{"id":"59dec710-11e4-40b2-892d-022b6d60eaf3","arxiv_id":"2607.16878","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For every n≥3 there exist bipartite states with identical local unitary Bargmann invariants that are not local-unitarily equivalent.","lead":"The authors construct pairs of two-qudit quantum states on C^n ⊗ C^n, for every n≥3, that share every local unitary Bargmann invariant yet are not related by any local unitary transformation. The result settles a conjecture by Zhang and Xie about whether these invariants give a complete classification of local equivalence.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the terse step in Lemma 2.4 is correct, so the counterexample survives scrutiny.","rationale":"The reader's verdict identifies the dim A=3 subcase of Lemma 2.4 as the weakest assumption, and I agree that this is the least-justified step in the proof. However, the concern does not land: the alleged gap is real in the sense that the authors did not spell out the argument, but the assertion is true and easily proved. The span of two orthogonal rank-one projections and the identity is a commutative algebra, so any two projections in it commute. The rest of the proof is sound: the dim A=4 case uses the invertibility of D1 versus non-invertibility of D2' via Lemma 2.1, which is correctly applied; the dim A=5 case gives D1=0, impossible for n≥3; and the final normalization preserves the counterexample because adding cI commutes with every unitary and with the flip, and if the normalized states were locally unitarily equivalent, the unnormalized ones would be too. Thus the central claim of Theorem 2.6 stands. The reader's ACCEPT verdict with moderate confidence is appropriate; no adjustment is needed.","tokens_in":8040,"tokens_out":19139,"duration_ms":157700,"concrete_test":"Symbolically verify the dim A=3 step: work in a basis where E1=diag(1,0,...,0) and E2=diag(0,1,0,...,0). Enumerate all rank-one projections in span{I,E1,E2}: for m>3 these are exactly E1 and E2; for m=3 also I-E1-E2. Check that every pair commutes by computing their commutator. This confirms the 'It follows' assertion in Lemma 2.4.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most delicate step is the dim A=3 subcase of Lemma 2.4, where the authors assert without proof that two rank-one projections lying in span{I,E1,E2} must commute. This is load-bearing because it closes the contradiction when the span is 3-dimensional; if false, the non-local-unitarity conclusion would fail. However, the assertion is true: E1 and E2 are orthogonal rank-one projections, so E1E2=E2E1=0 and E1^2=E1, E2^2=E2. Hence span{I,E1,E2} is a commutative algebra, and any two elements of it—in particular any two projections—commute. The remaining cases (dim A=4 and dim A=5) are handled correctly: the invertibility/non-invertibility contradiction from Lemma 2.1 is valid, and the linear-independence argument via Lemma 2.3 is sound. The normalization step in Theorem 2.6 preserves both the W-relations and non-local-unitary equivalence because adding a multiple of the identity commutes with all local unitaries and with the flip. No other load-bearing gap was identified.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies local unitary (LU) equivalence of bipartite quantum states on C^n ⊗ C^n and asks whether the local unitary Bargmann invariants—the Bargmann invariants of the triple (ρ_AB, ρ_A⊗I_n, I_n⊗ρ_B)—form a complete set of LU orbit invariants. The authors answer this negatively for every n ≥ 3. Their main result, Theorem 2.6, explicitly constructs two states ρ_AB, σ_AB with maximally mixed marginals and a global unitary W such that σ_AB = W ρ_AB W*, σ_A⊗I = W(ρ_A⊗I)W*, and I⊗σ_B = W(I⊗ρ_B)W*, hence the two states have identical local unitary Bargmann invariants, while σ_AB and ρ_AB are not locally unitarily equivalent. The construction uses rank-one projections, a shift by a multiple of the identity to normalize, and Sakai's flip automorphism. The paper also proves in Proposition 2.9 that for pure bipartite states, the same conditions do force local unitary equivalence, so the incompleteness is specific to mixed states.","tokens_in":8278,"tokens_out":14259,"duration_ms":125143,"significance":"If correct, the paper resolves an open problem posed by Zhang and Xie [15, Section 8.2] and gives a clean negative answer to the completeness of local unitary Bargmann invariants for n ≥ 3. This is a useful and nontrivial contribution. The construction is explicit and self-contained, with no free parameters and no circularity; the use of the flip automorphism and the normalization by adding the identity are elegant and make the counterexample robust. The paper also identifies a positive result for pure states, which helps delimit the boundary of the phenomenon. The main proof is sound, and the only weaknesses are terse justifications at a few local steps.","major_comments":[],"minor_comments":[{"comment":"The proof states 'It follows that E'_1E'_2 = E'_2E'_1' without justification. The assertion is true because span{I, E1, E2} is commutative: E1E2 = E2E1 = 0, E1^2 = E1, and E2^2 = E2. Since this step closes the contradiction in the 3-dimensional case, the authors should spell out this one-line reason explicitly.","section":"§2, Lemma 2.4, dim A = 3 case"},{"comment":"The final step 'From this, it is routine to see that σ_AB and ρ_AB are locally unitarily equivalent' is too terse. Given that the paper's main result is a non-LU construction, this positive pure-state claim deserves a short proof: the two marginal equations force W to map the Schmidt subspaces span{e_j⊗f_l : l} and span{e_l⊗f_j : l} to the corresponding primed subspaces, so W acts as U⊗V on the Schmidt subspace. This is especially pertinent when Schmidt coefficients are degenerate.","section":"§2, Proposition 2.9"},{"comment":"The bound ∥ρ_AB∥ ≤ 2n+2 is asserted without derivation. It is correct—for instance B1 and B2 have norm at most n−1 and B3 has norm at most 2—but a one-line justification would make the normalization step fully transparent.","section":"§2, Theorem 2.6"},{"comment":"The linear independence of {Q1, P2, I_n} is stated compactly. Since this is needed for the application of Lemma 2.3, a brief explanation (rank arguments for αQ1 + βI_n) would improve readability.","section":"§2, Example 2.8"}],"recommendation":"minor_revision","confidential_remarks":"The paper is mathematically sound and the negative result is significant. The terse steps noted in the minor comments are all true and locally fixable; I do not see any load-bearing gap. The manuscript is suitable for publication after a small revision that expands these justifications."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper settles an open question from the Zhang–Xie survey: for bipartite states on C^n ⊗ C^n with n ≥ 3, local unitary Bargmann invariants are not a complete set of invariants for local unitary equivalence. The authors give explicit two-qudit states ρ_AB and σ_AB, for every n ≥ 3, that have the same three Bargmann invariants (equivalently, are related by a global unitary W conjugating ρ_AB, ρ_A⊗I, I⊗ρ_B) yet are not locally unitarily equivalent. This is a new counterexample to the conjecture, as far as the citations show.\n\nThe construction is genuinely self-contained. Start with two rank-one projections P1, P2 that are orthogonal and Q1, Q2 that don't commute; build a self-adjoint matrix with maximally mixed marginals; apply Sakai's flip; add a multiple of the identity to make it a state. The W-relations hold by construction, and Lemma 2.4 rules out any local unitary intertwinement. I checked the operator-algebra steps: Lemma 2.1's invertibility criterion is correct, and the use of Lemma 2.3 is legitimate. There are no free parameters and no circularity.\n\nThe soft spots are presentation, not substance. The dim A = 3 case in Lemma 2.4 says two projections in span{I, E1, E2} must commute, which is stated as if obvious. It is true: since E1E2 = E2E1 = 0, the span is a commutative algebra. But the reader has to fill that in. The 'routine' step in Proposition 2.9 is also terse; the claim that Schmidt coefficients match is fine, but the local unitary connecting the two pure states is not written down. Neither threatens the main theorem. Example 2.8 is a nice simplification for n ≥ 4.\n\nWho should read this: anyone working on local unitary invariants or quantum state classification, especially people who cited the conjecture as a plausible route. It is a focused no-go theorem rather than a new framework, but the counterexamples are simple enough to be useful. I would send it to a serious referee, asking for a fuller write-up of Lemma 2.4 and Proposition 2.9. The math itself holds.","headline":"Explicit counterexamples show Bargmann invariants don't classify LU orbits for two qudits of dimension ≥3; the proof is sound, with only terse presentation to polish.","tokens_in":8731,"tokens_out":2280,"would_cite":true,"duration_ms":22151,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P45","81P40","15A30"],"pacs":[],"model":"deepseek-v4-flash","headline":"For dimension three and above, local unitary Bargmann invariants do not determine local unitary equivalence.","keywords":["Bargmann invariants","local unitary equivalence","two-qudit states","quantum state invariants","partial traces","rank-one projections","unitary orbits","mixed-state entanglement"],"falsifier":"Find four rank-one projections E1, E2, E'1, E'2 in M_3(C) such that E1E2=0, E'1 and E'2 do not commute, and both E'1 and E'2 lie in span{I, E1, E2}; such a quadruple would disprove the key subcase of Lemma 2.4 and invalidate the main counterexample.","tokens_in":7905,"feed_emoji":"🧩","tokens_out":5804,"duration_ms":53055,"temperature":0.7,"pith_summary":"The paper proves that the natural set of numbers called local unitary Bargmann invariants is not enough to distinguish two-qudit states up to local unitary transformations. For every dimension n≥3, the authors construct two density matrices on C^n⊗C^n that share all the same local unitary Bargmann invariants yet are not locally unitarily equivalent. This answers a question posed in the recent literature, which had conjectured that these invariants might be complete. The construction relies on states built from rank-one projections, arranged so that both partial traces are scalar, and on a tensor-flip automorphism that produces the second state from the first. The result means that any complete set of local unitary invariants for mixed states must contain information beyond these algebraic traces.","feed_headline":"Local Bargmann invariants fail for two-qudit states when n≥3","feed_subtitle":"A new pair of states shares all these invariants yet is not locally unitarily equivalent, so extra invariants are needed.","key_machinery":"The construction uses rank-one projections P_1,P_2 and Q_1,Q_2 chosen so that P_1P_2=0 while Q_1 and Q_2 do not commute, then assembles ρ_AB = P_1⊗B_1 + P_2⊗B_2 + I⊗B_3 with B_i defined from Q_1,Q_2 so that both partial traces are scalar multiples of the identity. The tensor-flip automorphism, which swaps the two tensor factors, supplies the global unitary W that relates ρ_AB to σ_AB. Lemma 2.4 is the technical heart: it says that if such a σ were locally unitarily equivalent to ρ, the projections appearing in the two expressions would have to satisfy incompatible commutation relations, forcing a contradiction.","core_discovery":"The central discovery is a family of counterexamples: for each n≥3 there are density matrices ρ_AB and σ_AB on C^n⊗C^n and a global unitary W such that σ_AB = Wρ_ABW*, σ_A⊗I_n = W(ρ_A⊗I_n)W*, and I_n⊗σ_B = W(I_n⊗ρ_B)W*, yet no local unitary U⊗V maps ρ_AB to σ_AB. Because equality of all local unitary Bargmann invariants is equivalent to the existence of such a W—a classical invariant-theory result the paper cites—this shows the invariants are not complete for n≥3. The counterexamples are built by choosing rank-one projections P_1,P_2 that are orthogonal and Q_1,Q_2 that do not commute, forming ρ_AB from tensor products with matrices made from Q_1,Q_2 so that both partial traces are scalar, a","pith_inferences":["A likely lesson is that any complete set of local unitary invariants for mixed states must encode the tensor-product structure more delicately than the partial-trace data used in the Bargmann triple.","The counterexample is a finite-dimensional analogue of classical operator-algebra phenomena where global equivalence of direct sums does not imply local equivalence; this paper gives a concrete quantum-information instance that may be used to test similar conjectures in other settings.","One could test whether the failure persists for states of restricted rank or fixed entanglement structure; the paper's examples use full-rank states after normalization, so it is open whether low-rank states are also resistant to this invariant set.","A practical extension is to search numerically for the smallest pair of states that are distinguishable only by invariants beyond the Bargmann set, which could guide the construction of practical nonlocal-basis detectors."],"forward_implications":["For any local dimension n≥3, the local unitary Bargmann invariants do not form a complete set of invariants for local unitary orbits; additional invariants are required.","The incompleteness extends to rectangular bipartite systems C^m⊗C^n whenever the greatest common divisor of m and n is at least 3 (as noted in Remark 2.7).","For pure states, the situation is different: the paper's Proposition 2.9 shows that local unitary Bargmann invariants do determine local unitary equivalence, so the failure is specific to mixed states.","For n≥4, the paper provides a simplified explicit example (Example 2.8) using projections of ranks 1 and 2, making the counterexample easier to verify.","The result closes a proposed problem in the literature and shifts attention to finding alternative complete descriptions of local unitary orbits in higher dimensions."],"fun_headline_variants":["Bargmann invariants not complete for n≥3 local unitary equivalence","Same Bargmann invariants, yet no local unitary map for these states","Counterexample: Bargmann invariants fail to distinguish local orbits","Local unitary orbits require extra invariants beyond Bargmann"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The counterexample rests on the unproved assertion in Lemma 2.4 that two rank-one projections lying in the three-dimensional span of the identity and two mutually orthogonal rank-one projections must commute; if this assertion were false, the argument that the constructed states are not locally unitarily equivalent would break.","fun_headline_variants_meta":{"raw":{"variants":["Bargmann invariants not complete for n≥3 local unitary equivalence","Same Bargmann invariants, yet no local unitary map for these states","Counterexample: Bargmann invariants fail to distinguish local orbits","Local unitary orbits require extra invariants beyond Bargmann"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000179,"raw_usage":{"total_tokens":1100,"prompt_tokens":668,"completion_tokens":432,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":412,"completion_tokens_details":{"reasoning_tokens":359}},"tokens_in":412,"tokens_out":432,"duration_ms":4361,"temperature":1.0,"reasoning_tokens":359,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T19:41:48.313234+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find four rank-one projections E1, E2, E'1, E'2 in M_3(C) such that E1E2=0, E'1 and E'2 do not commute, and both E'1 and E'2 lie in span{I, E1, E2}; such a quadruple would disprove the key subcase of Lemma 2.4 and invalidate the main counterexample.","supporting_citations":[],"review_version":1}