{"id":"de393040-6b84-4b6b-846d-1d87e7103eb4","arxiv_id":"2506.03029","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"High-order perturbation series for Bose-Hubbard rings are reduced to a small set of elementary matrix elements, yielding accurate ground-state energies up to ninth order when compared with exact diagonalization.","lead":"This paper simplifies high-order perturbation theory for one-dimensional ring-shaped Bose-Hubbard models, reducing the number of terms needed to compute energy corrections up to ninth order. It provides a convergence criterion and validates the simplified series against exact diagonalization for a three-site trimer.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (1) defines all-to-all hopping, but Eqs. (19)-(21) use the nearest-neighbor Bloch spectrum; the two models coincide only for M=3, so the claimed M-independent convergence radius is not established for the stated ring model with M>3.","rationale":"I read the paper as claiming two things: a symbolic simplification of Kato perturbation theory for Bose-Hubbard rings, and a convergence criterion Eq. (21). The simplification itself is well supported: the unit gap and the fact that a single tunneling event from the uniform ground state lands in the first-excited subspace justify Eq. (14), the reflection argument justifies rules (i)-(iv), and the ED agreement for N=9, 27 is a strong check of Table 1. The reader's weakest assumption (repulsive interactions, N a multiple of M) is explicitly assumed in the paper and flagged in the footnote, so it is a scope condition rather than a hidden defect. The load-bearing problem I see is not the spectral gap but the mismatch between the Hamiltonian in Eq. (1) and the operator whose norm is used in Eq. (20). Because the trimer is the only case tested, the M-dependence of the model is invisible in the validation. For any M>3 the stated radius 1/(4N) is not justified by the text; either Eq. (1) should be restricted to nearest-neighbor pairs or Eq. (21) should be replaced by the M-dependent bound. Since the simplification method survives either fix, the appropriate disposition is conditional acceptance rather than rejection.","tokens_in":11987,"tokens_out":17642,"duration_ms":197841,"concrete_test":"Build the M=4, N=8 Bose-Hubbard Hamiltonian in two variants: (a) exactly as in Eq. (1), with all ordered i != j hoppings, and (b) nearest-neighbor ring hopping. For each, evaluate the ninth-order simplified series from Table 1 with the corresponding EME values and compare with exact diagonalization for |Omega/kappa| <= 1/(4N), e.g. Omega/kappa = 0.02. If variant (a) deviates while variant (b) agrees, Eq. (21) is not a valid convergence estimate for the Hamiltonian defined in the paper. A cheaper check: compute the largest eigenvalue of the one-particle hopping matrix from Eq. (1) for M=4; it is 3, not 2, so ||V|| = 3N, contradicting Eq. (20).","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Eq. (1) defines the perturbation as a sum over all ordered pairs i != j, i.e. tunneling between every pair of sites. The convergence estimate, however, is built on Eq. (19), which is the Bloch spectrum of nearest-neighbor ring hopping, and on the resulting M-independent norm ||V|| = 2N in Eq. (20). These two operators are identical only for M=3, where every pair of sites is adjacent; the numerical validation (N=9, 27, trimer) therefore cannot distinguish them. If Eq. (1) is read literally for M>3, the maximum eigenvalue of V is N(M-1), so Kato's criterion (15) gives |Omega/kappa| <= 1/[2N(M-1)], not 1/(4N). The claimed M-independence of Eq. (21) is then false for the model as written. If nearest-neighbor hopping was intended, Eq. (1) must be corrected; as printed, the paper's central convergence claim is internally inconsistent for general rings. The simplification rules and Table 1 are unaffected, since both models have unit spectral gap and V|a> lies in the first-excited subspace, so this is a correctable definitional flaw rather than a failure of the perturbative construction.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper applies Kato's time-independent perturbation theory to one-dimensional ring-shaped Bose-Hubbard models. It exploits the unit spectral gap between the unperturbed ground state and the first excited manifold to reduce all higher-order energy corrections to products of elementary matrix elements built from S_1 and V, resulting in a compact guide (Table 1) for corrections up to ninth order. A convergence criterion is derived from Kato's norm estimate, giving |Omega/kappa| <= 1/(4N), and the simplified series is validated against exact diagonalization for Bose-Hubbard trimers with N=9 and N=27.","tokens_in":12189,"tokens_out":6638,"duration_ms":71163,"significance":"If the central claims hold, the paper provides a genuinely useful reduction of perturbation-theory bookkeeping for Bose-Hubbard rings: the number of required elementary matrix elements drops from 91 to 46 and the number of diagonal matrix elements from 490 to 50 up to ninth order (Table 2). The paper also ships a reproducible Python implementation (Ref. [25]) and makes explicit falsifiable predictions that are checked against independent exact diagonalization. The main significance, however, hinges on two things that need tightening: the definition of the tunneling operator in Eq. (1) is inconsistent with the convergence analysis, and the derivation of the simplification rules is sketched rather than proved. These issues are correctable within the manuscript's scope.","major_comments":[{"comment":"The model Hamiltonian in Eq. (1) is written as an all-to-all tunneling term (sum over all ordered pairs i != j), while Section 4.1 states that the perturbation operator V contains only nearest-neighbor couplings and Eq. (19) uses the nearest-neighbor Bloch spectrum. For M > 3 these operators are different: the spectral norm of the all-to-all hopping operator is N(M-1), not 2N, so Kato's criterion (15) would give |Omega/kappa| <= 1/[2N(M-1)] rather than Eq. (21). Because the numerical validation is restricted to M = 3, where the cycle graph and the complete graph coincide, the presented data cannot discriminate between the two definitions. This is a load-bearing inconsistency for the claimed M-independent radius of convergence. Please correct Eq. (1) to the intended nearest-neighbor ring hopping, or recompute the norm for the all-to-all model, and make the perturbation operator in Eq. (1) consistent with the rest of the text.","section":"§2, Eq. (1); §4.1; §4.2, Eqs. (19)–(21)"},{"comment":"The derivation of the simplification rules is too compressed for the central technical result. Equation (14) is stated after observing that V|a> lies in the first-excited subspace, but the explicit reduction of arbitrary powers S_k to S_1 and the sign factors in rules (i)–(iv) are not demonstrated. The rules require that every state reached by a single tunneling event from the unperturbed ground state has exactly the same unperturbed energy, one unit above the ground state; this holds only for repulsive interactions, uniform filling with M | N, and a perturbation that connects the ground state exclusively to that manifold. The footnote on attractive interactions acknowledges one failure mode, but the main text claims applicability to a wider class of Bose-Hubbard systems without stating these hypotheses. Please provide a complete derivation of Table 1 from Eq. (14) and state the precise conditions under which the simplification scheme applies.","section":"§4.1, Eq. (14) and rules (i)–(iv)"},{"comment":"The numerical validation is performed exclusively for the trimer, M = 3. As noted above, this case cannot test the M-independence of the convergence radius because the all-to-all and nearest-neighbor hopping operators coincide when M = 3. To support the claim that Eq. (21) applies to rings of arbitrary M, the manuscript should either provide a rigorous derivation with the corrected nearest-neighbor Hamiltonian or include at least one explicit check for M > 3 (for example, M = 4 with N = 8 or N = 16) within the estimated convergence radius. Without one of these, the M-independence claim is not established by the evidence presented.","section":"§4.3, Figs. 1–3"}],"minor_comments":[{"comment":"The dimension formula is written with an undefined symbol L: it should read D = (N+M-1)!/(N!(M-1)!). The subsequent trimer expression is consistent with this corrected formula.","section":"§5, Eq. (22)"},{"comment":"The proposed bijection B(j,k) = j(N - j - 3/2) + k is not integer-valued for j > 0 and does not reproduce Table A1 (for example, N = 3, j = 1, k = 0 gives 0.5 instead of 4). The correct mapping appears to be B(j,k) = j(2N - j + 3)/2 + k (or an equivalent form). Please correct the formula or the surrounding derivation.","section":"Appendix A, Eq. (A.15)"},{"comment":"The sentence claiming that the simplifications reduce the number of terms 'by almost a factor of two' is ambiguous: the number of diagonal matrix elements is reduced from 490 to 50 (about a factor of ten), while the number of elementary matrix elements is reduced from 91 to 46 (about a factor of two). Please clarify which quantity is meant.","section":"§5, discussion of Table 2"},{"comment":"There are several typographical errors that should be fixed in a revision, including 'constisting' in the Introduction, 'numerically aquired' in Section 5, and the missing article in 'with solely D ∝ N^2 and, thus, a diagonalization ...'.","section":"Throughout"},{"comment":"The sentence 'For one-dimensional Bose-Hubbard-Hamiltonians, the spectrum of the tunneling operator is best studied in the Bloch basis' should specify that this applies to the nearest-neighbor hopping operator; for the all-to-all operator of Eq. (1) as printed, the Bloch spectrum in Eq. (19) is not the correct spectrum when M > 3.","section":"§4.2, text before Eq. (19)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's citation of the author's own Zenodo code (Ref. [25]) is appropriate as a computational tool and does not constitute circular support for the physics claims. The fit to Physica Scripta seems reasonable. The main concern is the definitional inconsistency in Eq. (1); a straightforward correction to nearest-neighbor hopping, plus a more explicit derivation of the simplification rules, would make the paper's central claims defensible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This paper does something genuinely useful: for ring-shaped Bose-Hubbard models where the unperturbed ground state has a unit gap, it shows how to reduce the Kato perturbation series up to ninth order to a small set of diagonal matrix elements. The rules (i)-(iv) in Sec. 4.1, the collected DMEs in Table 1, and the counting in Table 2 are new and go beyond Eckardt and Teichmann et al. The code is on Zenodo, and the validation against exact diagonalization for the trimer with N=9 and 27 shows the series tracks the exact ground state inside the estimated convergence radius. The convergence criterion (21) is a practical, parameter-dependent bound. That is a solid methods contribution for people doing higher-order perturbation theory in Bose-Hubbard systems.\n\nThe main soft spot is the inconsistency between the Hamiltonian in Eq. (1) and the spectrum used for the norm estimate. Eq. (1) sums over all ordered pairs i≠j, i.e. all-to-all hopping. Eq. (19) is the Bloch spectrum for nearest-neighbor ring hopping, which gives ||V||=2N and hence the M-independent bound 1/(4N). For M>3 these are different operators; the all-to-all V has spectral norm (M-1)N, so the bound would be 1/[2N(M-1)]. The numerical validation only uses M=3, where the two coincide, so it cannot distinguish them. The text elsewhere clearly speaks of nearest-neighbor tunneling—“each lattice site only allows for two directions of travel”—so Eq. (1) is most likely a typo. It should be corrected before publication. This is a definitional flaw, not a failure of the perturbation scheme itself; the simplification rules only rely on the unit gap and V|a> lying in the first-excited subspace, which holds for both models.\n\nThe other caveats are minor. The proof of the simplification rules is informal; a rigorous argument for arbitrary order would strengthen the paper. The generalization to other lattice geometries is asserted but not demonstrated, which is fine as an outlook. The convergence bound is crude (Kato's own words), but it is honestly presented and the numerical tests show the series actually does better than the bound in some cases.\n\nBottom line: this is a practical, checkable contribution to the perturbation-theory toolbox. I would send it to review, but insist that the Hamiltonian be corrected and, ideally, that the authors add a more explicit justification of the reduction rules. It's not deep physics, but it is careful and reproducible work.","headline":"Useful and reproducible simplification of Kato perturbation theory for Bose-Hubbard rings, but Eq. (1) misstates the hopping and needs a fix.","tokens_in":12745,"tokens_out":7072,"would_cite":false,"duration_ms":75199,"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":"Ground-state energies of Bose-Hubbard rings can be computed to ninth order in perturbation theory from a fixed set of 50 reduced diagonal matrix elements, using the system's one-unit spectral gap and reflection symmetry.","keywords":["perturbation theory","Bose-Hubbard model","Kato formalism","ground-state energy","many-body physics","convergence radius","ring lattice","diagonal matrix elements"],"falsifier":"Apply the Table 1 series to a Bose-Hubbard ring with M=4, N=12 at $|\\Omega/\\kappa|=0.01$, which lies inside the claimed radius $1/(4N)=0.0208$, and compare with exact diagonalization; agreement would support the claimed M-independence, while disagreement would falsify it.","tokens_in":11749,"feed_emoji":"⚛️","tokens_out":15114,"duration_ms":138641,"temperature":0.7,"pith_summary":"The standard perturbation series for the ground-state energy of a Bose-Hubbard ring normally explodes in length with each order, but this paper shows that up to ninth order it can be evaluated from only 50 reduced diagonal matrix elements. The simplification comes from a spectral fact: for repulsive interactions and uniform filling, every state that one tunneling event can reach from the ground state lies exactly one energy unit above it, so all the energy-denominator factors in the series collapse to the same simple factor. Reflection symmetry from the hermiticity of the Hamiltonian then lets every elementary building block be written with outer exponents equal to one, producing the universal weight tables in the paper. A convergence criterion $|\\Omega/\\kappa| \\leq 1/(4N)$ follows from Kato's bound, and explicit trimer calculations for $N=9$ and $N=27$ reproduce exact diagonalization within that radius. If this holds, the scheme turns a formally unmanageable calculation into a small, reusable set of system-specific matrix elements.","feed_headline":"One energy gap cuts ninth-order Bose-Hubbard series to 50 matrix elements","feed_subtitle":"Simplified Kato series reproduces exact diagonalization on trimers and should extend to other lattices.","key_machinery":"The load-bearing object is the reduction identity $\\hat S_k \\hat V |a\\rangle = (-1)^{k-1}\\hat S_1\\hat V |a\\rangle$ for $k\\geq 1$, which follows from the one-unit spectral gap and collapses every resolvent-projector power in the Kato series to the first power. An elementary matrix element (EME) is a diagonal matrix element of the perturbation series that contains no zero projector $\\hat S_0$; the identity, together with the reflection symmetry $(k_1,\\dots,k_m)=(k_m,\\dots,k_1)$ from hermiticity, implies rules (i)--(iv): any EME reduces to one with outer exponents $1$, up to the sign $(-1)^{k_1+k_m-2}$. These rules are what generate the weighted diagonal-matrix-element families in Table 1, so the computational work for any specific ring is just evaluating the few EME values and combining them with the universal weights.","core_discovery":"The central claim is that the Kato form of perturbation theory can be restructured for Bose-Hubbard rings using the unit spectral gap. Because the perturbation $\\hat V$ acting on the ground state $|a\\rangle$ produces only first excited states with energy $\\varepsilon_a + 1$, the identity $\\hat S_k \\hat V |a\\rangle = (-1)^{k-1} \\hat S_1 \\hat V |a\\rangle$ holds for all $k \\geq 1$, so arbitrary powers of the resolvent projectors never need to be evaluated separately. Combined with the reflection symmetry of elementary matrix elements, this yields the rules of Section 4.1: in any elementary matrix element the outer exponents become one, with a sign carried by the original exponents. Table 1 lists all weighted products of elementary matrix elements needed through order 9, and Table 2 shows that the scheme reduces the number of diagonal matrix elements from 490 to 50, with only 18 elementary matrix elements that have not appeared in lower orders. The same simplification is argued to work for any Bose-Hubbard-type system whose perturbation is a single tunneling event and whose gap is one unit, including hypercubic or triangular lattices. The trimer validation with $N=9,27$ shows agreement with exact diagonalization inside the estimated convergence radius $|\\Omega/\\kappa|\\leq 1/(4N)$.","pith_inferences":["The same reduction could in principle be iterated beyond order nine, since the gap condition fixes the outer exponents; a symbolic generator could produce weight tables for higher orders without repeating the full Kato expansion.","The paper's norm bound is deliberately crude, and the N=9 trimer data follow exact diagonalization outside $|\\Omega/\\kappa|\\leq 1/(4N)$, suggesting a sharper convergence radius might be derivable from the low-energy spectrum or a subspace-restricted norm.","The pathway interpretation in the Discussion points toward a graph-theoretic reading in which the Table 1 weights count closed walks on the ring; making that explicit could yield a transfer-matrix method for arbitrary order.","Because the simplification depends only on the spectral gap, it should carry over to other single-tunneling bosonic lattice models, such as ladders or hypercubic lattices, with only the elementary matrix element values changing; the paper states this extension but does not demonstrate it numerically."],"forward_implications":["An order-by-order calculation up to order 9 for any Bose-Hubbard ring requires evaluating only 50 reduced diagonal matrix elements, built from 46 elementary matrix elements of which 18 have not appeared in lower orders.","The same weighted terms in Table 1 apply unchanged to other lattice geometries with a single tunneling event and a one-unit spectral gap, so only the elementary matrix element values are system-specific.","The convergence radius $|\\Omega/\\kappa| \\leq 1/(4N)$ is independent of the number of lattice sites M, so adding more sites does not shrink the perturbative regime for fixed particle number.","Within the estimated convergence radius, the trimer calculations show that already sixth order, which requires only six elementary matrix elements, gives ground-state energies in agreement with exact diagonalization."],"supporting_citations":[{"why":"supplies Kato's perturbation method and the convergence estimate that becomes Eq. (15).","marker":"[18]"},{"why":"introduces the reduction of diagonal matrix elements into elementary matrix elements and the exclusion of vanishing first-order corrections, which the simplification rules build on.","marker":"[19]"},{"why":"gives the detailed Kato series for eigenvalues and the spectral-norm definition used in the convergence criterion.","marker":"[22]"},{"why":"provides the Bloch-basis spectrum of the tunneling operator, giving its operator norm 2N used in the convergence radius.","marker":"[12]"},{"why":"supplies the Python implementation of the Kato formalism used to generate and check the simplified series.","marker":"[25]"},{"why":"states that the spectral norm of a hermitian operator equals its largest absolute eigenvalue, justifying the value 2N.","marker":"[26]"},{"why":"provides the explicit trimer Hamiltonian matrix elements and index mapping used in the validation against exact diagonalization.","marker":"[29]"}],"fun_headline_variants":["One spectral gap slashes Bose-Hubbard perturbation to 50 elements","Gap-based trick reduces ninth-order Bose-Hubbard series tenfold","Perturbation theory simplified for rings via unit spectral gap","Easy higher-order Bose-Hubbard: From 490 to 50 matrix elements","Unit gap unlocks simple perturbation series for Bose-Hubbard"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire reduction relies on every state that a single tunneling event can create from the unperturbed ground state having exactly the same unperturbed energy, one unit above the ground state, which requires repulsive interactions and uniform filling with N a multiple of M.","fun_headline_variants_meta":{"raw":{"variants":["One spectral gap slashes Bose-Hubbard perturbation to 50 elements","Gap-based trick reduces ninth-order Bose-Hubbard series tenfold","Perturbation theory simplified for rings via unit spectral gap","Easy higher-order Bose-Hubbard: From 490 to 50 matrix elements","Unit gap unlocks simple perturbation series for Bose-Hubbard"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000386,"raw_usage":{"total_tokens":2042,"prompt_tokens":949,"completion_tokens":1093,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":565,"completion_tokens_details":{"reasoning_tokens":1000}},"tokens_in":565,"tokens_out":1093,"duration_ms":9064,"temperature":1.0,"reasoning_tokens":1000,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T11:10:59.267073+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Apply the Table 1 series to a Bose-Hubbard ring with M=4, N=12 at $|\\Omega/\\kappa|=0.01$, which lies inside the claimed radius $1/(4N)=0.0208$, and compare with exact diagonalization; agreement would support the claimed M-independence, while disagreement would falsify it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies Kato's perturbation method and the convergence estimate that becomes Eq. (15)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"introduces the reduction of diagonal matrix elements into elementary matrix elements and the exclusion of vanishing first-order corrections, which the simplification rules build on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the detailed Kato series for eigenvalues and the spectral-norm definition used in the convergence criterion."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the Bloch-basis spectrum of the tunneling operator, giving its operator norm 2N used in the convergence radius."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the Python implementation of the Kato formalism used to generate and check the simplified series."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"states that the spectral norm of a hermitian operator equals its largest absolute eigenvalue, justifying the value 2N."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the explicit trimer Hamiltonian matrix elements and index mapping used in the validation against exact diagonalization."}],"review_version":1}