{"id":"b327a8fc-847b-4628-b067-b6e6e4f63c20","arxiv_id":"2512.11190","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A multiloop functional renormalization group in the single-boson-exchange representation reproduces parquet-approximation results for the 2D Hubbard model within a few percent when multi-boson rest functions are neglected.","lead":"This paper combines two techniques for computing the behavior of interacting electrons on a lattice—single-boson exchange and multiloop functional renormalization—and tests a simplified version against a more complete benchmark. In the weakly interacting regime the simplified version matches the benchmark to within a few percent, which could make such calculations much cheaper.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The mixed-form-factor bubble truncation is shared by the SBE-approximation and the parquet benchmark, so the claimed few-percent agreement may reflect a common systematic error rather than intrinsic accuracy.","rationale":"The paper is a careful methodological study, and the central finding is credible within the stated weak-coupling parameter window. The derivation in Section II is systematic, and the comparison of flow schemes in Section IV.A honestly documents residual cutoff dependence and incomplete 30-loop convergence. The most load-bearing weakness is not the formalism itself but the numerical validation: the same mixed-form-factor bubble truncation is used in both the SBE approximation and the parquet benchmark, and the stated justification for the truncation is only valid at Q=0. Since the finite-doping results are dominated by incommensurate Q, the few-percent agreement could in principle be an artifact of a shared systematic error. The reader's verdict of CONDITIONAL already captures this risk, so I do not recommend changing the verdict; the concern is addressable by including mixed-bubble terms or by benchmarking against an independent parquet solver. Secondary limitations—no released code or error bars, and a narrow parameter window—reinforce but do not replace this central concern.","tokens_in":41800,"tokens_out":9743,"duration_ms":99606,"concrete_test":"Recompute the converged SBE-approximation and parquet flows at U=2.5, β=5 for both half filling and finite doping (t′=−0.2, n=0.41), retaining the mixed s/d form-factor bubble entries Π_{X,sd} and ˙Π_{X,sd} in the Appendix A equations, i.e., using the full 2×2 form-factor bubble matrix instead of setting off-diagonal entries to zero. If the relative difference between SBE approximation and parquet remains within the reported 3–5% at the incommensurate magnetic peak and in χ_dSC(0), the shared-truncation concern is refuted; if it grows, the central claim must be restated as accuracy relative to a truncated parquet approximation rather than the full parquet approximation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III.E sets Π_{X,nm} = ˙Π_{X,nm} = 0 for n≠m, with the justification that mixed-form-factor contributions 'anyway at q=0 vanish.' That justification is local to Q=0, but the flow is evaluated for all bosonic momenta Q, including the incommensurate magnetic peaks at finite doping shown in Figs. 14–15. Because the parquet benchmark in Eqs. (81)–(83) and Appendix A employs the same truncation, the reported ≤3–5% agreement demonstrates that dropping M_X is inexpensive within this truncated form-factor space, not that the SBE approximation reproduces the full parquet approximation. The subsequent remark—'Within this approximation, the contributions from the d-wave component to the Yukawa couplings vanish identically in the SBE approximation'—makes the risk concrete: at finite doping, where d-wave pairing is a competing channel in Section IV.C, the SBE approximation may be missing vertex structure that the full parquet approximation would generate. If mixed-bubble terms are non-negligible at finite Q, both the SBE approximation and the benchmark parquet results shift together, so the reported few-percent error would not transfer to a full-parquet comparison.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a multiloop extension of the single-boson-exchange (SBE) functional renormalization group (fRG) in physical channels and applies it to the two-dimensional Hubbard model at weak coupling. The central claim is that the \"SBE approximation\" — obtained by neglecting the flow of the multi-boson rest functions M_X in Eqs. (81c), (82c), (83c) — reproduces the parquet approximation at loop convergence. This is supported by direct numerical comparisons for U=2.5, β=5 at half filling and at finite doping (n=0.41), with reported differences of at most 5% in the magnetic susceptibility peak at half filling and at most 3% at finite doping. The paper also provides a detailed derivation of the flow equations, discusses cutoff dependence across three flow schemes, and presents temperature-dependent results at β=5,10,20.","tokens_in":42062,"tokens_out":5178,"duration_ms":52720,"significance":"If the central claim holds, the paper offers a substantial computational simplification: the expensive flow of the SBE rest functions can be omitted without sacrificing quantitative accuracy in the weak-coupling regime, enabling applications to more challenging parameter regimes and more realistic models. The manuscript's strengths are its self-contained derivation of the multiloop SBE equations in physical channels, the direct parameter-free comparison with a parquet solver rather than a fitted benchmark, the explicit treatment of loop and self-energy convergence, and the systematic cutoff-dependence analysis. The main caveat is that the parquet benchmark itself is computed within a truncated form-factor space with off-diagonal bubble contributions set to zero; the headline claim therefore needs to be qualified accordingly.","major_comments":[{"comment":"The mixed-form-factor bubble truncation Π_{X,nm}=Π̇_{X,nm}=0 for n≠m is justified by the statement that these contributions 'anyway at q=0 vanish,' but the flow is evaluated for all bosonic momenta Q, including the incommensurate magnetic peaks at finite doping shown in Figs. 14–15. Since the parquet benchmark is computed with the same truncation, the reported ≤3–5% agreement demonstrates that dropping M_X is accurate within this truncated form-factor space, not necessarily that the SBE approximation reproduces the full parquet approximation. The abstract and conclusion should be qualified to avoid overstating the claim. If feasible, a numerical test with mixed-bubble terms retained, or at least an estimate of their size at finite Q, would strengthen the conclusion.","section":"Section III.E and Appendix A (Eqs. A1–A3)"},{"comment":"The text explicitly states that 'none of the three flow schemes are actually fully converged at U=2.5, β=5, and half filling for ℓ=30.' Yet Figs. 11–13 compare 'converged' SBE fRG results with the parquet approximation, using up to 36 loop corrections per self-energy iteration. The paper should quantify the difference between the 30ℓ results and the converged results for the susceptibility peaks. Without this number, the reader cannot judge whether the reported 3–5% differences between the SBE approximation and parquet are contaminated by residual loop-convergence error.","section":"Section IV.A and Fig. 10"},{"comment":"The central claim that the SBE approximation 'accurately reproduces the parquet approximation at loop convergence' is benchmarked only at β=5 (for half filling and n=0.41) and for U up to 2.5 at β=5. The temperature-dependent results at β=10 and β=20 in Sec. IV.C are not compared with the parquet approximation. The abstract and conclusion should either restrict the validation statement to the computed parameter window or add parquet benchmarks at the lower temperatures shown in Figs. 15–17.","section":"Section IV.B–IV.C and Abstract"}],"minor_comments":[{"comment":"The phrase 'since that their contributions anyway at q=0 vanish' is grammatically awkward and should be rephrased, e.g., 'since these contributions vanish at Q=0.' Also clarify that the truncation is applied for all Q, not only Q=0.","section":"Section III.E"},{"comment":"Typo: 'hilf-filled' should be 'half-filled' in the text preceding Fig. 7.","section":"Section IV.A"},{"comment":"Typo: 'fom' should be 'from' in the caption.","section":"Fig. 11 caption"},{"comment":"The phrase 'rest function like contributions' should be hyphenated: 'rest-function-like contributions.'","section":"Section V"},{"comment":"The phrase 'enclose the reduced Brillouin zone' is unclear; consider 'enclosed by the Γ-X-M-Γ symmetry path.'","section":"Fig. 6 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid methodological contribution and the numerical implementation appears careful. However, the headline claim should be qualified because the parquet benchmark shares the mixed-form-factor bubble truncation. This is a fixable issue via revised wording and, ideally, a selective numerical test. I also recommend asking for a quantitative statement of loop convergence at the reported parameters, since the paper notes that ℓ=30 is not fully converged. The scope of the validation (β=5 only) should be made explicit."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is a solid piece of methodology: the authors write down the multiloop SBE fRG equations in physical (M/D/SC) channels for SU(2)-symmetric systems and implement them for the 2D Hubbard model at weak coupling. The new content is the physical-channel derivation and the numerical test of the SBE approximation, where the flow of the multi-boson rest functions M_X is dropped. Second, the central quantitative claim—that the SBE approximation reproduces the parquet approximation to within a few percent at loop convergence—is credible but narrower than it looks. The comparison is made between two implementations that share the same truncated form-factor space, including the neglect of mixed-form-factor bubble contributions. So the paper convincingly shows that within that implementation, dropping M_X is cheap; it is less decisive about how well the SBE approximation tracks the full parquet approximation.\n\nThe derivation in Sec. II.E is clearly structured and the equations are worth having in the literature. The numerical implementation is described in enough detail to be reproducible in principle, and the authors are honest about the residual cutoff dependence and the fact that the 30-loop results are not fully converged at U=2.5, β=5. The self-energy flow is exact, and the only truncations are the neglect of the 2PI flow (standard parquet) and the SBE approximation itself. No parameter is fitted to the parquet data.\n\nThe soft spot is the mixed-bubble truncation, set to zero in Sec. III.E and Appendix A. The stated justification is that these terms vanish at q=0, but they are dropped at all bosonic momenta, including the incommensurate wave vectors relevant at finite doping. Since the parquet benchmark runs with the same truncation, the few-percent agreement could in part be a shared systematic error. The authors even note that within this truncation the d-wave components of the Yukawa couplings vanish in the SBE approximation, which is a concrete worry at finite doping where d-wave pairing is a competing channel. This is a real limitation, but it is not a fatal flaw: it means the claim should be stated more carefully, and the benchmark should be checked against a parquet solver that keeps the mixed-bubble terms, or at least estimates their size at finite Q. The parameter window (one U, β=5–20, two fillings) is narrow; broader benchmarking would strengthen the case.\n\nWho this is for: anyone using multiloop fRG or SBE-based methods for lattice fermions. It deserves a serious referee. I would send it to review and ask for a revision that either quantifies the mixed-bubble effects or softens the claim to 'within the standard truncated-unity implementation.'","headline":"A careful, useful multiloop SBE fRG derivation and benchmark; the few-percent agreement with parquet is real but is partly a test of two approximations sharing a form-factor truncation.","tokens_in":42629,"tokens_out":2921,"would_cite":true,"duration_ms":28563,"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":"Neglecting multi-boson exchange contributions, the multiloop single-boson exchange fRG reproduces the parquet approximation to within a few percent in the 2D Hubbard model.","keywords":["multiloop functional renormalization group","single-boson exchange decomposition","Hubbard model","parquet approximation","Yukawa couplings","susceptibilities","rest functions","weak coupling"],"falsifier":"Compute the magnitude of the neglected mixed form-factor bubble contributions Π_{X,nm}(Q) at the finite incommensurate wave vector that dominates the magnetic susceptibility at finite doping, or repeat the calculation without the truncation and compare the SBE approximation to parquet; if the mixed-bubble terms are not negligible at these Q, the claimed ≤3% agreement is unlikely to survive.","tokens_in":41617,"feed_emoji":"🧲","tokens_out":5950,"duration_ms":52348,"temperature":0.7,"pith_summary":"The paper establishes that the single-boson exchange (SBE) approximation of the multiloop functional renormalization group, which drops the flow of the multi-boson exchange rest functions, reproduces the parquet approximation accurately once loop corrections are converged. For the two-dimensional Hubbard model at weak coupling (U=2.5, β=5), the converged SBE results differ from parquet by at most 5% in the magnetic susceptibility peak, and by ≤3% at finite doping. This matters because the rest functions are the most expensive objects to compute, so dropping them makes quantitative multiloop fRG substantially cheaper. The paper also provides a complete derivation of the multiloop SBE fRG equations in physical channels, and analyzes cutoff dependence, loop convergence, and temperature dependence.","feed_headline":"Dropping multi-boson terms keeps fRG accurate to 5%","feed_subtitle":"In the 2D Hubbard model, the cheap single-boson exchange approximation matches the full parquet solution within a few percent.","key_machinery":"The central machinery is the single-boson exchange (SBE) decomposition of the two-particle vertex, which rewrites the vertex in terms of bosonic propagators w_X(Q), Yukawa couplings λ_X(Q,k), and rest functions M_X(Q,k,k') in three physical channels (magnetic, density, superconducting). The multiloop fRG flow equations for these objects are derived from the parquet/Bethe-Salpeter structure and are solved with a loop expansion (ℓ=1,2,3,...) together with a self-energy flow equation for the self-energy. The SBE approximation consists of discarding the flow equations for the rest functions M_X, thereby removing the most expensive part of the computation.","core_discovery":"The central claim is that the SBE approximation—setting the flow of the multi-boson exchange rest functions M_X to zero while keeping the flow of bosonic propagators and Yukawa couplings—accurately reproduces the parquet approximation at loop convergence in the weak-coupling 2D Hubbard model. Quantitatively, the relative difference in the magnetic susceptibility peak is at most 5% at half filling and ≤3% at finite doping, with frequency- and momentum-dependent vertices showing similar accuracy. Since the converged multiloop SBE fRG with the rest functions included is equivalent to the parquet approximation, this shows that the computationally expensive rest-function flow can be discarded wit","pith_inferences":["Because the SBE approximation already matches parquet at weak coupling, one might expect it to become less accurate at intermediate to strong coupling, where multi-boson exchange processes (the dropped rest functions) are more important; the paper does not test this regime, and the claim should not be extrapolated there.","The numerical implementation sets all mixed form-factor bubble contributions to zero at all momenta, not just at q=0 where they vanish; since the parquet comparison uses the same truncation, the observed agreement may in part reflect a shared systematic error rather than the intrinsic accuracy of the SBE approximation.","A testable extension would be to include the mixed-bubble terms selectively at the dominant finite-momentum wave vectors (or to compute their magnitude) to see whether the few-percent agreement persists; if it does, the truncation is innocent; if not, the reported accuracy is partly accidental.","Since the SBE approximation is a differential-equation-based resummation that generates rest-function-like contributions implicitly, it may be the basis for a cheaper parquet solver for models where the full parquet solution is numerically out of reach."],"forward_implications":["Multiloop fRG calculations for correlated electron systems become substantially cheaper, since the rest functions M_X, the most costly objects, no longer need to be flowed.","The SBE approximation provides a physically transparent picture in which the dominant fluctuations are carried by single-boson exchange processes, with multi-boson effects implicitly resummed through the flow.","The method opens the route to more challenging parameter regimes and more realistic models, where full parquet-equivalent multiloop fRG would be prohibitively expensive.","The residual cutoff dependence of the SBE approximation is small at loop convergence (few percent), and the Ω-flow converges fastest among the tested schemes.","The approach correctly captures the interplay of magnetic and d-wave superconducting fluctuations as temperature is lowered."],"fun_headline_variants":["Skip multi-boson flow, match parquet to 5%","Cheap SBE fRG hits parquet accuracy within 5%","Dropping multi-boson terms: fRG still accurate to 5%","Multiloop fRG: ignore rest functions, keep accuracy","Hubbard model: SBE fRG matches full parquet result"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The numerical comparison relies on setting all mixed form-factor bubble contributions to zero at every momentum, although the stated justification only guarantees they vanish at q=0; if those terms are sizable at the finite momenta that dominate the response, the claimed few-percent agreement with parquet could degrade.","fun_headline_variants_meta":{"raw":{"variants":["Skip multi-boson flow, match parquet to 5%","Cheap SBE fRG hits parquet accuracy within 5%","Dropping multi-boson terms: fRG still accurate to 5%","Multiloop fRG: ignore rest functions, keep accuracy","Hubbard model: SBE fRG matches full parquet result"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000229,"raw_usage":{"total_tokens":1306,"prompt_tokens":722,"completion_tokens":584,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":466,"completion_tokens_details":{"reasoning_tokens":487}},"tokens_in":466,"tokens_out":584,"duration_ms":5732,"temperature":1.0,"reasoning_tokens":487,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T16:55:21.729324+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the magnitude of the neglected mixed form-factor bubble contributions Π_{X,nm}(Q) at the finite incommensurate wave vector that dominates the magnetic susceptibility at finite doping, or repeat the calculation without the truncation and compare the SBE approximation to parquet; if the mixed-bubble terms are not negligible at these Q, the claimed ≤3% agreement is unlikely to survive.","supporting_citations":[],"review_version":1}