{"id":"33a53797-611f-4412-ab5c-ab83bcc11305","arxiv_id":"2608.11503","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A rigorous separation of global, symmetry-allowed, and dynamically accessible spectral gaps for compact Hamiltonian encodings of set cover, including an explicit even-cycle family with an Omega(n^-13) cyclic-gap certificate.","lead":"This paper defines which part of a Hamiltonian's spectrum a quantum adiabatic protocol can actually use: the cyclic space generated by the protocol, not just the full symmetry-allowed space. It proves on even-cycle set-cover instances that one interpolation has a dark gap closure while a separately constructed parent path has a uniformly polynomial accessible gap.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the gap-certificate chain in Props. 5-7 is internally consistent; the remaining risk is proof verification, not a demonstrated error.","rationale":"After reading the full manuscript, I cannot identify a load-bearing internal flaw. The paper's central distinction K_u ⊆ H_sym is definitionally sound and is supported by both the exact even-cycle multiplicity closure (Sect. 5.2) and the parent-path construction. The most intricate part, the uniform gap lower bound, is assembled from standard tools (Hermon-Salez comparison, log-concave couplings, Caputo-Sasada recursion), and the constants cohere when re-derived from the stated normalizations. I specifically checked Eq. (89): it follows from (86) and (82) without a hidden 1/N, and the substitution in (90) reproduces 1024 n^{-(a+6)}. The adiabatic derivative bounds in Cor. 8 are consistent with the row-sum estimates and the JRS bound. Therefore the Reader's conditional verdict should stand: the paper is careful, honest about limitations, and the only caveats are artifact pinning and the absence of machine-checked proofs, not a demonstrated error. Verdict adjustment: none.","tokens_in":14167,"tokens_out":46648,"duration_ms":450745,"concrete_test":"Independently verify the normalization chain by exact diagonalization: for N=3, m=1..10 and τ∈{1,0.5,0.1}, compute the gap of L^*_{3,m}=P-I for the three-box heat-bath kernel and confirm it is at least 1/3; for N=5, m=2 and the same τ, confirm that the bound in Eq. (81) is at least 1/N. If both hold, the proof chain is sound; if either fails, recompute Prop. 7 with the corrected constant and reassess the n^{-13} certificate.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is the uniform gap certificate for the Johnson/Metropolis parent path (Prop. 7 → Cor. 8). I stress-tested exactly the chain the Reader flagged: the zero-range reduction (Prop. 5), the three-box heat-bath contraction (Lemma 6 and App. A.2), and the Caputo-Sasada recursion in App. A.3. Recomputing from the stated normalizations, the conditional pair gap is at least τ/k^2 (App. A.4, Eq. 86), the heat-bath gap is at least 1/k (Eq. 82), and the mean-field decomposition (Eq. 83) gives gap_ZRP ≥ 2τ k^{-3}; together with 1 - cos(2π/k) ≥ 8/k^2 and k = n/2 this yields 1024 n^{-(a+6)}, so Prop. 7 is internally consistent. The W1-to-spectral-gap step in App. A.2 is valid because the reversible kernel contracts the d1-Lipschitz seminorm by 2/3 and every nonconstant eigenfunction on the finite connected metric space has positive Lipschitz seminorm. I did not find a concrete normalization error, a missing 1/N factor, or an exponent error. The remaining risk is the usual absence of machine-checked verification and the not-yet-pinned repository, not a known flaw in the argument.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies multi-register Hamiltonian encodings of Minimum Set Cover and argues that, for a fixed initial state and a symmetry-preserving interpolation, the relevant spectral object is the restriction of the Hamiltonian to the protocol's cyclic (Krylov) space, which can be strictly smaller than the joint symmetry-fixed space. It proves a sector-resolved localization bound, a stability theorem for dark symmetry-sector crossings, exact orbit-quotient diagnostics, and, on even cycles, constructs a Johnson/Metropolis parent path with endpoint cover probability 1-O(n^{-5}), a uniform D_n-fixed gap Omega(n^{-13}) obtained through a zero-range comparison and a log-concave three-box coupling, and a conditional adiabatic runtime O(epsilon^{-1} n^{41} log^2 n). The paper explicitly disclaims any quantum-speedup claim.","tokens_in":14508,"tokens_out":37733,"duration_ms":337504,"significance":"The conceptual separation of global, symmetry-allowed, and dynamically accessible spectra is clearly valuable, and the even-cycle family gives a concrete analytic witness. I checked the main proof chain: the constants in Prop. 7 (2 tau k^{-3} and 1024 n^{-(a+6)}) and the runtime exponents n^{27} and n^{41} in Cor. 8 are consistent with the stated normalizations. The W1-to-spectral-gap step in App. A.2 is valid because the heat-bath kernel contracts the relevant Lipschitz seminorm by 2/3, and the Caputo-Sasada recursion in Eq. (81) is used in its standard normalization. The paper is commendably honest about the conditional nature of the adiabatic corollary and the non-certified character of the finite-instance numerics.","major_comments":[],"minor_comments":[{"comment":"The statement that the physically relevant spectrum is the cyclic-space spectrum should be explicitly restricted to paths whose Hamiltonians lie in the algebra used to define K_u, or to a cyclic space generated by the full family of path Hamiltonians; as written, the sentence could be read as applying to arbitrary symmetry-preserving interpolations.","section":"Section 8"},{"comment":"Please state explicitly which version of the Jansen-Ruskai-Seiler bound is used and write out the error estimate, since the n^{41} runtime exponent is a headline result and the current proof only cites the reference.","section":"Corollary 8"},{"comment":"The repository is given only as an unversioned URL and the Zenodo DOI is promised but not yet assigned; these should be completed before publication for reproducibility.","section":"Code and data availability"},{"comment":"The definition of W is ambiguous as printed: it should read W = (11^T - I)/(n_B-1), with parentheses around 11^T - I.","section":"Eq. (32)"},{"comment":"The notation [z^k] for coefficient extraction is used without definition; please define it the first time it appears.","section":"Eq. (15)"},{"comment":"A formatting issue: the displayed definition of the metric appears as '1 2 sum |x_i - y_i|' and should read (1/2) sum |x_i - y_i|.","section":"Appendix A.2"},{"comment":"The finite numerical diagnostics are clearly labeled as non-certified, but the paper should state once more that they are illustrative only and are not used as an asymptotic theorem; this is already implied but would benefit from an explicit sentence near Table 1.","section":"Section 5.1"}],"recommendation":"minor_revision","confidential_remarks":"No circularity found. The stress-test concerns about App. A.2 and Eq. (81) did not land as errors after recomputation: the W1 contraction coefficient is 2/3 with the metric defined as half the componentwise l1 distance, and the Caputo-Sasada normalization gives the stated 1/k heat-bath gap. The remaining risk is purely the reliance on the cited Hermon-Salez and Caputo-Sasada bounds, which is standard practice. The paper's no-speedup disclaimer and conditional adiabatic framing are appropriate, and the three-way spectral distinction is a genuine contribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a serious theory paper and the central distinction it draws is worth taking seriously. The physically relevant spectrum is not the full spectrum, nor automatically the symmetry-fixed one, but the spectrum restricted to the cyclic space generated by the protocol. That claim is made precise, and the analytic even-cycle family supports it with an exact global multiplicity closure that is dynamically dark and a separate Johnson/Metropolis parent path with a uniform cyclic-gap certificate Ω(n^-13). I checked the gap-certificate chain against the appendices, as did the stress-test note, and it is internally consistent: the Hermon–Salez comparison, the three-box contraction, and the Caputo–Sasada recursion all line up with the stated normalizations. No concrete error surfaced.\n\nWhat the paper does well: it separates the conceptual claim from the quantitative one, it makes no speedup claim, and it flags the conditional nature of the adiabatic corollary (abstract Hamiltonian access, Dicke-state preparation, the parent endpoint is not the original problem Hamiltonian). The sector-resolved localization bound with the kinetic floor is a nice methodological point, and the stability theorem for transverse dark crossings is clean. The numerical finite-instance work is clearly labeled as diagnostics, not as proof, and the residuals are reported.\n\nThe soft spots are real but proportionate. The main one is that the uniform gap certificate rests on a chain of comparisons with a few proof sketches in Appendix A; the stress-test found no flaw, but this deserves a full line-by-line verification by a referee with expertise in reversible Markov chains and the Caputo–Sasada recursion. A second, minor, point is reproducibility pinning: the repository exists but has no commit hash or DOI yet, so the frozen instances are not fully pinned. The abstract Hamiltonian-access assumption is stated honestly, but it does limit direct practical relevance.\n\nWho this is for: researchers working on adiabatic quantum optimization, symmetry reduction in QAOA/quantum walks, and spectral-gap certificates for annealing paths. It deserves a serious referee. I would recommend sending it to peer review with a request for a careful checking of Appendix A and for the author to pin the artifact before publication.","headline":"A serious, honest theory paper; the three-way spectral separation and the even-cycle family with a proven polynomial cyclic gap are the real contributions, and the proof chain survives close checking.","tokens_in":15033,"tokens_out":1625,"would_cite":true,"duration_ms":15030,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"When a Hamiltonian evolution preserves symmetries, the physically relevant spectrum is that of the Hamiltonian restricted to the protocol's cyclic subspace—not the full spectrum or the entire symmetry-allowed space.","keywords":["cyclic space","symmetry-allowed space","Hamiltonian spectrum","adiabatic quantum computing","set cover","orbit quotient","zero-range process","spectral gap"],"falsifier":"Compute the exact spectral gap of $H_a$ restricted to the $D_n$-fixed space at $a=7$ for even $n$ from 20 to 200; if it falls below $1024\\, n^{-13}$ at any tested $n$, the comparison-chain normalization is wrong. Independently, diagonalize the three-box heat-bath kernel for several small totals $m$: if any nonconstant eigenvalue exceeds $2/3$ in absolute value, the claimed contraction constant and all subsequent gap bounds fail.","tokens_in":13937,"feed_emoji":"⚛️","tokens_out":8988,"duration_ms":79121,"temperature":0.7,"pith_summary":"The paper argues that for any Hamiltonian evolution whose initial state and interpolation preserve symmetries, the physically relevant spectrum is the spectrum of the Hamiltonian restricted to the cyclic space generated by the protocol, not the full spectrum and not even the whole symmetry-fixed space. It makes this concrete in a compact multi-register encoding of Minimum Set Cover, where register permutations and the faithful base action of the incidence automorphism group define a symmetry $S_k \\times G_B$, and the cyclic space is strictly smaller than the symmetry-allowed space in several small instances. A sector-resolved localization theorem certifies that the ground state of a sector has high cover probability whenever the sector contains feasible covers and the invalid-state separation exceeds the sector's kinetic floor. On an even-cycle family the original linear path has an exact global ground-multiplicity closure that is dynamically dark, while a constructed parent path carries a uniform cyclic gap $\\Omega(n^{-13})$ and yields a conditional polynomial adiabatic runtime; no quantum speedup is claimed.","feed_headline":"Only the cyclic-space spectrum governs a symmetric protocol's gap.","feed_subtitle":"A compact set-cover study separates full, symmetry-allowed, and dynamically accessible spectra.","key_machinery":"The central object is the cyclic space $K_u = \\mathcal{A}|u\\rangle$, where $\\mathcal{A}$ is the unital star-algebra generated by $H_{\\mathrm{init}}$ and $H_{\\mathrm{prob}}$; it is the minimal reducing subspace that contains the initial state and can be strictly smaller than the symmetry-fixed space $H_{\\mathrm{sym}}$. The argument runs through exact orbit quotients of the joint action $S_k \\times G_B$, a sector-resolved Schur-complement bound that inserts the sector kinetic floor into the invalid-state separation, and a parent Hamiltonian built as the symmetric discriminant of a reversible swap-chain generator with Gibbs-amplitude ground state, starting from the uniform symmetric Dicke state. The uniform gap certificate combines a zero-range spectral-gap comparison reducing cycle geometry to a mean-field kernel, a three-box heat-bath contraction with gap at least $1/3$, and a heat-bath gap recursion giving at least $1/k$; together these yield the accessible gap bound $\\Delta_{D_n}(a) \\ge 1024\\, n^{-(a+6)}$.","core_discovery":"The central claim is that protocol-dependent dynamics is governed by the cyclic space $K_u = \\mathcal{A}|u\\rangle$, the smallest common reducing subspace generated by the endpoint Hamiltonians acting on the initial state, and that $K_u \\subseteq H_{\\mathrm{sym}} \\subseteq H_{\\mathrm{valid}}$, with strict inclusions occurring in concrete cases. For the even-cycle family, the paper proves an exact multiplicity closure: at $s_c = (n-1)/(2n-1)$ the full linear Hamiltonian has $2k!$ degenerate ground states while the joint-fixed ground state is unique and its excitation gap is the constant $\\lambda (n-1)/(2n-1)$, so the global closure is dark to the symmetric protocol. An alternative parent path built from a reversible swap-chain generator has endpoint cover probability $1-O(n^{-5})$ and, by a zero-range comparison and a three-box log-concave coupling, cyclic gap at least $1024\\, n^{-13}$ for $0 \\le a \\le 7$, yielding the conditional adiabatic runtime $T \\ge C \\epsilon^{-1} n^{41} (\\log n)^2$ in the abstract Hamiltonian-access model.","pith_inferences":["If the cyclic-space principle generalizes, symmetry-based adiabatic or variational algorithms outside this encoding may need to certify dynamical subspaces, and reported spectral gaps based only on the fixed space might overestimate performance.","The exactly solvable even-cycle family with a dark multiplicity closure is a candidate stress test for adiabatic theorems and numerical gap estimators, since it separates global degeneracy from accessible gap in a controllable setting.","The kinetic-floor localization bound appears transferable to other constrained optimization Hamiltonians with kinetic hopping: any sector whose kinetic floor is computable could receive a similar leakage certificate without solving the full spectrum.","The comparison-chain proof technique for the gap certificate—zero-range reduction plus log-concave coupling—might yield explicit polynomial lower bounds for other reversible annealing Hamiltonians on symmetric state spaces, though this is not claimed in the paper."],"forward_implications":["A symmetry-based gap analysis for any adiabatic or QAOA-style protocol must be performed in the protocol's cyclic space; global gap closures can be invisible to the dynamics and do not by themselves obstruct adiabatic evolution.","For the even-cycle family, the original linear interpolation has a closed global multiplicity at $s_c$ while its joint-fixed excitation gap stays positive, so global degeneracy and accessible gap are independent structural quantities.","The constructed parent path gives a certified cover probability $1-O(n^{-5})$ and a uniform accessible gap $\\Omega(n^{-13})$, implying a polynomial adiabatic runtime in the abstract Hamiltonian-access model conditional on Dicke-state preparation and access to the parent Hamiltonian.","The sector-resolved localization bound provides a sufficient condition for high cover probability in a symmetry sector: feasible covers present and invalid-state separation larger than the sector kinetic floor.","Changing the initial state or breaking a preserved symmetry changes the cyclic space and can turn a dark crossing into an accessible one, so inaccessibility claims are always conditional on state preparation and generator algebra."],"supporting_citations":[{"why":"Supplies the orbit-quotient construction for symmetry-confined quantum walks used to define exact quotients of the joint action.","marker":"[7]"},{"why":"Supplies the symmetric-discriminant construction that turns the reversible swap chain into a parent Hamiltonian with the Gibbs-amplitude ground state.","marker":"[9]"},{"why":"Provides the annealing-time framework connecting such discriminants to quantum simulations of classical annealing.","marker":"[10]"},{"why":"Supplies the quantitative gapped adiabatic bound that converts the gap and derivative controls into the polynomial runtime.","marker":"[11]"},{"why":"Provides the Dirichlet-form comparison from the cycle zero-range chain to the mean-field kernel, giving the geometric factor in the gap bound.","marker":"[14]"},{"why":"Provides the stochastic monotonicity theorem used to prove the three-box heat-bath contraction with gap at least 1/3.","marker":"[15]"},{"why":"Supplies the heat-bath spectral-gap recursion in the normalization used to obtain the 1/k lower bound.","marker":"[16]"},{"why":"Extends the heat-bath recursion in the same normalization, supporting the uniform gap certificate.","marker":"[17]"}],"fun_headline_variants":["Cyclic space, not full spectrum, governs protocol gaps.","Set cover: symmetric gaps live only in cyclic space.","Global degeneracy is dark to symmetry-preserving dynamics.","Adiabatic runtime certified by cyclic-space gap.","Stability: dark crossings persist; cyclic gap constant."],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the normalization factors in the zero-range comparison and the three-box heat-bath contraction are exactly as stated; a single wrong constant there would destroy the gap certificate and the polynomial runtime.","fun_headline_variants_meta":{"raw":{"variants":["Cyclic space, not full spectrum, governs protocol gaps.","Set cover: symmetric gaps live only in cyclic space.","Global degeneracy is dark to symmetry-preserving dynamics.","Adiabatic runtime certified by cyclic-space gap.","Stability: dark crossings persist; cyclic gap constant."]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000349,"raw_usage":{"total_tokens":1970,"prompt_tokens":1068,"completion_tokens":902,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":684,"completion_tokens_details":{"reasoning_tokens":826}},"tokens_in":684,"tokens_out":902,"duration_ms":8931,"temperature":1.0,"reasoning_tokens":826,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:05:25.063730+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the exact spectral gap of $H_a$ restricted to the $D_n$-fixed space at $a=7$ for even $n$ from 20 to 200; if it falls below $1024\\, n^{-13}$ at any tested $n$, the comparison-chain normalization is wrong. Independently, diagonalize the three-box heat-bath kernel for several small totals $m$: if any nonconstant eigenvalue exceeds $2/3$ in absolute value, the claimed contraction constant and all subsequent gap bounds fail.","supporting_citations":[{"cited_title":"Quantum walks on quotient graphs","cited_arxiv_id":"quant-ph/0701173","evidence_quote":"Supplies the orbit-quotient construction for symmetry-confined quantum walks used to define exact quotients of the joint action."},{"cited_title":"Quantum speed-up of Markov chain based algorithms","cited_arxiv_id":null,"evidence_quote":"Supplies the symmetric-discriminant construction that turns the reversible swap chain into a parent Hamiltonian with the Gibbs-amplitude ground state."},{"cited_title":"A version of Aldous' spectral-gap conjecture for the zero range process","cited_arxiv_id":"1808.00325","evidence_quote":"Provides the Dirichlet-form comparison from the cycle zero-range chain to the mean-field kernel, giving the geometric factor in the gap bound."},{"cited_title":"Increasing properties of P´ olya frequency functions","cited_arxiv_id":null,"evidence_quote":"Provides the stochastic monotonicity theorem used to prove the three-box heat-bath contraction with gap at least 1/3."},{"cited_title":"On the spectral gap of the Kac walk and other binary collision processes","cited_arxiv_id":"0807.3415","evidence_quote":"Supplies the heat-bath spectral-gap recursion in the normalization used to obtain the 1/k lower bound."},{"cited_title":"On the spectral gap of the kac walk and other binary collision processes on $d$-dimensional lattice","cited_arxiv_id":"1308.5096","evidence_quote":"Extends the heat-bath recursion in the same normalization, supporting the uniform gap certificate."}],"review_version":1}