{"id":"e0be2d8f-94e1-4996-b5c6-29c687493f46","arxiv_id":"2509.04581","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Fixed-charge dualities imply symmetric entanglers between non-invertible SPT phases, and an explicit Rep(A4) entangler is constructed as a matrix product unitary.","lead":"A theoretical physics paper shows that certain 'non-invertible' quantum symmetries do support symmetric entanglers, circuits that connect two different symmetry-protected phases. This overturns a prior expectation and supplies the first explicit example, for the symmetry Rep(A4), as a matrix product unitary.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The explicit Rep(A4) entangler is not yet established: unitarity, zero index, and the state-mapping identity are asserted without demonstration, and Eq. (14) holds only up to an N-dependent normalization.","rationale":"I read the paper in good faith. The idea is plausible, the Rep(A4) construction is explicit and checkable, and the symmetry-preservation proof via conjugacy classes is a reasonable strategy. I do not see a demonstrated contradiction in the formulas. However, the central example is not yet verified: unitarity, zero index, and exact state mapping are asserted rather than shown, and the equation E|Psi1> = |Psi2> is at best projective because of the omitted 2^{-N} normalization. These are exactly the premises that would have to be true for the paper's claim of an explicit symmetric entangler to hold. Since the general theorem in Sec. II is also used to motivate the expectation but the MPU is admitted to be ad hoc, the explicit construction is the real load-bearing evidence. The reader's conditional verdict captures this accurately: the construction is plausible, but the omitted checks mean the central claim is not yet established. I therefore keep the verdict unchanged rather than accepting outright or rejecting. My main emphasis on the MPU checks rather than the symTFT-to-lattice dictionary is why my agreement with the reader's weakest-assumption diagnosis is only partial.","tokens_in":12422,"tokens_out":23240,"duration_ms":243003,"concrete_test":"Write an exact numerical implementation of the tensor T from Eqs. (12)-(13) for a small closed chain, e.g., N=3 or N=4, and check: (a) E^dagger E = I by applying E to every basis state; (b) the matrix elements of E|Psi1> are proportional, with a single N-dependent constant, to those of |Psi2> from Eq. (11); (c) [E, L_pi] = 0 and [E, L_omega] = [E, L_omega^2] = 0 by explicit multiplication with the MPO symmetry operators of Eq. (8); and (d) the MPU index defined in Definition IV.1 of Ref. [35] evaluates to zero. Any failure of (a), (c), or (d) disproves the claim that E is a symmetric entangler.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central existence claim rests on the explicit MPU E constructed in Sec. III C 3. Three properties are load-bearing: (i) E is unitary as a many-body operator for every N; (ii) E has zero MPU index, hence is an FDQC; (iii) E commutes with the Rep(A4) symmetry operators and maps the two SPT states to each other. None of these is actually demonstrated. Unitarity is dismissed by citing Theorem 1 of Ref. [34], but the local isometry conditions are not written down or checked against the sign matrices in Eq. (13). The zero-index statement is a single sentence with no computation. The state-mapping equation Eq. (14) is also not exact: because each tensor carries a factor 1/2, acting on |Psi1> gives E|Psi1> = 2^{-N} |Psi2> up to the sign factors, not |Psi2> as written; this is harmless projectively but shows the mapping is only verified up to an unstated normalization. Moreover, the sign matrices (s_e, s_x, s_{x^2}) require a fixed row/column ordering, and the parity argument that makes the signs cancel when acting on |Psi1> is not given. Since this MPU is the paper's only concrete evidence that a symmetric entangler for non-invertible SPT phases exists, a failure in any of these checks would falsify the central example. The general FCD-to-boundary argument also depends on an unproved dictionary, but the explicit construction is independent of that dictionary and therefore stands or falls on the MPU checks.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper argues that non-invertible SPT phases in 1+1d can be connected by symmetric entanglers despite the absence of a stacking structure. Section II proves, in the symTFT setup, that a fixed-charge duality D of the Drinfeld center Z(C) satisfies F(D(μ)) = F(μ) for the forgetful functor F, so D preserves all boundary symmetry operators. Section III discusses examples: Rep(D8) has no FCD and, consistently with Ref. [7], no symmetric entangler; G×Rep(G) cluster states are not connected by any FSD; and the two Rep(A4) SPT phases, which are connected by an FCD, are claimed to be connected by an explicit matrix product unitary E. The MPU E is presented with sign tables, and the paper asserts that E is unitary of order two, has zero MPU index, commutes with the Rep(A4) symmetry, and maps the product state |Ψ1> to the second SPT state |Ψ2>. Appendix A gives an L-symbol computation showing that the two states lie in distinct phases.","tokens_in":12752,"tokens_out":10292,"duration_ms":95877,"significance":"If the missing checks are supplied, the Rep(A4) construction would be the first explicit symmetric entangler for non-invertible SPT phases, directly contradicting the earlier no-go expectation of Ref. [7] and showing that the absence of stacking does not preclude entanglers. The paper's strengths are its clean formal statement in Sec. II, the explicit and parameter-free tensors in Eqs. (10)-(13), and the concrete L-symbol and torus partition function invariant in Appendix A that distinguishes the two phases. No fitted parameters enter; the only adjustable data are the sign tables, which are fixed. The main reservation is that the central example is not yet fully established, because unitarity, the zero-index property, and exact commutation are asserted rather than demonstrated.","major_comments":[{"comment":"Eq. (14) is not exact as written. Each tensor in Eq. (12) carries a factor 1/2, so acting on |Ψ1> = |e,e,...> gives E|Ψ1> = 2^{-N} Σ_{g_i} Tr[Q(g_1)...Q(g_N)] |g_1...g_N>, up to the signs s_e(g_i,e), not |Ψ2> as displayed. Since the MPS |Ψ2> in Eq. (11) has norm 2^N, the normalized state is consistent with unitarity of E, but the equality should be written with an explicit normalization and the sign factors should be checked to be constant on the relevant configurations.","section":"III C 3, Eq. (14)"},{"comment":"Unitarity of E for every N is load-bearing and is not demonstrated. The sentence that E satisfies the conditions of Theorem 1 of Ref. [34] does not state those conditions or verify them against the sign matrices Eq. (13). Because E is the only concrete evidence for the paper's central claim, the local isometry conditions, or an explicit calculation of E†E = 1, should be written out.","section":"III C 3, after Eq. (13)"},{"comment":"The claim that the MPU index is zero is a single sentence with no computation. The finite-depth property of the entangler rests entirely on this index; please provide the index calculation in the convention of Ref. [35], or give an explicit finite-depth circuit decomposition of E.","section":"III C 3, after Eq. (13)"},{"comment":"The commutation proof drops the sign factors from Eq. (12). After acting with E on a general basis state, the amplitudes are written 'up to some signs from Eq. 13', but those signs depend on the individual group elements g_i and h_i. To conclude that E commutes with L_π and the other symmetry operators, one must show that the signs do not distinguish the input charge class or otherwise cancel in the matrix elements of E L_π and L_π E. As written, the charge-preservation argument only uses the block structure and trace properties, not the signs.","section":"III C 3, Eqs. (15)-(16)"},{"comment":"The general claim that an FCD implies a boundary symmetric entangler is not a theorem in the present manuscript. Section II proves only F(D(μ)) = F(μ), i.e. preservation of boundary symmetry lines; the step from a bulk duality to a lattice unitary is cited to Refs. [21,25,26] and described as an expectation. The discussion in Sec. IV states that the explicit MPU was constructed ad hoc and does not directly use the bulk FCD, so the general statement is not supported by the explicit construction either. Please either provide a precise dictionary theorem or label this part as a conjecture, separate from the explicit Rep(A4) result.","section":"II, last paragraph; IV"}],"minor_comments":[{"comment":"The theorem statement writes F(D(µ)) = D(µ), which is not well typed since D(µ) is a bulk anyon while F outputs a boundary line; the intended statement, used in the proof, is F(D(µ)) = F(µ).","section":"II, Theorem statement"},{"comment":"The symbol (g,h) in T_{x^2g,x^2h} is never defined; presumably it is the 2-cocycle phase satisfying Q(g)Q(h) = (g,h)Q(gh), but this should be stated explicitly.","section":"III C 3, Eq. (12)"},{"comment":"The row and column ordering of the sign matrices is not specified; the reader cannot reproduce the MPU without knowing, for example, whether the ordering is (e,a,b,c).","section":"III C 3, Eq. (13)"},{"comment":"The relation T_{j,i}^* = U T_{i,j} U^{-1} with U = iY is stated without showing that the sign tables satisfy it; a short verification would make the order-two claim checkable.","section":"III C 3"}],"recommendation":"major_revision","confidential_remarks":"No additional concerns beyond the technical gaps listed above. The paper is within the scope of the journal, and the explicit Rep(A4) example is worth publishing once the unitarity, index, and commutation checks are supplied."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper claims two things: a general argument that fixed-charge dualities of the symTFT give symmetric entanglers for non-invertible SPT phases, and the first explicit such entangler, an MPU connecting the two Rep(A4) SPT phases. The second claim, if right, overturns the expectation from Seifnashri–Shao for Rep(D8). The author is honest that the MPU was built ad hoc and does not follow from the general theorem.\n\nWhat is genuinely good: the L-symbol computation in the appendix gives a concrete invariant distinguishing the two states, and the MPU construction itself is explicit enough to be checked. The general theorem is short and uses the standard symTFT machinery; the idea that FCDs preserve charges and hence boundary symmetries is natural.\n\nThe soft spots are real. The theorem statement contains a type error: 'F(D(µ)) = D(µ)' should be 'F(D(µ)) = F(µ)'. More substantively, the proof uses β_a(D(µ)) = β_a(µ) as though it follows from charge preservation, but β_a(µ) is a trace over the junction space W^µ_1, which is not determined by the charge alone. That step needs an argument or a citation; without it the theorem is not proved. The explicit MPU is where I would focus referee attention. The state-mapping identity E|Ψ1> = |Ψ2> is stated 'up to a possible overall sign,' but because each tensor carries a factor of 1/2, the equality is actually only up to a 2^{-N} normalization as well, which is harmless if the paper says so. The sign matrices in Eq. (13) appear to depend on g and h, so the claim that the signs collapse to an overall constant when acting on |Ψ1> requires a parity argument that is not given. Unitarity is dismissed with a citation to Theorem 1 of Ref. [34], but the local isometry conditions against the sign matrices are never checked. The zero-index claim is a single sentence with no computation. If any of these fail, the central example fails; they are all probably fixable, but they are currently assertions, not demonstrations.\n\nDespite this, the paper deserves a serious referee. The claim is important, the construction is sufficiently concrete that a referee can verify it, and the author has flagged the main caveat about the symTFT-to-lattice dictionary. I would send it to review with a request to supply the missing checks, not desk-reject it.","headline":"A plausible and important claim that non-invertible SPTs can have symmetric entanglers, with a concrete Rep(A4) example, but several load-bearing checks on the explicit MPU are asserted rather than demonstrated.","tokens_in":13299,"tokens_out":4330,"would_cite":false,"duration_ms":39022,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Fixed-charge dualities of the bulk symmetry TFT preserve boundary symmetries, so non-invertible SPT phases connected by them have symmetric entanglers; the paper builds one explicitly for the two Rep(A4) phases.","keywords":["non-invertible symmetry","SPT phases","symmetric entangler","fixed-charge duality","topological holography","symTFT","matrix product unitary","Rep(A4)"],"falsifier":"Evaluate the MPU index of the explicit tensor E on a finite chain using the index formula of the cited MPU structure theorem, and directly check [E, Lπ] = 0; a nonzero index or nonzero commutator would falsify the claim that E is a symmetric finite-depth circuit, and a fusion category with FCD-connected SPT phases but no boundary entangler would falsify the general theorem.","tokens_in":12214,"feed_emoji":"🔗","tokens_out":10035,"duration_ms":83625,"temperature":0.7,"pith_summary":"The paper takes on a widely repeated expectation: that non-invertible symmetry-protected phases cannot be connected by symmetric entanglers because they have no stacking structure. Its central claim is the opposite: whenever two 1+1d SPT phases of a non-invertible fusion-category symmetry are connected by a fixed-charge duality of the bulk symmetry TFT, a symmetric entangler must exist. The paper proves the key step, that a fixed-charge duality preserves the boundary symmetry operators, and then gives a concrete example: a matrix product unitary that maps the product state to the nontrivial Rep(A4) SPT state, commutes with the symmetry, and has zero index. If the claim holds, symmetric entanglers exist for non-invertible symmetries even though stacking does not, and the two notions are decoupled.","feed_headline":"Symmetric entanglers exist for non-invertible SPT phases","feed_subtitle":"Fixed-charge dualities guarantee them, and an explicit circuit is built for Rep(A4).","key_machinery":"The load-bearing object is the symTFT dictionary: a 1+1d system with fusion-category symmetry $C$ is encoded by the 2+1d TQFT $Z(C)$, with a reference Dirichlet boundary fixing the charges and a physical boundary supplying the dynamics. Within that dictionary, the named mechanism is the forgetful functor $F: Z(C) \\to C$, which turns bulk anyons into boundary symmetry lines, together with fixed-charge dualities (braided autoequivalences of $Z(C)$ that preserve the charges). The theorem's engine is the invariance of torus partition functions under an FCD, which forces the junction multiplicities $\\langle D(\\mu),a\\rangle$ to equal $\\langle\\mu,a\\rangle$ and hence $F(D(\\mu)) = F(\\mu)$. On the lattice, the explicit entangler is a matrix product unitary built from the unique degree-2 projective representation of A4; its zero index, verified by a cited criterion, turns it into a finite-depth quantum circuit.","core_discovery":"The core discovery is a theorem about the symmetry TFT (symTFT) of a fusion category $C$. In the Drinfeld center $Z(C)$, bulk anyons $\\mu$ flow to boundary symmetry operators through a forgetful functor $F$, and charges are the coefficients with which anyons appear in the Dirichlet Lagrangian algebra. The theorem states that if $D$ is a braided autoequivalence of $Z(C)$ that preserves every charge (a fixed-charge duality, FCD), then $F(D(\\mu)) = F(\\mu)$ for every anyon $\\mu$; so $D$ preserves the boundary symmetries. The proof runs through torus partition functions with symmetry lines inserted: an FCD leaves the twisted partition functions invariant, which forces the junction dimensions $\\langle\\mu,a\\rangle$ to be invariant, hence $F(D(\\mu))=F(\\mu)$. The paper therefore concludes that an FCD connecting two SPT phases yields a symmetric entangler, and exhibits one for the two Rep(A4) SPT phases as an explicit matrix product unitary.","pith_inferences":["A criterion the argument suggests but does not state: for a fixed fusion category, the SPT phases connected by fixed-charge dualities should be exactly the orbits of the symmetric-entangler action, so enumerating FCDs would classify which pairs admit entanglers.","The bulk-to-boundary dictionary is cited rather than derived; making it algorithmic would turn the theorem into a construction method and test it on categories beyond Rep(A4).","The sign flip in the torus partition function $(Z_{\\pi\\pi,\\pi})_{\\mu\\nu}$ computed in the appendix offers a concrete, gauge-invariant diagnostic that distinguishes the two Rep(A4) phases and could serve as a fingerprint for non-invertible SPT phases in other models."],"forward_implications":["The absence of symmetric entanglers is not a general property of non-invertible SPT phases; it depends on whether the phases are connected by a fixed-charge duality.","The two Rep(A4) SPT phases form a Z2 torsor under the entangler E, since E is order two and maps one phase to the other.","For Rep(D8) and for non-abelian G×Rep(G) cluster states there is no fixed-charge duality connecting the phases, so the earlier no-entangler results remain consistent.","Existence of symmetric entanglers is decoupled from the existence of a stacking structure, since Rep(A4) has no diagonal subalgebra yet admits an entangler.","The explicit MPU provides a finite-depth, globally symmetric circuit on chains of arbitrary length N that maps the product state to the nontrivial Rep(A4) SPT state."],"supporting_citations":[{"why":"Classifies dualities into FCD, FAD, and FSD, and identifies the Rep(A4) SPT phases as FCD-connected.","marker":"[27]"},{"why":"Introduced the Rep(D8) cluster-state SPT and the no-symmetric-entangler expectation this paper overturns.","marker":"[7]"},{"why":"Supplies the G×Rep(G) cluster states and the non-commuting unitary UC, the non-abelian counterpoint to an entangler.","marker":"[6]"},{"why":"Provides the on-site MPS/MPO construction for fusion-category-symmetric states adapted to build the Rep(A4) states.","marker":"[8]"},{"why":"Establishes the topological-holography dictionary between bulk symTFT symmetries and 1+1d dualities.","marker":"[21]"},{"why":"Shows how bulk dualities become MPO intertwiners in 1D lattice models, the cited step from bulk FCD to boundary entangler.","marker":"[25]"},{"why":"Companion derivation of dualities in 1D lattice models from topological sectors, also backing the boundary-descent step.","marker":"[26]"},{"why":"Provides the MPU/QCA structure theorem and the index-zero criterion that turns the MPU into an FDQC.","marker":"[35]"},{"why":"Supplies the unitarity condition (Theorem 1) used to verify that the explicit MPO is a unitary MPU.","marker":"[34]"},{"why":"Gives the L-symbol/MPO-symmetry formalism used to confirm the two Rep(A4) states are in inequivalent phases.","marker":"[33]"}],"fun_headline_variants":["Symmetric entanglers now exist for non-invertible SPTs","Fixed-charge dualities guarantee symmetric entanglers","Explicit symmetric entangler for Rep(A4) SPT phases","Non-invertible SPTs get symmetric entanglers via FCD","Symmetric entanglers proven for non-invertible SPTs"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a fixed-charge duality of the bulk symTFT descends to a symmetric entangler on the lattice; the paper cites this dictionary from earlier work rather than proving it, and the explicit Rep(A4) MPU was built ad hoc without using that FCD.","fun_headline_variants_meta":{"raw":{"variants":["Symmetric entanglers now exist for non-invertible SPTs","Fixed-charge dualities guarantee symmetric entanglers","Explicit symmetric entangler for Rep(A4) SPT phases","Non-invertible SPTs get symmetric entanglers via FCD","Symmetric entanglers proven for non-invertible SPTs"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000918,"raw_usage":{"total_tokens":3902,"prompt_tokens":869,"completion_tokens":3033,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":485,"completion_tokens_details":{"reasoning_tokens":2939}},"tokens_in":485,"tokens_out":3033,"duration_ms":21177,"temperature":1.0,"reasoning_tokens":2939,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T16:29:33.507429+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the MPU index of the explicit tensor E on a finite chain using the index formula of the cited MPU structure theorem, and directly check [E, Lπ] = 0; a nonzero index or nonzero commutator would falsify the claim that E is a symmetric finite-depth circuit, and a fusion category with FCD-connected SPT phases but no boundary entangler would falsify the general theorem.","supporting_citations":[{"cited_title":"Zhang and C","cited_arxiv_id":null,"evidence_quote":"Introduced the Rep(D8) cluster-state SPT and the no-symmetric-entangler expectation this paper overturns."},{"cited_title":"Moradi, S","cited_arxiv_id":null,"evidence_quote":"Companion derivation of dualities in 1D lattice models from topological sectors, also backing the boundary-descent step."}],"review_version":2}