{"id":"ff69092b-c102-456a-887b-a782cd183c14","arxiv_id":"2501.07084","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The convex roof of the square root of the three-tangle is solved exactly for all GHZ-W mixtures of three qubits, using a new zero-state locking theorem.","lead":"This paper claims to compute exactly how much genuine three-particle entanglement, measured by the square root of the three-tangle, remains in any mixture of the GHZ and W states of three qubits. If correct, it is one of the few exact solutions of the convex-roof problem in entanglement theory.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unproven 'intransparency' (non-intersection of optimal decompositions) is used to exclude (0,n_e) decompositions; the zero-state locking proof only shows a local minimum, not global optimality, so the claimed completeness of the convex roof is not established.","rationale":"I read the paper in good faith: it presents a plausible exact solution, with a genuine attempt at a zero-state locking theorem, a new inequality, and a detailed geometric classification. The strongest_claim is conditional on a complete proof and reproducible numerics. The weakest point is not the algebra of the tangle but the global structural assumption: the proof of zero-state locking establishes that a zero-state is a local minimizer among decompositions that include it, yet the exclusion of (0,n_e) decompositions requires a global statement that every visible zero-state is in every optimal decomposition. The non-intersection premise is not derived from the convex roof definition; it is stated as a property. Because the central claim is completeness over the whole Bloch sphere, one counterexample region with a lower (0,2) decomposition would invalidate the exactness. The paper's own inequality is only local and evaluated at a single point in the published calculation, so it does not close the gap. I therefore agree with the reader's call for a complete proof of the structural assumptions and would keep the verdict CONDITIONAL until a numerical or analytic test settles whether (0,2) decompositions ever beat the claimed ones. The proposed grid search is practical: rank-two convex roof with at most four pure states is a low-dimensional global optimization, and a match over a fine grid would materially increase confidence.","tokens_in":13851,"tokens_out":7560,"duration_ms":74492,"concrete_test":"On a dense grid of rho[p,phi] in the wedge phi in [-pi/3, pi/3], numerically minimize sum_i lambda_i sqrt(tau_3(psi_i)) over lambda_i >= 0 with sum lambda_i = 1 and psi_i on the Bloch sphere (i <= 4 by Carathéodory) subject to sum lambda_i rho(psi_i) = rho[p,phi]. Use multi-start global optimization (e.g., differential evolution) from random seeds plus the paper's claimed optimal decompositions. If any minimum is strictly below the paper's value, the intransparency/zero-locking assumption is falsified; matching minima would corroborate the structure.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — an exact convex roof for every state in the Bloch sphere — rests on the assertion (Section II) that optimal decompositions cannot intersect, making them 'intransparent', and that every visible zero-state must belong to the optimal decomposition. The proof of zero-state locking (Eqs. 1–2) demonstrates only that shifting a zero-state off the zero polytope locally increases the average tangle for m<d; it does not compare against a decomposition that omits the zero-state altogether. The step 'This property excludes the possibility of optimal (0,n_e) decompositions' relies on non-intersection, which is asserted rather than proved; Refs. [32,33] bound the number of states (Carathéodory/Uhlmann) but not disjointness of optimal decompositions. The inequality in Eq. 6 is local and is evaluated only at p0 (Eq. 9) for the specific plane; it does not rule out (0,2) optimality elsewhere. If a (0,2) decomposition were optimal in any region, the claimed tiling into (3,1), (2,1), and (1,1) types would miss it and the 'exact' roof would be too high.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims to give an exact solution of the convex roof of the square root of the threetangle for rank-two density matrices, in particular for all states in the Bloch sphere spanned by the GHZ and W states. The solution is described by a zero polytope, four three-dimensional tetrahedra (three of them new, with tips |N1>, |N2>, |N3>), (2,1) decompositions along certain circles, and (1,1) decompositions covering the rest. The two main tools are a \"zero-state locking\" theorem, proved by a Taylor-expansion argument in Section II, and an inequality (Eq. (6)) that is claimed to decide between (2,1) and (0,2) decompositions. The paper also claims that SL-invariance extends the result to all states in the SL-class of GHZ and W, and it presents numerical evidence from a grating with errors below 4 per mille.","tokens_in":14178,"tokens_out":4385,"duration_ms":45420,"significance":"If the claimed exact convex roof is correct, this is an important result: exact convex roofs for multipartite entanglement measures are rare, and the problem is NP-hard in general. The geometric classification into zero polytope, tetrahedra, curves, and remaining (1,1) decompositions is explicit and could be used to benchmark numerical convex-roof algorithms. The paper also gives a proof attempt of zero-state locking, a structural property that is interesting in its own right. However, the manuscript does not, as written, establish the global optimality that the word \"exact\" requires: the key arguments are local or rely on an unproved geometric assertion about non-intersection of optimal decompositions. The paper does not ship machine-checked proofs or reproducible code; the numerical grating, while suggestive, is not a proof.","major_comments":[{"comment":"The zero-state locking proof is local: it shows that if an optimal decomposition contains a zero-state, then shifting that state slightly increases the average tangle for m<d. It does not show that a decomposition that omits the zero-state entirely cannot have lower or equal average tangle. Therefore the statement that every visible zero-state must belong to an optimal decomposition is not established. This is load-bearing because the subsequent classification into (n_z,n_e) types with the maximal number of zero-states uses it.","section":"Section II, Eqs. (1)-(2)"},{"comment":"The exclusion of (0,n_e) decompositions with n_e>1 rests on the assertion that optimal decompositions cannot intersect and can be viewed as intransparent. This assertion is stated but not proved, and the cited Refs. [32,33] only bound the number of pure states in an optimal decomposition; they do not imply disjointness of different optimal decompositions. Without a proof of intransparency, the claimed completeness of the catalog (zero polytope plus four tetrahedra plus (2,1) curves plus (1,1) decompositions) is not established: other decomposition types could lower the convex roof in some regions.","section":"Section II, paragraph 'Optimal decompositions do not intersect...'"},{"comment":"Inequality (6) is derived from a second-order Taylor expansion around a zero-state and is evaluated only at p0 for one particular plane (Eq. (9)). It is not shown to be valid globally over the entire Bloch sphere. The paper itself later states that it would be interesting to find examples where the inequality points towards (0,2) decompositions in parts of the Bloch sphere. Consequently, the claim that (2,1) decompositions are always optimal and (0,2) decompositions never are optimal is not proven, and the exactness of the resulting convex roof remains conditional.","section":"Section 'n_z,n_e-decompositions with n_e>1', Eq. (6)"},{"comment":"The identification of the circular arcs and their distance 0.0711148 from the Bloch-sphere center is justified by a numerical grating with errors smaller than 4 per mille, together with a check of orthogonality of the derivative. Since these arcs are part of the claimed exact classification, the manuscript should either provide an analytic derivation or a rigorous error bound for these quantities. Numerical evidence alone does not establish an exact convex roof.","section":"Appendix C, 'Optimal (2,1) decompositions in between...'"}],"minor_comments":[{"comment":"The text switches between first-person singular and plural (\"I\" vs. \"we\") without a consistent convention; this should be harmonized.","section":"Section II"},{"comment":"The phrase \"the two complex conjugated yero states\" contains a typo: \"yero\" should be \"zero\".","section":"Figure 2 caption"},{"comment":"There is an unbalanced parenthesis in Eq. (B4), which makes the displayed formula difficult to parse; please check the typesetting.","section":"Appendix B, Eq. (B4)"},{"comment":"The text refers to \"the right panel in Fig. 1\" when marking the states M_i, but Fig. 1 in the main text appears to have no right panel; the intended reference may be to Fig. 4. Please correct the cross-references.","section":"Appendix C"},{"comment":"Several reference entries contain typographical artifacts, such as \"D– okovi´ c\" and \"A VS Quant. Sci.\"; these should be cleaned up.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript has a credible target result and the numerical evidence suggests the classification may be correct, but the published version must close the gap between local analysis and global optimality. In particular, the intransparency/non-intersection assertion and the global validity of Eq. (6) need proofs or at least a precise statement of which parts are conjectural. I would not reject the paper on this basis, because the gaps may be fixable within the scope of the manuscript, but I cannot accept it in its current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is a genuine extension of the known two-tetrahedron picture for the GHZ-W mixture to a full Bloch-sphere solution. The paper adds three new tetrahedra, the connecting (2,1) curves, and a (1,1) tiling, and it proves the zero-state locking theorem in-house rather than importing it. The inequality in Eq. 6 gives a concrete criterion for when a (2,1) decomposition beats a (0,2) one. If the construction is correct, it is the first exact convex roof for a genuine multipartite measure on a non-trivial family beyond the two-qubit concurrence case. That is a real result.\n\nSecond, the completeness claim is not as watertight as the abstract suggests. The load-bearing step is the assertion that optimal decompositions cannot intersect ('intransparency'), stated in Section II and used to rule out (0,n_e) decompositions. I do not see a proof of that in the paper; Carathéodory and Uhlmann bound the number of states, not disjointness of optimal decompositions. The zero-state locking proof is a local Taylor-expansion argument: it shows a first-order increase when a zero-state is shifted off the polytope for m<d. That plausibly becomes global by continuity, but the paper does not give the global argument. The inequality in Eq. 6 is evaluated at one point, p0, on one plane. And the small-circle curves are located numerically on a grating with errors below 4 per mille; there is no code or data to reproduce them. These are not fatal, but they mean 'exact' is conditional on a structural assumption that is plausible but unproven.\n\nThe citation pattern is healthy: the working horse is proved here, and the earlier results (zero polytope, GHZ tetrahedron) are independent inputs. No circularity.\n\nThe paper deserves a serious referee. The referee should push for a proof of non-intersection, a global version of zero-state locking, and a reproducible numerical pipeline. I would send it to review expecting major revisions. My own verdict: the construction is likely right, but the proof of completeness needs work.","headline":"A plausible exact convex roof for the GHZ-W family that extends the known two-tetrahedron picture, but the completeness argument leans on an unproved non-intersection assumption and a local proof of zero-state locking.","tokens_in":14619,"tokens_out":3699,"would_cite":true,"duration_ms":34010,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P40"],"pacs":["03.67.Mn"],"model":"deepseek-v4-flash","headline":"For every rank-two mixture of the three-qubit GHZ and W states, this paper determines the exact convex roof of the square root of the threetangle and describes every optimal decomposition.","keywords":["convex roof","threetangle","GHZ-W mixture","zero-state locking","entanglement measure","Bloch sphere","rank-two density matrix","SL-invariant tangle"],"falsifier":"Take any rank-two GHZ-W state and run a global numerical minimization of the average $\\sqrt{\\tau_3}$ over all pure-state decompositions. If any decomposition that excludes all visible zero-polytope states returns a value below the paper's (2,1)/(1,1) curve, the claim of completeness is false; the inequality $|\\tau''(0)|<\\tau_0\\rho/d_1$ identifies exactly the region where such a counterexample would have to live.","tokens_in":13667,"feed_emoji":"⚛️","tokens_out":9804,"duration_ms":87361,"temperature":0.7,"pith_summary":"Entanglement measures on mixed states are usually defined through a convex roof: a minimization over all pure-state decompositions of a state, a problem that is generally NP-hard. This paper proves that for the family of rank-two density matrices mixing the three-qubit GHZ state with the W state, the convex roof of the square root of the threetangle $\\sqrt{\\tau_3}$ can be evaluated exactly over the entire Bloch sphere. The solution is geometric: optimal decompositions must contain as many zero-tangle states as are visible, a property the paper calls zero-state locking, and the remaining optimal pure states are confined to four tetrahedra, a set of circles, and one-parameter families. This converts one nontrivial NP-hard minimization into an explicit closed-form prescription and yields a complete exact convex roof for a genuine multipartite entanglement measure on a full rank-two family.","feed_headline":"Exact entanglement found for every GHZ-W mixture","feed_subtitle":"A complete geometric solution for the square root of the threetangle covers the whole Bloch sphere.","key_machinery":"The working horse is the zero-state locking theorem: in any optimal decomposition, every pure state on the zero polytope that can be obtained as a convex combination of the density matrix's eigenstates must itself appear in the decomposition. This is proved from the scaling $\\tau_d \\propto (z-z_j)^{m/d}$ of an SL-invariant tangle near a root, which makes a shifted zero state raise the average tangle when the root multiplicity satisfies $m<d$. Together with the assertion that optimal decompositions are intransparent and cannot intersect, this reduces the candidates to $(n_z,n_e)$ simplices with at most four vertices. The paper also derives the inequality $|\\tau''(0)|>\\tau_0\\rho/d_1$ that decides whether a $(2,1)$ decomposition beats a $(0,2)$ one, and evaluates it for the GHZ-W Bloch sphere.","core_discovery":"The paper's central claim is that the convex roof of $\\sqrt{\\tau_3}$ for $\\rho[p] = p\\,|GHZ\\rangle\\langle GHZ| + (1-p)\\,|W\\rangle\\langle W|$ is completely determined by a small set of decomposition types. The optimal decomposition consists of the zero polytope, four three-dimensional tetrahedra — one with the GHZ state as its tip and three new ones with tips $|N_1\\rangle, |N_2\\rangle, |N_3\\rangle$ — three (2,1) curves on azimuthal grand circles, three (2,1) curves on small circles at distance 0.0711148 from the center of the Bloch sphere, and (1,1) decompositions covering the rest. The paper proves zero-state locking for root multiplicities $m<d$ and derives the inequality that decides when $(2,1)$ beats $(0,2)$; for GHZ-W mixtures the inequality is evaluated explicitly, ruling out optimal decompositions made only of entangled states. Because the measure is SL-invariant, the same structure transfers to every state in the SL class of the GHZ-W mixture.","pith_inferences":["The inequality (6) is derived in the GHZ-W setting but is a purely rank-two statement; applying it to other SL-invariant measures with different zero-root multiplicities could either confirm zero-state locking as a universal feature or produce the first known counterexamples.","If optimal decompositions are genuinely intransparent as assumed, exact convex roofs for rank-three and higher mixed states might be assembled from the rank-two faces of the state space, turning a global NP-hard search into local face-by-face convexification.","The reported curve data (circle distance 0.0711148 and normal vectors) provide a sharp quantitative benchmark; any numerical convex-roof routine should reproduce the linear tangle profile along exactly those curves."],"forward_implications":["Every rank-two GHZ-W mixture has an exactly computable $\\sqrt{\\tau_3}$, removing the need for numerical convex-roof searches for this family.","The optimal decomposition for a given $\\rho[p]$ is known in advance from its location on the Bloch sphere: which tetrahedron, circle, or one-parameter family supplies the minimizing pure states.","Zero-state locking gives a general necessary condition for optimality: any visible zero-tangle state must be in the decomposition whenever its root multiplicity is below the tangle's degree.","SL-invariance extends the solution from the single GHZ-W line to the whole SL-equivalence class, so the pattern covers all rank-two states in that class.","The same construction applies to symmetric mixtures of generalized GHZ and W states for any number of qubits."],"supporting_citations":[{"why":"It provides the known zero polytope and the first tetrahedron for the GHZ-W mixture that the new solution extends.","marker":"[28]"},{"why":"It establishes the (n_z,n_e) nomenclature and the behavior of optimal decompositions for rank-two density matrices on which the paper relies.","marker":"[23]"},{"why":"It supplies the zero-state locking concept that the paper proves and uses throughout.","marker":"[31]"},{"why":"It defines the threetangle whose square root is the entanglement measure being convexified.","marker":"[26]"},{"why":"It gives the SL-transformation rules for optimal decomposition states used to extend the solution to the full SL class.","marker":"[30]"},{"why":"It bounds the number of pure states in an optimal decomposition by four, restricting the search to three-dimensional simplices.","marker":"[32, 33]"},{"why":"It gives the root-multiplicity scaling (z-z_j)^{m/d} used in the zero-state locking proof.","marker":"[35]"},{"why":"It provides the exactly solvable concurrence convex roof whose decomposition structure serves as the model being generalized.","marker":"[24, 25]"}],"fun_headline_variants":["Exact threetangle roof for every GHZ-W mix","Full Bloch sphere: exact entanglement solution","Zero-state locking proves optimal decompositions","Entanglement measure solved for all GHZ-W states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole construction rests on the assertion that optimal decompositions are intransparent: they cannot intersect one another, and every zero-tangle pure state visible from the given density matrix must belong to the optimal decomposition for that state. If that fails, a decomposition built entirely from entangled states could sit below the claimed roof.","fun_headline_variants_meta":{"raw":{"variants":["Exact threetangle roof for every GHZ-W mix","Full Bloch sphere: exact entanglement solution","Zero-state locking proves optimal decompositions","Entanglement measure solved for all GHZ-W states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00065,"raw_usage":{"total_tokens":3004,"prompt_tokens":986,"completion_tokens":2018,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":602,"completion_tokens_details":{"reasoning_tokens":1959}},"tokens_in":602,"tokens_out":2018,"duration_ms":15399,"temperature":1.0,"reasoning_tokens":1959,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:49:38.203601+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take any rank-two GHZ-W state and run a global numerical minimization of the average $\\sqrt{\\tau_3}$ over all pure-state decompositions. If any decomposition that excludes all visible zero-polytope states returns a value below the paper's (2,1)/(1,1) curve, the claim of completeness is false; the inequality $|\\tau''(0)|<\\tau_0\\rho/d_1$ identifies exactly the region where such a counterexample would have to live.","supporting_citations":[{"cited_title":"Neveling and A","cited_arxiv_id":null,"evidence_quote":"It provides the known zero polytope and the first tetrahedron for the GHZ-W mixture that the new solution extends."},{"cited_title":"Pezze, A","cited_arxiv_id":null,"evidence_quote":"It establishes the (n_z,n_e) nomenclature and the behavior of optimal decompositions for rank-two density matrices on which the paper relies."},{"cited_title":"Osterloh and J","cited_arxiv_id":null,"evidence_quote":"It defines the threetangle whose square root is the entanglement measure being convexified."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It gives the SL-transformation rules for optimal decomposition states used to extend the solution to the full SL class."},{"cited_title":"Viehmann, C","cited_arxiv_id":null,"evidence_quote":"It gives the root-multiplicity scaling (z-z_j)^{m/d} used in the zero-state locking proof."}],"review_version":1}