{"id":"5564ae5f-ef25-4a3f-ac4d-4ec8efbf79f3","arxiv_id":"2608.07145","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For rank-2 mixtures of GHZ and W states, the convex-roof concurrence fill is exactly (5λ² - 4λ + 8)/9, derived through a new prism visualization of pure-state decompositions.","lead":"This paper introduces concurrence prisms as a geometric way to picture the pure-state decompositions of rank-2 three-qubit mixed states, and uses the picture to derive an exact entanglement formula for mixtures of GHZ and W states. The formula gives a clean benchmark for tripartite entanglement and shows how convex-roof minimization can be viewed as minimizing the total fill of prisms.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (26) converts complex-coefficient states (21b) to the real-amplitude formula (24) without proving phase-independence; the entire Jensen bound depends on this omitted step.","rationale":"I checked the rest of the derivation independently. The pure-state side-length formula Eq. (24) follows from the reduced density matrix with diagonal entries (4-q)/6 and (2+q)/6 and off-diagonal sqrt(q(1-q))/6, giving determinant (5q^2-4q+8)/36. The Jensen bound is valid because g(r)=(5 lambda^2 r^2 - 4 lambda r + 8)/9 is convex and E[r]=sum |v_{1k}|^2=1 by co-isometry; the V of Eq. (34) has r_k=1 and p_k=1/4, so it saturates. The numerical SLSQP check is supportive but not load-bearing. The only genuine gap is the unproven phase-independence in the step to Eq. (26); it is fixable and the appendix contains the needed trace values. Therefore the reader's ACCEPT stands, but the manuscript should state explicitly that the cross terms in Eq. (15) vanish for the GHZ-W pair before using Eq. (26). Agreement with the reader is partial: the reader identified the one-variable reduction as relying on the equilateral formula Eq. (24), but did not flag the complex-phase issue in the comparison between Eq. (21b) and Eq. (22).","tokens_in":11018,"tokens_out":21411,"duration_ms":182412,"concrete_test":"Directly evaluate the reduced density matrix for |psi>=alpha|GHZ>+beta|W> with alpha=sqrt(q)e^{i theta}, beta=sqrt(1-q); compute det rho_A and confirm it equals (5q^2-4q+8)/36 for all theta. Alternatively, insert the Appendix A values Tr(Lambda^{GHZ}_A X_A)=0, Tr(Lambda^W_A X_A)=0, Tr(X_A^2)=0 into Eq. (15) and verify all phase-dependent terms vanish, leaving C^2_{l(mn)}(|phi_k>)=(5 lambda^2 r_k^2 - 4 lambda r_k + 8)/9. If either check fails, Eqs. (26) and (35) do not follow.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Eq. (26) is load-bearing: it rewrites F(|phi_k>) for the complex-coefficient pure states of Eq. (21b) as a function of r_k=|v_{1k}|^2/p_k only, collapsing the convex roof to a one-variable Jensen problem. This conversion is not automatic. Eq. (22) is a real, nonnegative superposition sqrt(q)|GHZ>+sqrt(1-q)|W>, but Eq. (21b) allows arbitrary phases through v_{1k} and v_{2k}. The one-qubit determinant, via Eq. (15), generally contains phase-dependent terms -4Re((alpha beta*)^2 Tr(X^2)) and -8Re(|alpha|^2 alpha beta* Tr(Lambda^i X)) (and the beta-analogue). The main text asserts Comparing Eq. (22) with Eq. (21b), we set q=... without showing that these terms vanish for GHZ-W. Appendix A supplies Tr(X^2)=0 and Tr(Lambda^i X)=0, so the gap is fillable, but it is an omitted proof at exactly the point where the one-variable reduction happens. If the phase terms did not vanish, Eq. (26) would be false, the Jensen lower bound g(1) would not apply to the true F(|phi_k>), and the saturating co-isometry in Eq. (34) would not certify Eq. (35). Thus the closed form is only as secure as this phase-independence.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a geometric 'concurrence prism' picture for rank-2 three-qubit mixed states: each pure state in a decomposition gives a prism whose base is the concurrence triangle and whose height is the corresponding probability, and convex-roof concurrence fill is interpreted as the minimized total 'prism fill area.' The main technical result is a closed-form convex-roof evaluation for mixtures ρ(λ)=λ|GHZ⟩⟨GHZ|+(1−λ)|W⟩⟨W|, namely F_min(ρ(λ))=(5λ²−4λ+8)/9. The proof combines a one-variable reduction for pure states in the span of GHZ and W, Jensen's inequality, and an explicit co-isometry attaining the bound. Numerical SLSQP multi-start optimization is reported as corroboration, and a prism visualization of the optimal decomposition is presented.","tokens_in":11267,"tokens_out":6402,"duration_ms":61189,"significance":"Analytic convex-roof evaluations for multipartite mixed-state entanglement measures are rare, and the formula (5λ²−4λ+8)/9 is a concrete benchmark that correctly reproduces the pure-state endpoints F(GHZ)=1 and F(W)=8/9. The explicit saturating co-isometry and the explicit trace cancellations in Appendix A make the central result independently checkable. The prism picture is appealing as a geometric interpretation of the convex-roof minimization, although it is more a visualization than a new computational method. The numerical SLSQP check is supportive but not a substitute for the analytic proof.","major_comments":[{"comment":"The reduction of F(|φ_k⟩) to a function of r_k=|v_{1k}|²/p_k is asserted by comparing Eq. (22) with Eq. (21b), but Eq. (21b) has arbitrary complex coefficients v_{1k}, v_{2k}. Equation (15) contains explicitly phase-dependent terms −4Re((αβ*)²Tr X²) and −8Re(|α|²αβ*Tr Λ^i X); the equivalence to Eq. (26) holds only because Appendix A provides Tr(X²)=0 and Tr(Λ^i X)=0 for the GHZ-W pair. This cancellation is not shown at the point where the one-variable Jensen problem is set up. Please add an explicit evaluation of Eq. (15) for |ψ⟩=α|GHZ⟩+β|W⟩, demonstrating that C²=(5q²−4q+8)/9 with q=|α|², independent of the phase of αβ*. This step is load-bearing: without it, Eqs. (28)–(33) are not justified.","section":"Sec. IV, Eq. (26)"},{"comment":"The paper claims that dividing the eigenprisms into horizontal sections and adding the corresponding heights yields a geometric construction of the general prisms (Fig. 2). As written, this is a schematic description rather than a derivation: no theorem states that the base of the prism so constructed is the concurrence triangle of |φ_k⟩ from Eq. (8b). The algebraic content is Eq. (15); the geometric algorithm should either be proved or explicitly labeled a visualization heuristic. Since the title and abstract present the prism construction as a central contribution, this gap should be addressed.","section":"Sec. III, Eqs. (19a)–(19e)"}],"minor_comments":[{"comment":"The notation '|v^2_{1k}|' should be '|v_{1k}|²'.","section":"Sec. IV, Eq. (26)"},{"comment":"The subscript '123' in F^avg_123 should be 'ABC' for consistency with the rest of the paper.","section":"Sec. IV, Eq. (28)"},{"comment":"The quantity F=h√A is called 'prism fill area', but the optimization minimizes Σ_k p_k F(|φ_k⟩), which is an average prism-fill contribution; please define 'total prism fill area' unambiguously.","section":"Sec. II, Definition of prism fill area"},{"comment":"State explicitly that V in Eq. (34) is a co-isometry with orthonormal rows and that it gives p_k=1/4 and r_k=1 for all k, so that Jensen's inequality is saturated.","section":"Sec. IV, Eq. (34)"},{"comment":"The numerical SLSQP description is useful; please also report the spread of the minima over the 40 restarts, since SLSQP is a local optimizer and the reported points alone do not certify global optimality.","section":"Sec. IV, numerical section"}],"recommendation":"major_revision","confidential_remarks":"The main closed-form result appears sound once the missing phase-independence calculation is added; the Appendix A identities are sufficient to fill the gap. My main concern is the mismatch between the strong geometric claims in Sec. III and the schematic level of the construction given there. If the authors add the explicit cancellation proof and either prove or carefully reframe the geometric algorithm, the paper could be acceptable. The numerical check is reasonable but should not be presented as the primary evidence for the closed form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The real content in this paper is the exact convex-roof formula for the GHZ-W rank-2 mixture: F_min(λ) = (5λ² - 4λ + 8)/9. The derivation is correct: Jensen's inequality gives the lower bound, and the explicit rank-2 co-isometry in Eq. (34) saturates it. I verified the endpoints and the determinant algebra; everything lines up. The numerical SLSQP check is consistent, though they don't ship code.\n\nWhat's genuinely new is that closed-form formula. To my knowledge it doesn't appear in the cited literature. The prism visualization is a geometric repackaging of the convex-roof definition; it doesn't add predictive content, but it's a fair pedagogical tool and the paper doesn't sell it as more.\n\nThe soft spots are minor but real. The main one is Eq. (26). The transition from the real-coefficient state (22) to the complex-coefficient state (21b) is asserted with 'Comparing...', and a reader will legitimately wonder whether phase terms in Eq. (15) spoil the one-variable reduction. They don't: for this particular pair, Tr(X²)=0 and Tr(ΛX)=0 (Appendix A), and a direct determinant calculation gives exactly the same function of |α|². So the stress-test concern is fillable, but it should have been filled in the main text. This is an omitted proof at a load-bearing point, not a mistake.\n\nSecond, Sec. III is schematic. It describes the prism construction with words and figures rather than a formal statement. Since the actual result doesn't depend on the geometric construction's rigor, this is a presentation issue, not a correctness issue.\n\nThird, Eq. (15) is stated without derivation. For a specialized readership that formula is standard enough; still, a supplementary derivation would help.\n\nOverall: this is a sound, modest contribution to the convex-roof toolbox for three-qubit mixed states. It is not a breakthrough and the novelty is bounded, but the central formula is correct and useful for people quantifying tripartite entanglement. I'd send it to a serious referee and expect it to be accepted after minor revision. Bring to reading group? Maybe, if the group works on entanglement measures.","headline":"Solid closed-form convex roof for GHZ-W rank-2 mixtures; the prism picture is a repackaging, and the one omitted phase-independence step is fillable, so the result stands.","tokens_in":11835,"tokens_out":3277,"would_cite":true,"duration_ms":29130,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P40","81P45"],"pacs":["03.67.Mn"],"model":"deepseek-v4-flash","headline":"This paper extends the concurrence-triangle geometry to rank-two three-qubit mixed states, proving a closed-form concurrence fill for mixtures of GHZ and W states.","keywords":["concurrence triangle","concurrence fill","convex roof extension","three-qubit entanglement","GHZ state","W state","mixed states","rank-two states"],"falsifier":"The formula predicts that for $\\lambda = 0.4$ the concurrence fill of $\\rho = 0.4 |\\mathrm{GHZ}\\rangle\\langle\\mathrm{GHZ}| + 0.6 |\\mathrm{W}\\rangle\\langle\\mathrm{W}|$ is exactly $0.8$. A global numerical search over all $2 \\times 4$ co-isometries that finds any ensemble with average concurrence fill below $0.8$ would refute the claim; independently computing the convex roof by a different method at several values of $\\lambda$ would test the entire curve.","tokens_in":10767,"feed_emoji":"🧊","tokens_out":9388,"duration_ms":79459,"temperature":0.7,"pith_summary":"The paper sets out to carry the geometric picture of the concurrence triangle from pure three-qubit states into the mixed-state regime. It associates with every pure-state decomposition of a rank-two three-qubit state a set of concurrence prisms, one prism per ensemble element with the concurrence triangle as base and the probability as height, and shows that the convex-roof concurrence fill equals the minimum over decompositions of the total prism fill area, defined as the product of height and square root of base area. For the canonical family $\\rho(\\lambda) = \\lambda |\\mathrm{GHZ}\\rangle\\langle\\mathrm{GHZ}| + (1-\\lambda) |\\mathrm{W}\\rangle\\langle\\mathrm{W}|$, the minimization becomes a one-variable convex problem and yields the closed form $\\min F = (5\\lambda^2 - 4\\lambda + 8)/9$, together with an explicit optimal ensemble. This gives an exact, checkable value for a benchmark family and a geometric language for convex-roof optimization in three-qubit entanglement.","feed_headline":"Exact entanglement formula for mixed GHZ–W states","feed_subtitle":"A new prism picture reduces tripartite entanglement of mixed states to geometry, with an exact formula for GHZ–W mixtures.","key_machinery":"The machinery is the concurrence prism: a right prism whose base is the concurrence triangle of a pure state, with side lengths equal to the three squared two-qubit concurrences, and whose height is the state's probability in a decomposition; its fill area, height times the square root of the base area, equals that state's contribution $p_k F(|\\varphi_k\\rangle)$ to the convex-roof sum. A $2 \\times 4$ co-isometry parametrizes all pure-state decompositions of a rank-two state relative to its eigendecomposition, reducing the GHZ–W problem to averaging $g(r) = (5\\lambda^2 r^2 - 4\\lambda r + 8)/9$ over $r_k = |v_{1k}|^2/p_k$. Because $g$ is convex and the probabilities weight the $r_k$ so that their average is 1, convexity supplies the lower bound $g(1)$, and the explicit co-isometry achieves it by making every $r_k = 1$. The geometric prerequisite is that every superposition of $|\\mathrm{GHZ}\\rangle$ and $|\\mathrm{W}\\rangle$ has an equilateral concurrence triangle, collapsing the whole calculation to a single convex function of one variable.","core_discovery":"The central claim is that the convex-roof concurrence fill of a rank-two three-qubit mixed state is a geometric minimization: over all pure-state decompositions $\\{p_k, |\\varphi_k\\rangle\\}$ one minimizes $\\sum_k p_k F(|\\varphi_k\\rangle)$, and this quantity is visualized as the total prism fill area of prisms whose bases are the concurrence triangles of the $|\\varphi_k\\rangle$ and whose heights are the probabilities $p_k$. For the mixture $\\rho(\\lambda) = \\lambda |\\mathrm{GHZ}\\rangle\\langle\\mathrm{GHZ}| + (1-\\lambda) |\\mathrm{W}\\rangle\\langle\\mathrm{W}|$, the paper proves this minimum exactly. Every pure state $|\\psi\\rangle = \\sqrt{q}|\\mathrm{GHZ}\\rangle + \\sqrt{1-q}|\\mathrm{W}\\rangle$ has an equilateral concurrence triangle with squared side $(5q^2 - 4q + 8)/9$, hence concurrence fill equal to that same value; substituting the decomposition coefficients through a $2 \\times 4$ co-isometry turns the average fill into an expectation of a convex function, and convexity together with the normalization of the coefficients gives a lower bound that is saturated by the explicit co-isometry with all entries $\\pm 1/2$. The conclusion is $\\min F(\\rho(\\lambda)) = (5\\lambda^2 - 4\\lambda + 8)/9$, verified numerically and realized by four equal-probability, identical-fill pure states.","pith_inferences":["If the equilateral-triangle property holds for other pairs of orthogonal eigenstates, for example other permutation-symmetric pairs, the same convexity argument would immediately produce closed-form concurrence fills for those rank-two families; the paper does not establish this, but the structure invites the test.","The prism picture suggests a concrete numerical strategy for arbitrary rank-two states: minimize total prism fill area over the co-isometry parameters, benchmarking global optimizers against the GHZ–W closed form before applying them elsewhere.","The dip below both pure-state values at intermediate $\\lambda$ indicates that by this measure a GHZ–W mixture is not a simple interpolation of its components; that behaviour could matter for resource estimates in protocols consuming noisy GHZ or W states, though the paper does not discuss applications."],"forward_implications":["For the GHZ–W family, the exact concurrence fill is the parabola $(5\\lambda^2 - 4\\lambda + 8)/9$, so the value at any mixing ratio is known without numerical optimization.","The minimizing ensemble is explicitly constructible: four equal-probability pure states generated by the co-isometry with all entries $\\pm 1/2$, so the convex roof is not only evaluated but realized.","Mixing can suppress genuine tripartite entanglement below both components: the minimum of the parabola occurs at $\\lambda = 0.4$ with value $0.8$, below the pure W value $8/9$ and the GHZ value $1$.","Every rank-two three-qubit mixed state acquires an associated prism picture, giving a geometric interpretation of convex-roof optimization as balancing total prism fill area over decompositions.","The analytic curve provides a benchmark against which numerical convex-roof algorithms for this family can be tested."],"supporting_citations":[{"why":"Defines the concurrence fill as the genuine-tripartite-entanglement measure whose mixed-state extension is the subject of the paper.","marker":"[22]"},{"why":"Supplies the co-isometry parametrization of pure-state decompositions used to reduce the convex roof to an optimization over matrix elements.","marker":"[29]"},{"why":"Provides the inequality for convex functions that yields the lower bound for the average concurrence fill.","marker":"[30]"},{"why":"Establishes GHZ and W as the two inequivalent genuinely entangled classes of three-qubit pure states, the two components of the mixture studied.","marker":"[3]"},{"why":"Gives the bipartite concurrence formula used to compute the side lengths of the concurrence triangle from reduced density matrices.","marker":"[2]"},{"why":"Guarantees that four pure states suffice in any decomposition of a rank-two state.","marker":"[28]"}],"fun_headline_variants":["Exact entanglement formula for mixed GHZ–W states via prisms","Convex-roof prism picture gives exact tripartite entanglement","Closed-form concurrence fill for all GHZ–W rank-2 mixtures","From triangles to prisms: exact entanglement for rank-2 mixed states","Exact convex-roof fill for GHZ–W mixtures"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation rests on the equilateral-triangle property of all superpositions of |GHZ> and |W>, namely that the three sides of the concurrence triangle are always equal with length $(5q^2 - 4q + 8)/9$, and on the general prism construction of Section III remaining a valid representation of every decomposition; if either gave way, the closed-form minimum and its geometric interpretation would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Exact entanglement formula for mixed GHZ–W states via prisms","Convex-roof prism picture gives exact tripartite entanglement","Closed-form concurrence fill for all GHZ–W rank-2 mixtures","From triangles to prisms: exact entanglement for rank-2 mixed states","Exact convex-roof fill for GHZ–W mixtures"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001333,"raw_usage":{"total_tokens":5515,"prompt_tokens":1129,"completion_tokens":4386,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":745,"completion_tokens_details":{"reasoning_tokens":4293}},"tokens_in":745,"tokens_out":4386,"duration_ms":28365,"temperature":1.0,"reasoning_tokens":4293,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T13:58:50.364100+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"The formula predicts that for $\\lambda = 0.4$ the concurrence fill of $\\rho = 0.4 |\\mathrm{GHZ}\\rangle\\langle\\mathrm{GHZ}| + 0.6 |\\mathrm{W}\\rangle\\langle\\mathrm{W}|$ is exactly $0.8$. A global numerical search over all $2 \\times 4$ co-isometries that finds any ensemble with average concurrence fill below $0.8$ would refute the claim; independently computing the convex roof by a different method at several values of $\\lambda$ would test the entire curve.","supporting_citations":[{"cited_title":"Sen(De) and U","cited_arxiv_id":null,"evidence_quote":"Defines the concurrence fill as the genuine-tripartite-entanglement measure whose mixed-state extension is the subject of the paper."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the co-isometry parametrization of pure-state decompositions used to reduce the convex roof to an optimization over matrix elements."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the inequality for convex functions that yields the lower bound for the average concurrence fill."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes GHZ and W as the two inequivalent genuinely entangled classes of three-qubit pure states, the two components of the mixture studied."},{"cited_title":"Aggarwal, International Journal of Theoretical Physics 64, 292 (2025)","cited_arxiv_id":null,"evidence_quote":"Guarantees that four pure states suffice in any decomposition of a rank-two state."}],"review_version":1}