{"id":"ffc7624e-5d40-4a2b-8a31-58a8a1637fc3","arxiv_id":"2608.03861","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"A rank-ℓ local operator on the three-leg AKLT ladder reads boundary-encoded qubits only through the L=ℓ transfer-matrix sector, with decay length ξ(L=ℓ), and leg exchange generates a multi-qubit SWAP gate.","lead":"This paper builds the exact ground state of a three-leg AKLT ladder and shows that the geometric leg-exchange symmetry creates a logical qubit permutation gate, while a new local-probe measure reveals how fast bulk operators can read out boundary-encoded quantum information. The results give a quantitative link between SPT symmetry, lattice geometry, and edge-state protection.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (80) rests on unproven Assumptions (A1)–(A3) in Appendix G4a: injectivity, transfer-matrix diagonalizability, and invertibility of the dominant fixed-point matrices r̂0, l̂0. If r̂0 or l̂0 were singular, the P̃ normalization, the trace-subtraction selection, and the resulting decay law would not","rationale":"The reader's weakest_assumption identifies exactly the premise I find most load-bearing: the unproven Assumptions (A1)-(A3) in Appendix G4a. I agree with the reader's assessment. The concern does not change the verdict, for several reasons. First, the assumptions are directly checkable and very plausibly satisfied: the AKLT-type construction yields injective MPS, the transfer-matrix sector decomposition is complete to 1.03×10^{-14}, the Gram matrix has full rank K=64, and wrong-sector reduced matrix elements sit at 10^{-16}. Second, the central claim is not only derived but also numerically verified: the fitted decay lengths reproduce ξ(L=1) and ξ(L=2) for every probe class, and the chain's rank-2 blindness is confirmed with an exact closed-form finite-size residue (Eq. 81), a falsifiable prediction. Third, the logical gate O(P_tb)=SWAP13⊗SWAP13 is verified by its eigenvalue multiplicities (40/24) and numerical non-factorizability. The unproven Hamiltonian gap (Sec. II A: 'expected to be gapped') is a related rigor gap, but it is not load-bearing for Eq. (80), which is an MPS/transfer-matrix statement independent of the gap. The concrete test above — condition numbers of r̂0, l̂0 plus the trace-subtraction check — would settle whether the derivation's premises hold. Since the concern is a verification gap rather than a detected error, and the numerical support for the central claim is strong, the reader's ACCEPT verdict should stand unchanged.","tokens_in":32878,"tokens_out":38859,"duration_ms":356748,"concrete_test":"Directly reconstruct the 64×64 transfer matrix T = Σ_s A^s ⊗ A^s from the explicit three-leg ladder MPS tensor (D=8, d_phys=80). Extract the dominant left/right eigenvectors, reshape into 8×8 matrices l̂0, r̂0, and test (A3): compute the relative smallest singular values σ_min(r̂0)/σ_max(r̂0) and σ_min(l̂0)/σ_max(l̂0), and verify the fixed-point equation Σ_s A^s r̂0 (A^s)^† = λ_0 r̂0 to machine precision. If either ratio is at or below the 10^{-10} level, Assumption (A3) fails and Eqs. (G13), (G21)-(G22) — hence Eq. (80) — do not follow. If both ratios are well above zero, also evaluate Tr(r̂0^{-1}r̂i) over the subleading eigenvectors of each L≠0 sector and check that it vanishes at O(η_1^N), confirming the trace subtraction that selects only the L=ℓ channel.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline quantitative result, the selection-rule decay law δ_edge(dist) ≃ A e^{-dist/ξ(L=ℓ)} (Eq. 80), is derived in Appendix G4 from a spectral/channel decomposition of the transfer matrix. The derivation is built on three explicit assumptions stated but not proven for the three-leg ladder: (A1) the MPS is injective and T = Σ_s A^s ⊗ A^s is diagonalizable with unique dominant eigenvalue; (A2) biorthogonal eigenvectors; (A3) the reshaped dominant fixed points r̂0, l̂0 are invertible. The most load-bearing step is (A3). Invertibility of Ω00 = r̂0^T ⊗ l̂0 is needed (i) for the normalization P̃ = λ_0^{N/2}P + O(η_1^N) in Eq. (G13), and (ii) for the trace-subtraction identity Tr bΩ_{ij} = Tr(r̂0^{-1}r̂i)Tr(l̂j l̂0^{-1}) in Eq. (G21), whose vanishing for L_i≠0, L_j≠0 (Eq. G22) is exactly what eliminates all channels except the L=ℓ one. If r̂0 or l̂0 were singular, the thermodynamic-limit Gram matrix G_N ≈ λ_0^N Ω00 would have rank < 64, the claimed K=64 code space would be ill-defined, and the channel argument for Eq. (80) would break down. The paper itself flags the gap by writing 'We assume the following' at the start of Appendix G4a. The numerical evidence (K=64, sector completeness 1.03×10^{-14}, wrong-sector matrix elements at 10^{-16}) makes the assumptions very plausible, but none of it is a direct check of the fixed-point condition or of (A3).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs an exact matrix product state (MPS) representation for the three-leg AKLT ladder with on-site symmetry group SO(3)×Z2, and uses it to study boundary-encoded quantum information. It derives the parent Hamiltonian, verifies the nontrivial Haldane phase via cohomology (H^2 = Z2), string order, and entanglement spectrum. It shows that global SO(3) rotations induce a continuous family of logical rotations on the K = 64 edge code space, while the geometric top–bottom leg exchange induces the logical gate O(P_tb) = SWAP13⊗SWAP13, a genuinely multi-qubit permutation not decomposable into single-qubit unitaries. The main quantitative result is a symmetry-resolved decay law for the distinguishability measure δ_edge of local probes: a rank-ℓ single-site operator couples to the edge subspace only through the L = ℓ sector of the transfer matrix, giving δ_edge(dist) ≈ A e^{-dist/ξ(L=ℓ)} (Eq. 80). The derivation uses an exact channel decomposition of the transfer matrix (Appendix G4), and the predicted decay lengths are confirmed numerically for rank-1 and rank-2 operators. A finite-size, position-independent channel is also identified and analyzed.","tokens_in":33381,"tokens_out":15004,"duration_ms":142950,"significance":"If correct, the paper provides a clean, exactly solvable demonstration that lattice geometry (leg exchange) expands the set of symmetry-protected logical gates beyond what single-chain SPT systems offer, and that the SO(3) tensor rank of a local probe controls its exponential reach into the bulk. The analytic channel decomposition in Appendix G is a valuable technique, and the numerical verification is thorough: sector completeness at 10^-14, wrong-sector matrix elements at 10^-16, and fitted decay lengths agreeing with the sector correlation lengths to within a few percent. The paper is honest about the scope (gate set not universal, no finite code distance, passive protection). These are strong positives.","major_comments":[{"comment":"The central decay law (80) is derived under Assumptions (A1)–(A3), stated but not proved for the three-leg ladder. In particular, (A3) requires the reshaped dominant fixed points r̂0 and l̂0 to be invertible; this is used to normalize P̃ in (G13) and to factor the traces in (G21), whose vanishing in (G22) eliminates all channels except L = ℓ. If r̂0 or l̂0 were singular, the Gram matrix would have rank < 64 for large N and the selection-rule decay law would not follow in the stated form. The numerical evidence (rank K = 64, sector completeness 1.03×10^-14) is suggestive but is not a direct check of the thermodynamic fixed-point matrices. Please either prove these assumptions for this AKLT construction—e.g., injectivity and positivity of the fixed point—or provide direct numerical verification (e.g., determinants/condition numbers of r̂0 and l̂0, and semisimplicity of T). This is the load","section":"Appendix G4a, Eqs. (G13), (G21)–(G22), Eq. (80)"}],"minor_comments":[{"comment":"The paragraph beginning 'The scalar channel (a) is proportional to I_K...' appears twice, verbatim. Please remove the duplicate.","section":"Appendix G4 (after Eq. G29 and Table IV)"},{"comment":"The notation '1/√4 2' in the displayed derivation is ambiguous; it should read (1/√K) × 2 with K = 4, or an explicit multiplication sign should be inserted.","section":"Appendix G4i, Eq. (G29)"},{"comment":"In the sentence 'According to the Knill–Laflamme conditions [3], When considering...', the word 'When' should be lowercase.","section":"Sec. V A"},{"comment":"The spectral decomposition should state explicitly that the eigenbasis is chosen to be symmetry-adapted with respect to J²_adj, so that the L labels are well defined even when an eigenvalue is shared across sectors (e.g., λ/λ0 = −0.0341 appears in three sub-sectors in Table II).","section":"Appendix G4a, Eq. (G8)"},{"comment":"The fitted decay lengths differ from the theoretical sector lengths by 1–6% (e.g., 1.3459 vs 1.3630 for S^z_rung; 0.5090 vs 0.5418 for ∑(S^z_leg)^2). The text explains this as a fitting-window effect but does not report the restricted-fit values that 'can bring every probe to the sector prediction.' Including those values would quantitatively substantiate the claim of agreement.","section":"Sec. V B, Table III"}],"recommendation":"major_revision","confidential_remarks":"The paper is well within scope and the central physical claims are likely correct. The main technical gap is the unproven Assumptions (A1)–(A3) in Appendix G4; this is fixable with a direct proof or numerical verification and does not require reworking the results. The reader's report recommended acceptance; my assessment is slightly more cautious because of this load-bearing premise, but I expect the revision to be straightforward."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper is a solid, self-contained piece of exact MPS analysis. The two genuinely new things are the leg-exchange logical gate O(P_tb)=SWAP13⊗SWAP13, which is a non-factorizable qubit permutation with no single-chain analogue, and the rank-resolved selection rule δ_edge(dist) ≃ A e^{-dist/ξ(L=ℓ)}, derived from the Wigner–Eckart structure of the transfer matrix rather than just fitted. The derivation in Appendix G is careful and the numerical fits agree with the predicted sector correlation lengths. The finite-size channel (position-independent, system-size suppressed) is a nice touch that explains the plateau in the rank-2 ladder probes. Credit is due: the MPS construction, the cohomology argument, the entanglement-spectrum diagnostics, and the channel decomposition are all worked out in detail, and the paper is honest about its limits—the gate is only a permutation, not universal, and the protection is passive.\n\nThe soft spot is the one flagged in the stress-test note: Assumptions (A1)–(A3) in Appendix G4a are stated but not proved for this specific tensor. If the dominant fixed-point matrices r̂0 and l̂0 were singular, the normalization P̃ and the trace-subtraction identity would break, and Eq. (80) would not follow. That said, for an injective MPS these properties are typically theorems, not assumptions, and the numerical evidence (Gram rank 64, sector completeness at 1e-14, wrong-sector matrix elements at 1e-16) makes the assumptions very plausible. A referee should ask the authors to justify or prove injectivity explicitly, but this is a fixable gap in exposition, not a load-bearing flaw. The parent Hamiltonian gap is also asserted rather than proved, but that is standard for AKLT-type constructions.\n\nMy overall read: the central claims hold up, the paper is worth reading for anyone working on SPT edge encoding or transfer-matrix selection rules, and it deserves a serious referee rather than a desk reject. I would send it to review and expect the authors to tighten the injectivity discussion and perhaps say more about how generic the assumptions are.\n\nRecommendation: engage with it; accept with revisions in mind.","headline":"A careful, exact-solvable analysis of the three-leg AKLT ladder that delivers a concrete geometric gate and a selection-rule decay law; the central claims hold up, with only the usual unproven injectivity caveat.","tokens_in":33845,"tokens_out":1646,"would_cite":true,"duration_ms":19413,"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":"In the three-leg AKLT ladder, a local probe's reach into the edge code is fixed by its spin rank: rank-ℓ operators decay with the ℓ-th transfer-matrix sector's correlation length, and leg exchange yields a six-qubit permutation gate.","keywords":["AKLT ladder","symmetry-protected topological order","matrix product state","logical gate","local probe","Wigner–Eckart selection rule","transfer matrix","Haldane phase"],"falsifier":"Compute δ_edge for the quadrupole probe (S^z)^2 in the spin-1 AKLT chain at N=20: the paper predicts the position-independent value (2√6/9)3^{-(N-1)} ≈ 4.68×10^{-10}. If an independent MPS calculation gives a different value or a position-dependent profile, the trace-subtraction and channel decomposition behind Eq. (80) are wrong. A second check: in the three-leg ladder, a P_tb-odd rank-1 probe must decay with ξ(L=1,−)=0.7907 rather than the even-sector length 1.3630.","tokens_in":1740,"feed_emoji":"⚛️","tokens_out":1858,"duration_ms":74348,"temperature":0.7,"pith_summary":"The paper studies the three-leg AKLT ladder, an exactly solvable spin ladder whose open-boundary ground space encodes six logical qubits protected by SO(3)×Z2 symmetry. It establishes two linked facts. First, the leg-exchange symmetry acts on the encoded space as a logical permutation, SWAP13⊗SWAP13, that cannot be factored into single-qubit gates—something no single-chain AKLT system offers. Second, a single-site local probe with spherical-tensor rank ℓ can reach the edge code only through the L=ℓ angular-momentum sector of the transfer matrix, so its distinguishability decays as A e^{-dist/ξ(L=ℓ)} with the sector correlation length. This gives a quantitative, symmetry-resolved law for how much local operators can see of boundary-encoded quantum information.","feed_headline":"Spin rank sets how deep a local probe sees the quantum edge","feed_subtitle":"In the AKLT ladder, rank-ℓ probes decay by the ℓ-th transfer-matrix sector; leg-swap gives a permutation gate.","key_machinery":"The transfer matrix of the MPS, acting on the doubled virtual space (C²)⊗3⊗(C²)⊗3, together with its SO(3)×Z2 sector decomposition. Because the transfer matrix commutes with the adjoint Casimir, it block-diagonalizes into angular-momentum sectors L=0,1,2,3, refined by leg-exchange parity. The Wigner–Eckart selection rule then forces a rank-ℓ probe to propagate only through the L=ℓ sector, so its decay length is ξ(L=ℓ) = -1/ln|λ_L/λ_0|.","core_discovery":"The central claim is that local accessibility of the SPT edge code is governed by a Wigner–Eckart selection rule: in the exact MPS representation, a rank-ℓ single-site operator couples the L=0 sector of the transfer matrix only to the L=ℓ sector, so the distinguishability measure δ_edge decays as δ_edge(dist)≃A e^{-dist/ξ(L=ℓ)} for 1≪dist≪N. For the three-leg ladder the relevant sectors L=1,2,3 coexist with correlation lengths 1.3630, 0.5418, and 0.2960, making the hierarchy observable; in the single chain no L=2 sector exists, so a quadrupole probe is exactly blind in the thermodynamic limit, with only a size-dependent residue (2√6/9)3^{-(N-1)}. Independently, the leg-exchange symmetry is s","pith_inferences":["If the selection rule holds for arbitrary local error sets, passive protection can be tuned by symmetry labels: noise composed solely of rank-2 or higher tensors is exponentially suppressed in the bulk with short decay lengths, while rank-1 dephasing remains the dominant accessible channel.","The leg-exchange gate, although not universal, points to lattice geometry as a resource for implementing multi-qubit permutations in SPT codes without addressing individual qubits; combining such geometric gates with measurements could enlarge the accessible logical gate set.","Extending the same tensor construction to five- and seven-leg ladders or to cyclic tubes would test whether the hierarchy ξ(L=1)>ξ(L=2)>⋯ persists and whether cyclic leg permutations generate further discrete logical gates—a direction the paper explicitly leaves open.","A practical consequence of the boundary-probe behavior is that any useful passive code must treat the first unit cell as fully exposed; the protection is inherently about bulk operators, so realistic error models should separate boundary and bulk noise channels."],"forward_implications":["Local probes with the same SO(3) tensor rank produce identical distinguishability profiles; the operator's strength and microscopic form do not enter the decay rate.","In the spin-1 AKLT chain, rank-2 operators are exactly blind to the edge code in the thermodynamic limit, leaving only the position-independent finite-size residue (2√6/9)3^{-(N-1)}.","In the three-leg ladder, rank-2 noise decays with ξ=0.5418 before leveling off onto a size-dependent finite-size plateau, while rank-1 probes decay with ξ=1.3630.","The leg-exchange symmetry yields the logical permutation gate O(P_tb)=SWAP13⊗SWAP13, which is not a product of single-qubit unitaries; the gate set remains non-universal and no finite code distance is defined because boundary probes act with full strength.","The same selection-rule machinery predicts that multi-site error operators couple through Clebsch–Gordan composition, |ℓ1−ℓ2|≤L≤ℓ1+ℓ2, so two separated local errors should show exponential decay of distinguishability with their separation."],"supporting_citations":[{"why":"Supplies the Knill–Laflamme error-correction conditions that motivate the distinguishability measure δ_edge for how much a local operator resolves code states.","marker":"[3]"},{"why":"Establishes the H2(SO(3)×Z2,U(1)) classification and symmetry-protected edge degeneracy used to place the ladder in the Haldane phase.","marker":"[4]"},{"why":"Provides the group-cohomology framework and Künneth decomposition used to compute the SPT invariant of the full on-site symmetry group.","marker":"[6]"},{"why":"Gives the MPS intertwining relation that converts global on-site symmetries into virtual-boundary logical gates.","marker":"[7]"},{"why":"Supplies the AKLT valence-bond construction—Clebsch–Gordan projectors and the parent Hamiltonian—on which the exact tensor is built.","marker":"[13–15]"},{"why":"Provides the transfer-matrix formalism and sector correlation lengths used in the channel decomposition and in deriving Eq. (80).","marker":"[18, 19]"},{"why":"Gives the Wigner–Eckart selection rule that restricts a rank-ℓ operator to the L=ℓ transfer-matrix sector.","marker":"[35, 36]"}],"fun_headline_variants":["Rank-ℓ probes only see ℓ-sector in AKLT ladder","Leg-exchange symmetry gives logical gate in AKLT ladder","Wigner–Eckart sets probe decay length in AKLT ladder","AKLT edge code: probe rank selects decay sector","Three-leg AKLT: leg swap is a logical gate"],"cache_read_input_tokens":35456,"weakest_assumption_plain":"The argument assumes the ground-state tensor is injective and the two dominant boundary matrices derived from it are invertible; if either failed, the normalization step that converts raw boundary states into an orthonormal code basis would break, and the exponential decay law would not follow in its stated form.","fun_headline_variants_meta":{"raw":{"variants":["Rank-ℓ probes only see ℓ-sector in AKLT ladder","Leg-exchange symmetry gives logical gate in AKLT ladder","Wigner–Eckart sets probe decay length in AKLT ladder","AKLT edge code: probe rank selects decay sector","Three-leg AKLT: leg swap is a logical gate"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000657,"raw_usage":{"total_tokens":2882,"prompt_tokens":818,"completion_tokens":2064,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":562,"completion_tokens_details":{"reasoning_tokens":1980}},"tokens_in":562,"tokens_out":2064,"duration_ms":16885,"temperature":1.0,"reasoning_tokens":1980,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T10:53:10.475741+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute δ_edge for the quadrupole probe (S^z)^2 in the spin-1 AKLT chain at N=20: the paper predicts the position-independent value (2√6/9)3^{-(N-1)} ≈ 4.68×10^{-10}. If an independent MPS calculation gives a different value or a position-dependent profile, the trace-subtraction and channel decomposition behind Eq. (80) are wrong. A second check: in the three-leg ladder, a P_tb-odd rank-1 probe must decay with ξ(L=1,−)=0.7907 rather than the even-sector length 1.3630.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Knill–Laflamme error-correction conditions that motivate the distinguishability measure δ_edge for how much a local operator resolves code states."},{"cited_title":"The many-body state can be written as |Ψ⟩= DX α=1 |Lα⟩ ⊗ |Rα⟩,(E1) where|L α⟩and|R α⟩are the half-chain states associated with virtual boundary indexα","cited_arxiv_id":null,"evidence_quote":"Establishes the H2(SO(3)×Z2,U(1)) classification and symmetry-protected edge degeneracy used to place the ladder in the Haldane phase."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the group-cohomology framework and Künneth decomposition used to compute the SPT invariant of the full on-site symmetry group."},{"cited_title":"The 64×64 matrix is reshaped into a rank-12 tensor with six input and six output indices, each of dimension 2","cited_arxiv_id":null,"evidence_quote":"Gives the MPS intertwining relation that converts global on-site symmetries into virtual-boundary logical gates."}],"review_version":1}