":"Appendix A, Lemma 6"},{"comment":"Several proofs in the appendices (e.g., Proposition 20(iii) and Lemma 23) are lengthy but do not state the simplifying assumption that local unitaries have been applied; adding a one-line 'up to LU equivalence' at the start of each would improve readability.","section":"General exposition"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a number of promising ideas and the low-rank characterizations of product states may be correct and valuable. However, the central additivity theorem (Theorem 29) is not proved, and Proposition 26 has a significant gap. I would encourage the authors to supply the missing nine-case analysis for Theorem 29 and a rigorous proof of the algebraic-geometry step in Proposition 26. If those can be provided, the paper could be suitable for publication; in its current form the main claims are not established."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. The paper's headline additivity theorem for three-qubit states (Theorem 29) is not proved: the proof analyzes only α1⊗β1 and dismisses the other nine cases with 'Similarly.' The theorem's conditions are intricate enough that this is a real gap, not a stylistic shortcut. Second, Lemma 6 — used in the four-qubit characterization — has a purification argument that doesn't work as written.\n\nWhat's actually new and good: Proposition 9 gives a clean iff for bipartite 1-UDA: one marginal pure. Theorem 15's three-qubit product-state characterization is a solid piece of linear algebra (assuming Propositions 13 and 14 hold; I checked the flow, not every determinant). Theorem 28's additivity result for bipartite states is neat and fully proved. The general lemmas about ranges, real/imaginary parts, and reversible channels are elementary but useful. The full-rank half of Proposition 26 is correct and enough for Corollary 27 — almost all mixed states are not UDA.\n\nSoft spots: The missing nine cases in Theorem 29 are load-bearing. The theorem claims a complete characterization; without the case analysis, the claim is unsupported. Lemma 6's claim that an n-UDA second factor forces range inclusion is not established by the proof. The step 'Since ρ_{A_{n+1}...A_m} is n-UDA, we have σ_{A_{n+1}...A_m}=ρ_{A_{n+1}...A_m}' confuses compatibility with UDA; and the channel on the purifying system doesn't automatically map the |β_k> into the range of ρ_{A_{n+1}...A_m}. This matters because Proposition 20(iii) and Theorem 22(iv) ride on it. I'd expect the lemma to be repairable — the claim is plausible and might follow from a more careful argument — but as written it's a gap. Proposition 26(i) for rank D-1 also relies on an unstated algebraic-geometry fact from [29]; the full-rank case is fine and is what the 'almost all' corollary actually needs.\n\nWho is this for: people working on UDA, local tomography, and marginal problems. The paper deserves a serious referee — the bipartite and three-qubit product characterizations are worth having — but the referee should demand a complete proof of Theorem 29 and a fix for Lemma 6 before publication. I'd send it to review, not desk reject.","headline":"Several UDA characterizations are plausible and the bipartite additivity theorem is clean, but the three-qubit additivity theorem is not proved and Lemma 6 is shaky.","tokens_in":35528,"tokens_out":6684,"would_cite":false,"duration_ms":60294,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P45","81P40"],"pacs":["03.67.-a"],"model":"deepseek-v4-flash","headline":"This paper establishes exact conditions under which a multipartite mixed quantum state is the only state compatible with its k-partite marginals, and proves a sharp rank threshold beyond which uniqueness is impossible.","keywords":["UDA states","quantum marginals","multipartite mixed states","unique determination","additivity","local quantum state tomography","rank threshold","quantum certification"],"falsifier":"An explicit rank-7 three-qubit mixed state whose three two-qubit marginals admit a unique compatible state would refute the high-rank claim of Proposition 26; generating random rank-7 states and solving the marginal polytope program is a direct search. For the additivity theorem, a concrete counterexample would be two bipartite 1-UDA states, each with one pure marginal but neither a pure product, whose tensor product is 1-UDA — Theorem 28 says such a pair cannot exist.","tokens_in":34679,"feed_emoji":"⚛️","tokens_out":6055,"duration_ms":77230,"temperature":0.7,"pith_summary":"Mixed quantum states are rarely fixed by their local marginals. The paper proves complete necessary-and-sufficient conditions for several families: a bipartite state is 1-UDA exactly when one of its 1-marginals is pure; product mixed states of three, four, and five qubits are 2-UDA under explicit rank-and-range conditions; and an n-fold tensor product is k-UDA exactly when at least n−k factors are pure. It also shows that any n-partite mixed state of rank at least d1...dn−1 is not k-UDA, which implies almost all mixed states fail this kind of certification. Finally it gives additivity criteria: the tensor product of two bipartite 1-UDA states is 1-UDA if and only if one factor is a pure product state, and it states the analogous three-qubit additivity rule. Why care: these conditions are directly checkable, so they turn the abstract UDA notion into a practical tool for local tomography and certification.","feed_headline":"Almost all mixed states are not uniquely fixed by their marginals","feed_subtitle":"Exact conditions reveal which low-rank product states are the exceptions, and when UDA additivity survives.","key_machinery":"The load-bearing mechanism is range inclusion for compatible states. Lemma 6 asserts that if ρ = ρ_A1...An ⊗ ρ_An+1...Am and the second factor is n-UDA, then every compatible σ satisfies R(σ) ⊆ R(ρ); combined with the vanishing of the k-marginals of χ = ρ−σ, this forces χ = 0 in the characterized cases. Lemma 5 (if a state in the face of R(ρ) is not UDA, ρ is not UDA) drives the high-rank negative results, and Proposition 9 (bipartite 1-UDA iff one marginal pure) seeds the additivity analysis. The explicit form D⊕0 for rank-two two-qubit states (Lemma 12) is the concrete normal form that makes the four- and five-qubit conditions checkable.","core_discovery":"On its own terms, the paper's central discovery is that UDA-ness of a mixed state is controlled by purity of its marginals and by a rank threshold. Proposition 9 shows a bipartite state is 1-UDA iff one of its 1-marginals is pure. Theorems 15, 22, and 24 give the full 2-UDA characterizations for three-, four-, and five-qubit product states: each condition is a statement about which factors or marginals are pure and about the rank and special form (D⊕0) of the non-product factors. Proposition 26 proves that rank ≥ d1...dn−1 excludes k-UDA for every k, with Corollary 27 saying that almost all n-partite mixed states are not k-UDA under Hilbert–Schmidt or Bures measures. The additivity theorems","pith_inferences":["A natural extension is to conjecture that the rank threshold d1...dn−1 holds for every n-partite state, not just product states; testing random high-rank states for unique compatibility would provide evidence.","If the range-inclusion lemma survives scrutiny, it yields a stronger rigidity statement: any state sharing the same k-marginals must have support contained in the original state's support; this could be probed experimentally by checking whether reconstructed states have nested supports.","The additivity theorems suggest an operational test for 'non-classicality' of the certification: if a bipartite 1-UDA state's tensor with any other 1-UDA state fails to be 1-UDA, then the state must be a pure product state, giving a device-like criterion.","The complete characterization of 2-UDA product states might be lifted to graph states or symmetric states by combining the range-inclusion argument with the symmetry constraints used for Dicke states."],"forward_implications":["For the characterized product families, deciding whether a state is k-UDA reduces to checking the rank and purity of its marginals, so local tomography with polynomial observable counts becomes possible exactly for those states.","The rank bound d1...dn−1 means any certification scheme that uses fixed-order k-marginals will fail on all sufficiently high-rank mixed states; such states require either more marginals or additional structural assumptions.","UDA additivity fails generically: combining two UDA mixed states by tensor product almost always loses uniqueness, so scalable certification must keep at least one factor pure or fall in the special classes of Theorem 29.","The three-step procedure (reduce systems, reduce rank, check components) gives a decision algorithm for k-UDA for the product families, and a construction method for non-k-UDA examples.","The additivity criteria imply that entanglement of a UDA state is certified by marginals alone; for tensor products of two mixed UDA states the marginals cannot certify genuine multipartite entanglement."],"fun_headline_variants":["High rank kills unique determination from margins","Pure marginal is key to uniquely fixed mixed states","UDA additivity fully characterized for product states","Almost all mixed states fail marginal uniqueness","Rank threshold decides when margins pin down state"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the second tensor factor of a product state can serve as a purification of the first factor, so that every compatible state can be written as I⊗Λ(|ψ⟩⟨ψ|) with the Kraus images inside the second factor's range; this purification compatibility is assumed rather than demonstrated.","fun_headline_variants_meta":{"raw":{"variants":["High rank kills unique determination from margins","Pure marginal is key to uniquely fixed mixed states","UDA additivity fully characterized for product states","Almost all mixed states fail marginal uniqueness","Rank threshold decides when margins pin down state"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000216,"raw_usage":{"total_tokens":1252,"prompt_tokens":712,"completion_tokens":540,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":456,"completion_tokens_details":{"reasoning_tokens":488}},"tokens_in":456,"tokens_out":540,"duration_ms":5909,"temperature":1.0,"reasoning_tokens":488,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T14:29:13.160949+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"An explicit rank-7 three-qubit mixed state whose three two-qubit marginals admit a unique compatible state would refute the high-rank claim of Proposition 26; generating random rank-7 states and solving the marginal polytope program is a direct search. For the additivity theorem, a concrete counterexample would be two bipartite 1-UDA states, each with one pure marginal but neither a pure product, whose tensor product is 1-UDA — Theorem 28 says such a pair cannot exist.","supporting_citations":[],"review_version":1}