{"id":"b79b4c5d-852a-4114-b05a-59f64a5d6cab","arxiv_id":"2506.14480","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Central maps that are Lorentz-entanglement breaking or annihilating are exactly those whose underlying operator has 2-dominated norm gamma_2* <= lambda or 2-summing norm pi_2 <= lambda; an explicit non-entanglement-breaking Lorentz-entanglement-breaking map on M3(C)+ is constructed.","lead":"This mathematics paper defines and analyzes maps that break or annihilate entanglement when the reference system is a Lorentz cone, and connects these maps to classical operator ideal norms in Banach space theory. It gives exact characterizations for central maps in terms of the norms gamma_2*, gamma_2, and pi_2, and constructs an explicit map on 3x3 matrices that breaks Lorentz entanglement but not ordinary entanglement.","discovery_kind":"unification","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.6's general-case step rests on an unproved gamma_2* factorization through ell_infinity^n and ell_1^m; it is likely derivable from (4)/(8), but as written the general case of Theorem 1.4(2) is not established.","rationale":"The reader's strongest_claim is exactly Theorem 1.4, and the reader's weakest_assumption identifies the same point I would flag: the unproved factorization in Theorem 4.6 that reduces general X, Y to ell_infinity^n and ell_1^m. I attempted to test whether the assertion is actually false; it appears to be true and derivable from standard 2-nuclear factorization, equation (8), and the dual formula (4), because a diagonal contraction from ell_infinity^k to ell_2^k has 2-summing norm at most 1. Thus the concern is a missing proof rather than a counterexample. Still, as written, the general case of Theorem 1.4(2) depends on an unstated lemma, so a conditional verdict is appropriate until the proof is supplied. The other central claims -- the gamma_2 side in Corollary 4.7, the pi_2 characterization in Theorem 3.9, and the explicit non-entanglement-breaking Lorentz-entanglement-breaking example in Section 5.2 -- do not appear threatened by this gap. I would retain the CONDITIONAL verdict and ask the authors to add a proof or citation for the factorization step.","tokens_in":27528,"tokens_out":23400,"duration_ms":230754,"concrete_test":"Analytical check: insert into Theorem 4.6 an explicit lemma proving the asserted factorization. For v with gamma_2*(v) <= 1, use (4) to choose v = v2 v1 with pi_2(v1) <= 1 and pi_2(v2*) <= 1, then use (8) to factor v1 = B Delta A and v2* = C Delta_2 D with contractions A: X -> ell_infinity^n and D*: ell_1^m -> Y and diagonal contractions Delta, Delta_2; set v' = Delta_2* C* B Delta and verify gamma_2*(v') <= pi_2(B Delta) pi_2((Delta_2* C*)*) <= 1. If this chain cannot be completed, the reduction is invalid; if it succeeds, add the lemma with a reference to [TJ89, Sections 9 and 13].","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing gap is in the proof of Theorem 4.6, which establishes the central characterization in Theorem 1.4(2). After proving the case X = ell_infinity^n and Y = ell_1^m, the proof reduces the general finite-dimensional case by asserting: gamma_2*(v) <= 1 if and only if v = beta v' alpha for contractions alpha : X -> ell_infinity^n and beta : ell_1^m -> Y with gamma_2*(v') <= 1. This assertion is stated without proof or citation. It is exactly what transfers the ell_infinity/ell_1 case to arbitrary X and Y; if it failed, the 'if' direction of Theorem 1.4(2) for general spaces would not be established. The assertion is likely true and can be derived from (4) and the 2-nuclear factorization (8): write v = v2 v1 with pi_2(v1) <= 1 and pi_2(v2*) <= 1, factor v1 = B Delta A and v2* = C Delta_2 D with contractions A, B, C, D and diagonal contractions Delta, Delta_2, then set alpha = A, beta = D*, and v' = Delta_2* C* B Delta; the bound gamma_2*(v') <= pi_2(B Delta) pi_2((Delta_2* C*)*) <= 1 uses contraction properties and pi_2(Delta), pi_2(Delta_2) <= 1 for diagonal contractions. But this argument is absent from the manuscript, so the general case of Theorem 4.6 is a genuine proof gap that should be repaired before the theorem is accepted as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies linear maps between proper cones that annihilate or break entanglement when the reference cone is a Lorentz cone. For central maps λ⊕u between cones C_X and C_Y over finite-dimensional normed spaces, it claims exact characterizations: λ⊕u is Lorentz-factorizable iff γ_2(u)≤λ (Theorem 1.4(1)), and λ⊕u is Lorentz-entanglement breaking iff γ_2^*(u)≤λ (Theorem 1.4(2)); central Lorentz-entanglement annihilating maps into Lorentz cones are characterized by π_2(u)≤λ (Theorems 1.2 and 3.9). It also proves a no-closure duality theorem for LorEB and LorFact (Theorem 4.3), establishes structural properties of maxEA_2 between Lorentz cones, and constructs an explicit Lorentz-entanglement breaking map between 3×3 matrix cones that is not entanglement breaking, using the VB14 counterexample to Peres' conjecture.","tokens_in":27933,"tokens_out":10827,"duration_ms":110094,"significance":"If the proofs are completed, this is a substantial bridge between tensor products of ordered vector spaces and Banach-space operator ideal norms. The paper gives parameter-free norm criteria with no fitted constants, supplies a clean no-closure duality statement, and provides explicit examples separating LorEB from EB. The main theorems are likely correct, but the proof of Theorem 1.4(2) currently contains a load-bearing unproved factorization, and Lemma 3.1 has undefined quantities in cases that occur at boundary contractions.","major_comments":[{"comment":"In the final paragraph of the proof of Theorem 4.6, the reduction from general finite-dimensional X,Y to X=ℓ∞^n and Y=ℓ_1^m relies on the assertion that γ_2^*(v)≤1 if and only if v=βv'α with α:X→ℓ∞^n and β:ℓ_1^m→Y contractions and γ_2^*(v')≤1. This factorization is exactly what transfers the ℓ∞/ℓ_1 case to arbitrary spaces, and it is therefore load-bearing for Theorem 1.4(2), but it is stated without proof or citation. The statement is plausible and likely follows from (4) and the 2-nuclear factorization (8), but as written the general-case 'if' direction of Theorem 4.6 is not established. Please insert a proof or a precise reference for this factorization.","section":"§4.2, proof of Theorem 4.6"},{"comment":"In the proof of Lemma 3.1, the vector z_2 is defined using divisions by ∥x_1∥_2, ∥y_1∥_2, and hs(V_1), which may vanish. Lemma 3.1 is used in Theorem 3.2 to characterize central max-entanglement-annihilating maps between Lorentz cones, so the proof needs a complete treatment of these zero cases, for example by a perturbation/continuity argument or by splitting into cases before defining z_2. As it stands, the proof is incomplete exactly in boundary cases that are needed for the full statement.","section":"§3.1, proof of Lemma 3.1"}],"minor_comments":[{"comment":"The statement quantifies r,s∈R_+^0 but the conclusion uses s and t; r does not appear. The intended quantification is presumably s,t∈R_+^0.","section":"§3.1, Lemma 3.1 statement"},{"comment":"The text contains the typo 'Lorenz-entanglement annihilating'; it should be 'Lorentz-entanglement annihilating'.","section":"§1.2"},{"comment":"Proposition 2.2 is applied to the set maxEA_2(L_n,L_m) before the closedness of this set is explicitly stated; a sentence noting that closedness follows from continuity of the defining inclusion would remove ambiguity.","section":"§3.1, proof of Theorem 3.3"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi [Name],\n\nRead the La Piana–Müller-Hermes paper. Short version: it is a good paper and I would send it to a referee. The main results are the characterizations of central maps: λ⊕u factors through Lorentz cones iff γ2(u)≤λ, breaks Lorentz entanglement iff γ2*(u)≤λ, and annihilates Lorentzian tensor entanglement into a Lorentz cone iff π2(u)≤λ. Those are clean, non-obvious, and connect ordered vector spaces to classical Banach-space operator ideals. The Lorentzian tensor product and its non-associativity are nice. The explicit example in Section 5.2—a Lorentz-entanglement-breaking map on M3(C)+ that is not entanglement breaking, built from the VB14 counterexample to Peres' conjecture—is quite an achievement.\n\nThe reader's conditional verdict is fair. The main gap is real: in Theorem 4.6, after proving the ℓ∞^n → ℓ1^m case, the reduction to general X,Y relies on a factorization γ2*(v)≤1 iff v = β v' α with contractions α:X→ℓ∞^n, β:ℓ1^m→Y and γ2*(v')≤1. This is stated without proof or citation. It is plausible from the 2-nuclear factorization (8) and the dual factorization (4), and the stress-test note gives a reasonable derivation path. But as written, the 'if' direction of Theorem 1.4(2) for general spaces is not fully established. That needs to be filled in—not fatal, but a genuine gap.\n\nMinor issues: Lemma 3.1 defines z2 with divisions by ∥x1∥, ∥y1∥, hs(V1), which can be zero; an easy continuity argument or separate case handles it. There are also cross-reference slips (e.g., in the proof of Theorem 3.4 it cites 'Theorem 3.4' for the mixed-product property that is actually Theorem 3.3). None of this undermines the overall structure.\n\nThe citation pattern looks fine. The self-citations to AMH23/AMH24 are for context and strict inclusions, not load-bearing for the central theorems. No data fitting, no invented entities.\n\nBottom line: a referee should ask for the gap in Theorem 4.6 to be repaired and the zero-division issue cleaned up, but the core mathematics is sound and novel. I'd take it seriously. Yes to peer review.","headline":"A strong paper that gives exact norm characterizations for Lorentz-entanglement breaking and annihilating central maps; the main gap in Theorem 4.6 is likely repairable and should not block refereeing.","tokens_in":28430,"tokens_out":2100,"would_cite":true,"duration_ms":18926,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46B28","47L20","81P40"],"pacs":[],"model":"deepseek-v4-flash","headline":"For central maps between cones over normed spaces, Lorentz-entanglement breaking is equivalent to $\\gamma_2^*(u)\\le\\lambda$; factoring through a Lorentz cone is equivalent to $\\gamma_2(u)\\le\\lambda$.","keywords":["Lorentz cone","entanglement breaking","entanglement annihilating","operator ideal norms","2-summing norm","Hilbert-space factorization norm","2-dominated norm","cones over normed spaces"],"falsifier":"Test the unproved factorization step numerically: take $X=\\ell_p^n$ and $Y=\\ell_q^m$ with $p,q$ not of the $\\infty/1$ type, compute $\\gamma_2^*(\\mathrm{id})$ by its definition, and check whether a factorization through $\\ell_\\infty^N\\to\\ell_1^M$ exists with contractions and unchanged norm. Alternatively, for any candidate $u$ with $\\gamma_2^*(u)\\le1$, compute $\\sup|\\operatorname{Tr}[wvu]|$ over contractions $v:\\ell_2^k\\to X$ and $w:Y\\to\\ell_2^k$; if this supremum exceeds 1, then $1\\oplus u$ cannot be Lorentz-entanglement breaking, contradicting Theorem 1.4.","tokens_in":27298,"feed_emoji":"🍦","tokens_out":12695,"duration_ms":124328,"temperature":0.7,"pith_summary":"Lorentz cones are the \"ice-cream cones\" $\\{(t,x):t\\ge\\|x\\|_2\\}$ used as reference systems in entanglement theory. The paper studies positive maps between cones $C_X=\\{(t,x):t\\ge\\|x\\|\\}$ and $C_Y$, focusing on central maps $\\lambda\\oplus u$ that act as $\\lambda$ on the $t$-coordinate and $u$ on the space coordinate. It establishes that for such maps two Lorentz-related properties are exact norm thresholds: breaking Lorentz-cone entanglement is equivalent to $\\gamma_2^*(u)\\le\\lambda$, and factoring through a Lorentz cone is equivalent to $\\gamma_2(u)\\le\\lambda$. It also shows that when the target cone is a Lorentz cone, annihilating Lorentz-cone entanglement is governed by the 2-summing norm condition $\\pi_2(u)\\le\\lambda$. The reader should care because these equivalences turn qualitative cone-theoretic properties into computable Banach-space quantities and connect quantum-inspired entanglement to a classical operator-ideal theory.","feed_headline":"One norm threshold decides Lorentz entanglement breaking","feed_subtitle":"For central maps, Lorentz-entanglement breaking is exactly the 2-dominated norm bound $\\gamma_2^*(u)\\le\\lambda$; factoring is…","key_machinery":"The load-bearing device is the central-map decomposition $P=\\lambda\\oplus u$, namely $P(t,x)=(\\lambda t,u(x))$ on $\\mathbb{R}\\times X\\to\\mathbb{R}\\times Y$. Positivity of this map is exactly the operator-norm condition $\\|u\\|\\le\\lambda$; the paper shows that replacing the operator norm by $\\gamma_2$, $\\gamma_2^*$, and $\\pi_2$ respectively detects factorization through Lorentz cones, Lorentz-entanglement breaking, and Lorentz-entanglement annihilation. The proofs use the Sinkhorn-type normal form for positive maps between Lorentz cones, which reduces arguments to diagonal contractions, together with the retract property of Lorentz cones cut by subspaces and the standard operator-ideal machinery of 2-nuclear and Hilbert-space factorizations. Theorem 4.3, showing that the trace-dual of $\\mathrm{LorEB}$ is exactly the Lorentz-factorizable cone, is the bridge that makes the norm characterizations dual.","core_discovery":"The central discovery is a dictionary between two worlds. On one side are two families of positive maps defined by Lorentz cones: Lorentz-entanglement-breaking maps $\\mathrm{LorEB}(C_X,C_Y)$, which send the maximal tensor product with any Lorentz reference cone into the minimal tensor product, and their trace-dual Lorentz-factorizable maps $\\mathrm{LorFact}(C_X,C_Y)$. On the other side are two classical operator ideal norms on the underlying linear map $u:X\\to Y$: the Hilbert-space factorization norm $\\gamma_2(u)$ and its trace-dual, the 2-dominated norm $\\gamma_2^*(u)$. The paper proves that $\\lambda\\oplus u\\in\\mathrm{LorFact}(C_X,C_Y)$ if and only if $\\gamma_2(u)\\le\\lambda$, and $\\lambda\\oplus u\\in\\mathrm{LorEB}(C_X,C_Y)$ if and only if $\\gamma_2^*(u)\\le\\lambda$. It also proves that $\\lambda\\oplus u$ annihilates Lorentz-cone entanglement into a Lorentz cone if and only if $\\pi_2(u)\\le\\lambda$. A further theorem gives the dual-cone identity $\\mathrm{LorEB}(C_B,C_A)^*=\\mathrm{LorFact}(C_A,C_B)$ without requiring a closure, and the paper constructs a Lorentz-entanglement-breaking map on $3\\times3$ positive semidefinite matrices that is not entanglement breaking, using a known bound-entanglement state.","pith_inferences":["Inference: because $\\gamma_2$ and $\\gamma_2^*$ have Hilbert-space factorization descriptions, Lorentz-entanglement-breaking and Lorentz-factorizable central maps may be certifiable by semidefinite programming; the paper itself does not discuss algorithms.","Inference: if the unproved factorization step in Theorem 4.6 fails for some pair of normed spaces, the 'if' direction of the main equivalence would still hold for maps that admit the compression, but not for arbitrary $X,Y$; a counterexample would narrow the theorem without affecting the $\\ell_\\infty/\\ell_1$ case or the constructed examples.","Inference: the square-base cone $C_{\\ell_\\infty^2}$ behaves like a cone-theoretic analogue of a space with the 2-summing property; testing other polyhedral cones, such as $C_{\\ell_\\infty^d}$ for $d>2$, would show whether this property is special to the square or extends to higher-dimensional cubes."],"forward_implications":["Central Lorentz-entanglement breaking becomes a closed, convex, norm-bounded property: on central maps the cone is cut out by $\\gamma_2^*(u)\\le\\lambda$, so membership is a convex condition rather than a search over all reference cones.","Because $\\gamma_2^*$ is the trace-dual of $\\gamma_2$, Lorentz-entanglement breaking and Lorentz factorizability are dual cones, mirroring the operator-ideal duality in Banach space theory.","When the target cone is a Lorentz cone, $\\mathrm{LorEA}_2(C_X,\\mathbb{L}^n)$ is a closed convex cone whose central slice is the unit ball of $\\pi_2$; in the square-base case the whole positive cone $\\mathrm{Pos}(C_{\\ell_\\infty^2},\\mathbb{L}^m)$ annihilates Lorentz entanglement.","There exist strict inclusion examples: on $M_3(\\mathbb{C})_+$ there is a Lorentz-entanglement-breaking map that is not entanglement breaking, and the identity map on $M_3(\\mathbb{C})_+$ does not factor through a Lorentz cone.","The known implication from $\\infty$-max-entanglement annihilation to Lorentz-entanglement breaking now becomes, for central maps, exactly the condition $\\gamma_2^*(u)\\le\\lambda$.",""],"supporting_citations":[{"why":"Supplies the operator ideal norms $\\gamma_2$, $\\gamma_2^*$, and $\\pi_2$, their factorization formulas, and trace dualities used in Theorems 1.2-1.4.","marker":"[TJ89]"},{"why":"Establishes the dual-cone formula for $\\mathrm{LorEB}$ and the fact that $\\infty$-max-entanglement-annihilating maps are Lorentz-entanglement breaking, framing Theorem 1.4.","marker":"[AMH23]"},{"why":"Provides the Sinkhorn-type normal form for positive maps between Lorentz cones used to reduce proofs to diagonal contractions.","marker":"[Hil11]"},{"why":"Introduced the central-map template and the norm characterizations of positivity and entanglement breaking that the paper generalizes.","marker":"[Lam18]"},{"why":"Shows that tensor-power norm bounds imply $\\gamma_2^*\\le1$, connecting asymptotic entanglement annihilation to the Lorentz-entanglement-breaking condition.","marker":"[AMH24]"},{"why":"Shows the $3\\times3$ positive semidefinite cone cannot be represented through second-order cones, supporting the strictness claims.","marker":"[Faw19]"},{"why":"Supplies the bound-entanglement state and measurement vectors used to construct an explicit Lorentz-entanglement-breaking map that is not entanglement breaking.","marker":"[VB14]"},{"why":"Gives the criterion that completely positive and completely copositive maps break entanglement with $M_2$, used in the explicit example.","marker":"[CMHW19]"}],"fun_headline_variants":["Lorentz entanglement breaking: one norm to check","Operator ideals meet Lorentz cone maps","Dual identity links Lorentz factorizing and breaking","Lorentz-breaking map not entanglement breaking found"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes, without showing, that every linear map whose 2-dominated size is at most 1 can be re-expressed as a map between an $\\ell_\\infty$-space and an $\\ell_1$-space without increasing that size; the main equivalence for arbitrary normed spaces rests on this unstated compression step.","fun_headline_variants_meta":{"raw":{"variants":["Lorentz entanglement breaking: one norm to check","Operator ideals meet Lorentz cone maps","Dual identity links Lorentz factorizing and breaking","Lorentz-breaking map not entanglement breaking found"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000548,"raw_usage":{"total_tokens":2740,"prompt_tokens":1189,"completion_tokens":1551,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":805,"completion_tokens_details":{"reasoning_tokens":1494}},"tokens_in":805,"tokens_out":1551,"duration_ms":14763,"temperature":1.0,"reasoning_tokens":1494,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:52:30.653117+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Test the unproved factorization step numerically: take $X=\\ell_p^n$ and $Y=\\ell_q^m$ with $p,q$ not of the $\\infty/1$ type, compute $\\gamma_2^*(\\mathrm{id})$ by its definition, and check whether a factorization through $\\ell_\\infty^N\\to\\ell_1^M$ exists with contractions and unchanged norm. Alternatively, for any candidate $u$ with $\\gamma_2^*(u)\\le1$, compute $\\sup|\\operatorname{Tr}[wvu]|$ over contractions $v:\\ell_2^k\\to X$ and $w:Y\\to\\ell_2^k$; if this supremum exceeds 1, then $1\\oplus u$ cannot be Lorentz-entanglement breaking, contradicting Theorem 1.4.","supporting_citations":[],"review_version":2}