{"id":"8da8eac0-f806-44ad-b0f8-727cfccdb93f","arxiv_id":"2501.13618","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Wavelet-based flow equations drive free and quadratically coupled scalar field Hamiltonians toward block diagonal form by resolution, and the coarse-resolution block reproduces the low-lying normal mode frequencies of the truncated theory.","lead":"This paper uses a wavelet basis and a similarity renormalization group flow to separate length scales in simple 1+1 dimensional scalar field theories, and shows the coarse-scale Hamiltonian reproduces the low-energy spectrum. It extends a prior calculation by one resolution and adds a two-field quadratic interaction as a test case.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Load-bearing concern: the flow is implemented as a matrix flow on Ω² only, relying on the unproven no-φπ-mixing assumption delegated to [1]; if this fails, the effective Hamiltonian misses couplings and the quoted normal-mode frequencies do not represent the full truncated theory.","rationale":"The reader's CONDITIONAL verdict is well-founded. The manuscript's central claim rests on the assumption that the flow can be reduced to a matrix flow on Ω² with a common orthogonal rotation of fields and momenta and no φπ mixing. This assumption is explicitly acknowledged in Sec. VIII, and the justification is delegated to [1] without a derivation in the present paper. Since the paper extends the method to a new two-field model and to a higher resolution, the transfer of the earlier result is an unverified step. The proposed check—integrating the full Hamiltonian flow—would settle whether the assumption holds. Independent of that concern, the numerical demonstration is internally consistent under the stated algorithm, so there is no reason to reject; the result should remain conditional pending the check.","tokens_in":25898,"tokens_out":2301,"duration_ms":20231,"concrete_test":"Implement the full second-order Hamiltonian flow, including φ, ψ, πφ, πψ, with generator K(λ) = [G(λ), H(λ)] for the two-field model truncated at resolution 1 (and/or the free-field resolution-2 case). Check whether any φπ or π-only quadratic terms are generated at any λ. If they are, compare the low-lying eigenvalues of the full H(λ) at λ = 20 with the eigenvalues from the Ω²_ss block to determine whether the quoted frequencies represent the full theory. If no φπ terms appear, the assumption is validated and the numerical claims stand.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim depends on the flow remaining in the class H(λ) = (1/2)(πᵀπ + φᵀΩ²(λ)φ), with the unitary implemented as an identical orthogonal rotation O(λ) on φ and π and no φπ cross terms generated. But this is assumed, not derived, in Sec. IV, Eqs. (74)-(79), with the justification 'based on the results of [1]'; the assumption is acknowledged in Sec. VIII and delegated to [1]. The paper extends to a new two-field model and to resolution 2 for the free field, yet the transfer of the [1] result to these new settings is asserted without demonstration. All numerical evidence—Tables III and V, and the HS norms—is computed from the matrix flow dΩ²/dλ = [K,Ω²] with K = [G,Ω²], so if the assumption fails, the eigenvalues reported from the Ω²_ss block may not correspond to the true effective Hamiltonian of the full truncated theory. Since the manuscript never checks for generated φπ or π-only terms, the central claim is not secured by the paper alone.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript applies similarity renormalization group (SRG) flow equations to scalar field theories represented in a Daubechies wavelet basis. It first extends the single free-field analysis of Michlin and Polyzou to one higher resolution (k=2), presenting numerical evidence that the double-commutator flow drives the coefficient matrix Ω² to a block-diagonal form with blocks labelled by resolution, and that the eigenvalues of the coarsest block converge to the eigenvalues of the full truncated matrix as the flow parameter λ increases. It then introduces a model of two real scalar fields coupled by a mass-mixing term, formulates it in the wavelet basis, and applies the same matrix flow. Tables and figures show the off-diagonal Hilbert-Schmidt norms decaying and the coarse-block eigenvalues approaching the exact truncated eigenvalues. The central claim is that the specifically chosen generator produces a block-diagonal effective Hamiltonian whose coarsest-resolution block reproduces the normal-mode frequencies of the truncated model.","tokens_in":26158,"tokens_out":12176,"duration_ms":108399,"significance":"If the central claim holds, the work provides a concrete multiresolution truncation scheme for low-energy effective Hamiltonians in wavelet-based field theory, extending the prior single-field result to one higher resolution and to a two-field model. The numerical evidence is internally consistent: the matrix-flow plots, Hilbert-Schmidt norms, and eigenvalue tables behave as expected, and the exact normal-mode frequencies are obtained from closed-form expressions that are independent of the flow. The eigenvalue comparisons are clear and the HS-norm diagnostics are appropriate. The main caveat is that the physical interpretation of the Ω²_ss block as an effective Hamiltonian rests on an assumption about the allowed form of the unitary flow that is asserted but not proven in this manuscript. The work is an incremental but useful step; its significance would be substantially strengthened by a proof or explicit numerical verification of that assumption.","major_comments":[{"comment":"The reduction of the full Hamiltonian flow to a matrix flow on Ω² relies on the assumption that the unitary transformation acts as an identical orthogonal rotation on the fields and their conjugate momenta, so that no φπ mixing terms or changes to the momentum terms are generated. This is stated as an assumption and justified only by reference to [1], which treats a different system (a single free field at lower resolution). Since the two-field model and the resolution-2 extension are new to this paper, the transfer of the [1] result to these settings is not automatic. Because the central claim — that the Ω²_ss block after the flow is the effective Hamiltonian of the coarsest-resolution sector — depends on this assumption, the claim is not secured by the manuscript alone. Please provide a proof that the chosen generator leaves the class H(λ) = (1/2)(πᵀπ + φᵀΩ²(λ)φ) invariant, or alternatively add a numerical diagnostic for the truncated models that directly computes the φπ and π-only components of the flowed Hamiltonian and shows they vanish or are negligible.","section":"Sec. IV, Eqs. (74)-(79); Sec. VIII"},{"comment":"The paper states that the flow 'drives the Hamiltonian into a block diagonal form' and that the effective Hamiltonian 'correctly reproduces the normal mode frequencies.' As presented, the numerical evidence demonstrates block-diagonalization of the matrix Ω² and the preservation of its spectrum by the double-commutator flow. Since the spectrum of Ω² is unitarily invariant under this matrix flow, the agreement between the Ω²_ss eigenvalues and the full-matrix eigenvalues is primarily a consistency check of the numerical integration of the flow, not a falsifiable physical prediction. The stronger physical claim — that the coarse-block eigenvalues are the actual low-lying normal-mode frequencies of the full truncated field theory — requires the no-φπ-mixing assumption discussed above. The text should state this distinction explicitly and temper the abstract and conclusion accordingly.","section":"Abstract; Sec. VII; Tables III and V"}],"minor_comments":[{"comment":"The statement 'There is no coupling between oscillators across locations and resolutions' is contradicted by the banded structure of the kinetic matrices D^k_ss,mn and D^kq_sw,mn, which couple neighbouring locations, and by the presence of the H_sw cross-resolution blocks. Please rephrase to describe the actual coupling structure, e.g., 'the interaction term couples only same-location, same-resolution modes, while the kinetic terms have limited inter-location couplings within each resolution.'","section":"Abstract; Sec. II, VI"},{"comment":"The exponential solutions as written are incorrect: the right-hand sides should involve H_bmn(0) and H_cmn(0), not H_bmn(λ) and H_cmn(λ), and the derivation assumes that the eigenbases of H_c and H_b are λ-independent, which is generally false. Since the numerical method in Sec. IV uses the full commutator flow, this pedagogical section would benefit from a clarifying remark that Eqs. (45)-(46) are only schematic.","section":"Sec. III, Eqs. (45)-(46)"},{"comment":"The last term in Eq. (24) appears to have a typo: it should likely be G^t_nm w^{k-1}_m rather than w^{k-1}_n, to match the index structure of the preceding terms.","section":"Sec. II, Eq. (24)"},{"comment":"The caption of Fig. 3 lists 's0_1(x)' twice; the third curve is presumably s0_2(x) or another resolution-0 scaling function.","section":"Sec. II, Fig. 3"},{"comment":"Several summation indices are inconsistent; for example, in H_ss the term X_n ϕs,k_m ϕs,k_n D^k_ss,mn should be summed over m and n, and similar index fixes are needed in Eqs. (59) and (60).","section":"Sec. IV, Eqs. (58)-(60)"}],"recommendation":"major_revision","confidential_remarks":"The key structural assumption is delegated to [1] without examining whether it transfers to the new two-field model and the higher-resolution free-field case; I would urge the editor to require the authors to address this directly, either by a proof or by a numerical check of the generated φπ terms. The paper also does not specify the numerical integrator or tolerances used for the flow equations, which weakens reproducibility. The eigenvalue comparisons are clearly presented but are essentially a unitarity consistency check of the matrix flow; the framing should be adjusted so that the physical claim rests on the effective-Hamiltonian interpretation rather than on the spectrum-preservation property."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nRead the Basak–Ratabole paper. My take: it is a genuine but modest increment over Michlin–Polyzou, and the numerical story it tells is internally consistent. The free-field flow at resolution k=2 and the two-field quadratic model are real extensions of the formalism, not repackaged claims. Tables III and V show the coarse-block eigenvalues converging to the exact truncated eigenvalues as lambda grows, and the Hilbert–Schmidt norms decay as expected. For exactly solvable quadratic theories, spectrum preservation under a unitary flow is mathematically guaranteed, so the eigenvalue tables are really checking the numerical integration rather than testing new physics. That is fine, and the paper mostly says so.\n\nWhere I part company with the reader's softer framing is the no-phi-pi-mixing assumption. It is not a minor technicality. The whole flow is implemented as a matrix flow on Omega^2, with the Hamiltonian assumed to stay of the form (1/2)(pi^T pi + phi^T Omega^2(lambda) phi). The paper explicitly delegates the justification to [1] in Sec. VIII, and in Sec. IV just asserts it. The two-field model is new, and the resolution-k=2 free-field calculation goes beyond [1], yet no new check is provided that the flow does not generate terms coupling phi and pi or renormalize the momentum term. If that assumption fails, the Omega^2_ss eigenvalues quoted as the effective low-resolution spectrum do not represent the full truncated Hamiltonian. The reader's concern lands, and it is load-bearing. The fixed-point of the flow may still be block-diagonal, but the paper has not shown that diagonalizing only Omega^2_ss gives the true effective Hamiltonian.\n\nOther soft spots are smaller but real. Daubechies order K is never specified for the numerical runs; flow-integration details (step size, solver) are absent; code and data are not provided. Those are reproducibility issues the paper could fix. The models are quadratic, so no new physics is claimed, and I do not think the lack of a nonlinear interaction is itself a flaw—this is a benchmark paper.\n\nWho is this for? People working on wavelet-based Hamiltonian truncation and SRG implementations, especially the Polyzou program. It deserves a serious referee: the method is clearly described, the extension is honest, and the central open question—whether the momentum-sector assumption transfers to new settings—is exactly the kind of thing a referee can push on. I would not desk-reject it. I would also want the authors to either prove the no-mixing property for their generator or at least numerically check for generated phi-pi terms before publication.\n\nRecommendation: send to peer review, conditional on the authors addressing the assumption or explicitly weakening the claim.","headline":"A careful, honest extension of the wavelet-SRG program to one higher resolution and to a two-field quadratic model; the numerics are internally consistent, but the load-bearing no-momentum-mixing assumption is delegated to [1] rather than checked here.","tokens_in":26671,"tokens_out":698,"would_cite":false,"duration_ms":8574,"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":"With the right generator, wavelet-basis flow equations drive a truncated quantum Hamiltonian into blocks labelled by resolution, and the coarsest block reproduces the low-lying normal mode frequencies.","keywords":["Hamiltonian flow equations","Similarity Renormalization Group","Daubechies wavelet basis","scalar field theory","normal mode frequencies","resolution truncation","quadratic interaction","block diagonalization"],"falsifier":"Track the full unitary flow on both $\\phi$ and $\\pi$ operators for the two-field model and compute the size of any generated $\\phi\\pi$ terms; if they are not negligible compared with the reported eigenvalue accuracy, then diagonalizing only the $\\Omega^2_{ss}$ block misses part of the effective low-resolution dynamics.","tokens_in":25655,"feed_emoji":"⚛️","tokens_out":8237,"duration_ms":65540,"temperature":0.7,"pith_summary":"This paper tries to establish that the flow equations of the similarity renormalization group, implemented in a Daubechies wavelet basis, separate the length scales of a quantum field theory without losing the physics at any scale. The demonstration covers a free real scalar field in 1+1 dimensions at one resolution higher than an earlier study, and a model of two scalar fields coupled by a generally quadratic mass-mixing interaction. In both cases the chosen generator flows the truncated Hamiltonian into a block diagonal form, with each diagonal block belonging to a fixed resolution. The effective Hamiltonian of the coarsest block is shown to reproduce the low-lying normal mode frequencies of the full truncated theory, with the agreement improving as the flow parameter grows.","feed_headline":"Wavelet flow splits field theory into clean resolution blocks","feed_subtitle":"Diagonal blocks are labelled by resolution, and the coarsest one reproduces the low-lying normal mode spectrum.","key_machinery":"The machinery is the flow equation $d\\Omega^2(\\lambda)/d\\lambda = [K(\\lambda),\\Omega^2(\\lambda)]$ with the anti-Hermitian generator $K(\\lambda)=[G(\\lambda),\\Omega^2(\\lambda)]$, where $G(\\lambda)$ is the block-diagonal part of $\\Omega^2(\\lambda)$ containing only same-resolution couplings. This choice makes cross-resolution matrix elements decay exponentially in the flow parameter while the diagonal blocks are renormalized, turning the coarsest block into an effective Hamiltonian that carries the high-resolution effects. The underlying basis is the Daubechies wavelet basis, an orthonormal, compactly supported family in which each function is labelled by a location and a resolution index; the entries of $\\Omega^2$ are overlap integrals of derivatives of these functions.","core_discovery":"The central claim is that the resolution-truncated coupling matrix $\\Omega^2(\\lambda)$ evolves under the flow so that the off-diagonal blocks representing couplings between different resolutions decay to zero, while the diagonal blocks saturate to nontrivial matrices that encode the influence of the eliminated sectors. For the interacting two-field model, the interaction couples the two fields only within the same resolution, so the flow preserves those same-resolution couplings while suppressing cross-resolution couplings. As a result, diagonalizing just the coarsest-resolution block $\\Omega^2_{ss}(\\lambda)$ yields normal mode frequencies that converge to the exact eigenvalues of the truncated theory; for the lowest eigenvalue the discrepancy falls from about $10^{-3}$ at $\\lambda=0$ to $10^{-12}$ at $\\lambda=20$.","pith_inferences":["We infer that the same generator would need to be tested on anharmonic interactions, where the Hamiltonian gains higher-order terms; whether the resolution-block diagonal form survives beyond quadratic couplings is left open by the paper.","We infer that the band widening observed in the effective $\\Omega^2_{ss}$ block is the mechanism by which high-resolution physics is encoded locally into the coarsest sector, a feature that could be probed by tracking how the band width grows with $\\lambda$.","We infer that a direct check of the no-mixing assumption, by evolving both fields and momenta under the full unitary rather than only the coupling matrix, would settle whether the quoted normal-mode frequencies exhaust the effective low-resolution dynamics."],"forward_implications":["The flow provides a concrete numerical route from a resolution-truncated Hamiltonian to an effective low-resolution Hamiltonian whose spectrum tracks the full truncated spectrum to high precision.","The quadratic interaction between the two fields survives inside each resolution block after the flow, so scale separation does not erase the coupling that defines the model.","The free-field calculation at one resolution higher than the earlier study shows that the block-diagonal structure persists when additional resolution sectors are included.","If the no-mixing assumption holds, the same generator can be applied to larger truncations while avoiding the cost of evolving the momentum sector along with the configuration sector."],"supporting_citations":[{"why":"It supplies the earlier wavelet-basis flow-equation construction at resolution 1 and the justification that no field-momentum mixing is induced.","marker":"[1]"},{"why":"It introduces the Hamiltonian flow-equation method that the paper uses to evolve the coupling matrix.","marker":"[6]"},{"why":"It provides the wavelet field theory construction and the analytic evaluation of the derivative overlap integrals that define the entries of $\\Omega^2$.","marker":"[42]"},{"why":"It gives the representation of operators in compactly supported wavelets used to compute the derivative matrices.","marker":"[62]"},{"why":"It introduces the Daubechies wavelets, the orthonormal compactly supported basis family on which the whole calculation rests.","marker":"[51]"},{"why":"It supplies the filter-coefficient and refinement identities used to construct the scaling and wavelet functions.","marker":"[52]"}],"fun_headline_variants":["Wavelet flow splits theory into resolution blocks","Flow equations isolate low-energy modes in wavelets","Daubechies SRG suppresses cross-scale coupling","Coarse wavelet block matches exact eigenvalues","Resolution-block flow yields precise normal modes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the unitary flow rotates fields and momenta by the same orthogonal matrix, so the Hamiltonian keeps the form of a kinetic term plus a quadratic potential and no $\\phi\\pi$ mixing terms are generated.","fun_headline_variants_meta":{"raw":{"variants":["Wavelet flow splits theory into resolution blocks","Flow equations isolate low-energy modes in wavelets","Daubechies SRG suppresses cross-scale coupling","Coarse wavelet block matches exact eigenvalues","Resolution-block flow yields precise normal modes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000224,"raw_usage":{"total_tokens":1445,"prompt_tokens":911,"completion_tokens":534,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":527,"completion_tokens_details":{"reasoning_tokens":467}},"tokens_in":527,"tokens_out":534,"duration_ms":5625,"temperature":1.0,"reasoning_tokens":467,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T15:46:39.636277+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Track the full unitary flow on both $\\phi$ and $\\pi$ operators for the two-field model and compute the size of any generated $\\phi\\pi$ terms; if they are not negligible compared with the reported eigenvalue accuracy, then diagonalizing only the $\\Omega^2_{ss}$ block misses part of the effective low-resolution dynamics.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It supplies the earlier wavelet-basis flow-equation construction at resolution 1 and the justification that no field-momentum mixing is induced."},{"cited_title":"Flow-equations for hamiltonians.Annalen der Physik, 506(2):77–91, 1994","cited_arxiv_id":null,"evidence_quote":"It introduces the Hamiltonian flow-equation method that the paper uses to evolve the coupling matrix."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"It gives the representation of operators in compactly supported wavelets used to compute the derivative matrices."},{"cited_title":"Orthonormal bases of compactly supported wavelets","cited_arxiv_id":null,"evidence_quote":"It introduces the Daubechies wavelets, the orthonormal compactly supported basis family on which the whole calculation rests."}],"review_version":1}