{"id":"6113c673-7e68-4c7b-b72b-00e77a1eeb32","arxiv_id":"2608.04497","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Monte Carlo reconstruction of the 2D phi^4 two-particle vertex shows a multidimensional soft sector, a near-contact fully irreducible vertex, and validates local DGammaA only away from criticality.","lead":"Using Monte Carlo data for the two-dimensional phi^4 lattice model, this paper reconstructs the two-particle vertex and follows it through the Ising transition. It gives diagrammatic many-body theories a numerical benchmark, showing where a local-contact approximation works and where it fails.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The contact-only Lambda claim rests on treating Lambda(r1)/Lambda(r0)=0.0250±0.0064 at beta=0.68 as noise, and the cited 0.1% self-energy agreement is a k-averaged test that is blind to exactly such a tail.","rationale":"The reader identified the same load-bearing soft spot: the 4-sigma nearest-neighbor weight in Lambda is interpreted as zero, although the paper's own Appendix F warns that Lambda is a difference of large terms and that only the r=0 amplitude is clean. My concern sharpens this by noting that the headline 0.1% self-energy reproduction is k-averaged and therefore insensitive to a zero-momentum-sum nearest-neighbor tail: such a tail is invisible to lambda_eff and to the local self-energy by construction. The paper is transparent about the limitations and does not claim exactness; the DGammaA comparison in Figures 4 and 5 is a legitimate benchmark and would likely survive even if the tail is real. However, because the conclusion that Lambda is a contact is what upgrades the reconstruction into a DGammaA benchmark, the tail deserves a direct sensitivity test rather than a sigma-based dismissal. The proposed tail-injection test settles this: if the k-resolved self-energy moves by less than the statistical errors, the contact approximation is quantitatively supported; if it moves more, the correct claim is that Lambda is nearly local for local observables, and the benchmark should be qualified accordingly. Same-data consistency remains a secondary limitation: the 0.1% agreement is not an independent prediction, but the comparison with the local impurity vertex is a fair test, so the conditional verdict is appropriate and no verdict change is needed.","tokens_in":20224,"tokens_out":11149,"duration_ms":103261,"concrete_test":"At beta=0.64, L=16, where Lambda(r1)/Lambda(r0)=0.0150±0.0056 and the fixed-G parquet sweep converges, solve the parquet equations at the Monte Carlo G twice: once with the contact lambda_eff and once with lambda_eff plus a nearest-neighbor relative-coordinate tail normalized so that |Lambda_tail(r1)|/|Lambda_contact(r0)| equals the measured 1.5%, and also its 1-sigma upper bound. Recompute the k-resolved sunset self-energy via Eq. (D4) and compare each momentum component with the Monte Carlo self-energy and its bootstrap errors. If the local self-energy stays within 0.1% but individual k components deviate beyond errors, the contact truncation is validated only for the local self-energy, not for the benchmark as claimed. If all k components remain within errors, the tail is demonstrably irrelevant and the concern is settled.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central benchmark claim is that the fully irreducible vertex Lambda is a local contact and that replacing it by a constant lambda_eff reproduces the Monte Carlo self-energy to better than 0.1%. The load-bearing step is Section V's treatment of the L=16 beta=0.68 ratio Lambda(r1)/Lambda(r0)=0.0250±0.0064 as a \"pure contact term within 4 sigma\". A 4-sigma deviation is evidence against an exact contact, not within one; the text itself concedes that a short-ranged low-amplitude contribution may develop at larger L. The supporting self-energy test is less discriminating than stated: the 0.1% figure is for the k-averaged (local) self-energy, and lambda_eff is the Brillouin-zone sum of Lambda, so any nearest-neighbor tail with zero momentum sum drops out of lambda_eff and cannot affect the local self-energy. The k-resolved 4e-4 agreement quoted in Appendix D is obtained with the measured full F, not with the contact Lambda. Thus the reported numbers do not actually test whether the residual tail is negligible for the DGammaA benchmark; the tail is labeled noise on statistical grounds rather than shown to be harmless.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper uses Monte Carlo measurements of the connected two-particle correlator of the two-dimensional lattice phi^4 model to reconstruct the two-particle-irreducible vertex Gamma(k,p;q) by Bethe-Salpeter inversion. The vertex is decomposed into C_4v irreducible representations, its leading eigenvalues are tracked across the Ising transition, and the fully irreducible vertex Lambda is obtained by stripping crossed-channel ladders via the parquet equation. The central claims are that Lambda is a local contact, that inserting its effective value lambda_eff into the parquet and Schwinger-Dyson equations reproduces the Monte Carlo self-energy to better than 0.1%, and that this provides a benchmark for DGammaA. The paper also analyzes the stability of the parquet self-consistency in the critical region, identifying a repulsive fixed point and a single soft A_1 mode at the onset of the convergence wall.","tokens_in":20523,"tokens_out":6409,"duration_ms":56458,"significance":"The paper is valuable as one of the first direct, channel-resolved reconstructions of the 2PI vertex and the fully irreducible vertex for a lattice phi^4 theory, and it provides a concrete numerical comparison with DGammaA. The appendices document the technical machinery--null-space projection, threshold selection, bootstrap errors, and finite-size checks--in unusual detail, and the data/code availability statement is a strength. If the contact claim is established, the result would be a useful benchmark for parquet, fRG, and DGammaA communities. The DGammaA comparison via an independent single-site impurity calculation is a genuine strength of the paper.","major_comments":[{"comment":"The statement that Lambda(r1)/Lambda(r0)=0.0250 +/- 0.0064 at L=16, beta=0.68 is a \"pure contact term within 4 sigma\" is statistically misleading. The measured ratio is approximately 3.9 sigma from zero, so the data are evidence for a small nearest-neighbor tail, not evidence for an exact contact. The same applies to the beta=0.64 value, which is about 2.7 sigma from zero. The paper should either weaken the claim to \"consistent with a contact at the present resolution\" or provide a quantitative bound on the tail's contribution to the observables used in the benchmark.","section":"Section V, Fig. 3 inset"},{"comment":"The reported 0.1% agreement of the contact parquet closure with the Monte Carlo self-energy does not test the contact-only assumption for k-resolved quantities. The effective coupling lambda_eff = N^{-2} sum_{Q,k,p} Lambda(k,p;Q) is a Brillouin-zone average; any r != 0 Fourier component of Lambda sums to zero in this average, so lambda_eff contains only the on-site part of Lambda. The local self-energy <Sigma(k)>_k is consequently insensitive to a tail with zero total momentum sum. The k-resolved agreement of 4 x 10^{-4} quoted in Appendix D is obtained with the measured full F, not with the contact Lambda. Thus the numerical evidence does not establish that a residual short-range tail is harmless for the parquet closure; it establishes only that the on-site part of Lambda is sufficient for the local self-energy.","section":"Section V, Appendix D, Eq. (D4)"},{"comment":"The paper's own Appendix F states that lambda_eff is orthogonal to the noisy tail and that extracting a range xi_Lambda from the tail is not meaningful on current statistics. This is an important limitation that should be stated in the main text when the \"pure contact\" conclusion is drawn. Given this, the abstract's assertion that the vertex is \"a local contact\" overstates the evidence; the supported statement is that the fully irreducible vertex is contact-dominated at the achieved resolution.","section":"Section V, Appendix F"},{"comment":"The DGammaA comparison in Figs. 4-6 is a genuine independent test, and the agreement of Lambda_DMFT with lambda_eff away from criticality is a positive result. However, the \"first-principles benchmark\" language is too strong for the 0.1% self-energy claim because that claim uses lambda_eff and G from the same Monte Carlo dataset; it is a same-data closure check rather than an out-of-sample prediction. The paper should distinguish the independent DGammaA benchmark from the internal consistency check and rephrase the abstract accordingly.","section":"Abstract and Section V"}],"minor_comments":[{"comment":"The caption text \"Gamma X M\" should read \"Gamma, X, and M\" for clarity.","section":"Fig. 2 caption"},{"comment":"The ratios Lambda(r1)/Lambda(r0) are reported with \"within 1 sigma/3 sigma/4 sigma\" phrasing; consider reporting p-values or confidence intervals instead, since these statements are not standard evidence for a zero value.","section":"Section V"},{"comment":"The notation for the two particle-hole channels, Gamma_ph and Gamma_ph with an overbar, is easy to confuse; using an explicit overline or a different symbol consistently would improve readability.","section":"Appendix D"},{"comment":"The sentence \"It is worth noting that the term Q=0 predominantly contributes to this summation; however, it exceeds lambda_eff by an unacceptable 2-3% when beta >= 0.6\" is unclear and should be expanded or moved to an appendix.","section":"Section V"},{"comment":"The Fig. 3 inset is described in the text for L=8, while the nearby discussion in Section V reports L=16 ratios; please clarify which dataset is shown in the inset.","section":"Fig. 3 and Section V"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of cond-mat.stat-mech and the central numerical work appears sound. The main barrier is the overstatement of the contact nature of Lambda and the benchmark nature of the self-energy comparison; these issues are fixable in revision. I do not see grounds for rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does something genuinely new: it reconstructs the full two-particle-irreducible vertex for the 2D lattice phi^4 model from Monte Carlo data, resolves it in C4v channels, follows it across the Ising transition, and strips the crossed-channel ladders to extract the fully irreducible vertex via the parquet equation. That extraction, and the resulting DGammaA benchmark, are the contributions. The appendices are unusually careful about inversion details, null-space projection, conditioning, and bootstrap errors. The single-site DMFT comparison is independent and fair. Credit where it's due: this is the first full channel- and transfer-resolved map for a standard lattice model, and it will be a useful reference object for parquet and fRG practitioners.\n\nThe soft spots are real but narrower than the abstract suggests. Section V reports Lambda(r1)/Lambda(r0) = 0.0250 ± 0.0064 at L=16, beta=0.68, and calls it a pure contact within 4 sigma. That is statistically backwards: a 4-sigma deviation is evidence against an exact contact, not within one. The text honestly concedes a short-ranged low-amplitude tail may develop at larger L, but the abstract's \"local contact to a very good approximation\" overstates what the data show. More importantly, the 0.1% self-energy agreement is a k-averaged (local) quantity, and lambda_eff is the Brillouin-zone sum of Lambda, so any tail with zero momentum sum cancels out of that test. The k-resolved 4e-4 agreement in Appendix D uses the measured full F, not the contact Lambda. So the benchmark claim does not actually demonstrate that the residual tail is negligible; it labels the tail as noise on statistical grounds rather than showing it is harmless.\n\nThose are framing problems, not fatal ones. The core reconstruction, the multidimensional soft sector (A1+B1+B2), the collapse of the q=0 eigenvalue in the ordered phase, and the power-law tail at criticality are new and appear solid. The paper deserves a serious referee. I would send it to review, and I would ask the author to fix the statistical language, make the data and code available, and either demonstrate the tail's irrelevance for nonlocal self-energy or soften the contact-only claim accordingly.\n\nWho gets value: diagrammatic many-body theorists, parquet/DGammaA people, and anyone benchmarking two-particle vertices. I'd bring it to reading group, and would cite it if I worked on vertex-based methods.","headline":"A serious, well-documented Monte Carlo reconstruction of the 2PI vertex for 2D phi^4 gives the first full channel-resolved map and a sensible local-contact picture, but the 'pure contact' claim overreaches a 4-sigma tail and the 0.1% self-energy test is partly blind to it.","tokens_in":21024,"tokens_out":2022,"would_cite":true,"duration_ms":19657,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["05.50.+q","05.10.Ln"],"model":"deepseek-v4-flash","headline":"By stripping the crossed-channel ladders from Monte Carlo vertex data, the fully irreducible vertex of the two-dimensional phi^4 lattice model reduces to a local contact that reproduces the exact self-energy to within one-tenth of a…","keywords":["two-particle-irreducible vertex","phi^4 lattice field theory","Ising transition","Bethe-Salpeter equation","parquet formalism","DGammaA","Monte Carlo","emergent rotational symmetry"],"falsifier":"Run higher-statistics Monte Carlo at $L=32$ or $L=64$ near $\\beta\\simeq0.685$ and measure $\\Lambda(r_1)/\\Lambda(r_0)$: if the ratio settles at a nonzero value of order 2–3 percent rather than falling toward zero, the pure-contact picture is wrong. A complementary check is to include a short-range tail of that size in the parquet closure and test whether the resulting self-energy shifts by more than the claimed one-tenth of a percent.","tokens_in":20012,"feed_emoji":"🧲","tokens_out":17701,"duration_ms":124108,"temperature":0.7,"pith_summary":"The paper sets out to establish that the two-particle-irreducible (2PI) vertex of the two-dimensional $\\phi^4$ lattice model can be reconstructed from Monte Carlo measurements of the connected two-particle correlator, and that once the crossed-channel ladders are stripped away by the parquet equation, the fully irreducible vertex is a local contact to a very good approximation. This matters because it provides a first-principles benchmark for diagrammatic many-body theories: with that contact inserted into the parquet and Schwinger–Dyson equations, the Monte Carlo self-energy is reproduced to better than one-tenth of a percent, and the dynamical vertex approximation D$\\Gamma$A is shown to be nearly exact for local quantities in the disordered phase. The reconstruction also exposes the soft sector at the Ising transition as multidimensional: the ferromagnetic $A_1$ mode drives the instability while the nematic $B_1$ and diagonal-nematic $B_2$ modes cooperate, locked together by emergent rotational symmetry.","feed_headline":"One contact vertex reproduces phi^4 self-energy to 0.1%","feed_subtitle":"Closed with that contact, the parquet equations match Monte Carlo self-energy to better than one-tenth of a percent.","key_machinery":"The central object is the fully irreducible vertex $\\Lambda$, the part of the two-particle vertex that cannot be torn apart by cutting two particle lines, defined through the parquet equation $\\Lambda=\\Gamma_{\\mathrm{ph}}+\\Gamma_{\\mathrm{pp}}+\\Gamma_{\\overline{\\mathrm{ph}}}-2F$. Because the field is real and the interaction is a single on-site $\\phi^4$ term, the connected four-point function is fully crossing symmetric, so the three channels reduce to one function evaluated at three different momentum transfers and the parquet loop collapses to an iteration over the transfer momentum alone. The argument is carried by the observation that $\\Lambda$ behaves as a contact: the effective value $\\lambda_{\\mathrm{eff}}=N^{-2}\\sum_{Q,k,p}\\Lambda(k,p;Q)$ alone, fed into the Schwinger–Dyson sunset diagram, reproduces the Monte Carlo self-energy to better than a tenth of a percent. The instability analysis runs through the symmetrized Bethe–Salpeter kernel $K=\\chi_0^{1/2}\\Gamma\\chi_0^{1/2}$, whose leading eigenvalues in the $C_{4v}$ sectors track the ferromagnetic, nematic, and diagonal-nematic channels.","core_discovery":"On its own terms, the paper claims that the fully irreducible vertex $\\Lambda$, defined by the parquet equation $\\Lambda = \\Gamma_{\\mathrm{ph}} + \\Gamma_{\\mathrm{pp}} + \\Gamma_{\\overline{\\mathrm{ph}}} - 2F$, is a local contact: after the crossed-channel ladders are removed from the Monte Carlo-reconstructed 2PI vertex, the relative-coordinate weight at one lattice spacing is only 0.2% to 3% of the on-site value, and at $L=16$, $\\beta=0.68$ the ratio $\\Lambda(r_1)/\\Lambda(r_0)=0.0250\\pm0.0064$ is read as a pure contact within four standard deviations. Inserting the effective contact $\\lambda_{\\mathrm{eff}}=N^{-2}\\sum_{Q,k,p}\\Lambda(k,p;Q)$ into the parquet and Schwinger–Dyson equations while holding the propagator at its Monte Carlo value reproduces the measured self-energy to better than one-tenth of a percent across the convergent window, which the paper presents as a first-principles benchmark of D$\\Gamma$A. Away from criticality the fully self-consistent single-site D$\\Gamma$A vertex agrees with $\\lambda_{\\mathrm{eff}}$ within a few percent, while near criticality the measured contact is enhanced by about 15% relative to the bare interaction rather than screened. The paper also claims that the critical soft sector is multidimensional—$A_1$, $B_1$, and $B_2$ all grow with system size—and that in the critical region the physical solution of the parquet equations becomes a repulsive fixed point driven initially by a single order-parameter mode.","pith_inferences":["If the roughly 2.5% nearest-neighbor weight seen in $\\Lambda$ at criticality is a real remnant rather than noise, the contact-only parquet closure could still reproduce the self-energy to 0.1% only because that tail contributes negligibly to the sunset diagram; a testable extension is to include a short-range tail and compute when the benchmark breaks.","The same Monte-Carlo-reconstruction route could be applied to $O(N)$ models, where the $N=2$ quasi-long-range-ordered phase would be a sharper test of the contact picture because a gapless Goldstone sector may feed nonlocal weight into the fully irreducible vertex.","The finding that Dyson feedback destabilizes the self-consistent iteration before the fixed-propagator map does suggests that any self-consistent closure whose ladder eigenvalue approaches unity may lose convergence before the physical transition, independent of the specific parquet equations.","Because the center-of-mass structure of the 2PI vertex lives almost entirely in the ladders, the locality of the fully irreducible vertex is what makes local approximations viable; whether this persists in fermionic systems with multivalued Luttinger–Ward functionals remains an open question."],"forward_implications":["If the fully irreducible vertex really is a contact, the dynamical vertex approximation D$\\Gamma$A is validated as a near-exact closure for local observables in the disordered phase of $\\phi^4$ lattice field theory.","The measured vertex becomes a controlled benchmark: parquet-type two-particle theories can be tested against it on the disordered side of the transition, and the boundary of their validity is located where the correlation length exceeds one lattice spacing.","A faithful minimal description of the near-critical vertex must include the $B_1$ and $B_2$ stress-tensor channels alongside the $A_1$ energy channel, and the $B_1/B_2$ eigenvalue ratio locks to a finite value as the lattice grows, reflecting emergent rotational symmetry.","Parquet self-consistency loses convergence below the thermodynamic transition because the physical fixed point becomes repulsive; the loss is a finite-size effect whose onset drifts toward $\\beta_c$ as the system size increases.","In the ordered phase the zero-transfer eigenvalue collapses because the ferromagnetic weight condenses into the order parameter, while finite-momentum fluctuations persist well into the ordered phase."],"supporting_citations":[{"why":"Supplies the Bethe–Salpeter equation used to invert the connected two-particle correlator into the 2PI vertex.","marker":"[4]"},{"why":"Provides the exact Ising solution that identifies the $A_1$ energy channel and the $B_1$, $B_2$ stress-tensor channels.","marker":"[5]"},{"why":"Provides the cluster algorithm used to generate the Monte Carlo data.","marker":"[15]"},{"why":"Defines the parquet formalism from which the fully irreducible vertex is extracted by removing crossed-channel ladders.","marker":"[6–8]"},{"why":"Gives the functional-derivation form of the parquet equations that collapses the three channels into a single momentum-transfer iteration.","marker":"[31]"},{"why":"Fixes the Schwinger–Dyson sunset diagram through which the vertex enters the self-energy.","marker":"[32]"},{"why":"Establishes uniqueness of the Luttinger–Ward functional on Euclidean lattices, justifying the invertibility assumptions behind the vertex reconstruction.","marker":"[19]"},{"why":"Defines the dynamical vertex approximation that the local-contact result benchmarks.","marker":"[14]"},{"why":"Supplies the rank-one sign-flip stabilization used to reach the repulsive fixed point at the parquet convergence wall.","marker":"[25]"}],"fun_headline_variants":["2PI vertex of phi^4 becomes local contact, parquet matches to 0.1%","Parquet with contact vertex hits MC self-energy to 0.1%","Critical sector of phi^4 is multidimensional: A1, B1, B2 cooperate","Local contact Lambda benchmarks DGammaA in phi^4"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that the fully irreducible vertex is exactly a local contact rests on treating the nearest-neighbor weight $\\Lambda(r_1)/\\Lambda(r_0)=0.0250\\pm0.0064$ measured at the critical point as zero; if that few-percent tail is real, the contact-only parquet closure is an approximation rather than an exact benchmark.","fun_headline_variants_meta":{"raw":{"variants":["2PI vertex of phi^4 becomes local contact, parquet matches to 0.1%","Parquet with contact vertex hits MC self-energy to 0.1%","Critical sector of phi^4 is multidimensional: A1, B1, B2 cooperate","Local contact Lambda benchmarks DGammaA in phi^4"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001567,"raw_usage":{"total_tokens":6400,"prompt_tokens":1229,"completion_tokens":5171,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":845,"completion_tokens_details":{"reasoning_tokens":5084}},"tokens_in":845,"tokens_out":5171,"duration_ms":31756,"temperature":1.0,"reasoning_tokens":5084,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:38:40.562104+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run higher-statistics Monte Carlo at $L=32$ or $L=64$ near $\\beta\\simeq0.685$ and measure $\\Lambda(r_1)/\\Lambda(r_0)$: if the ratio settles at a nonzero value of order 2–3 percent rather than falling toward zero, the pure-contact picture is wrong. A complementary check is to include a short-range tail of that size in the parquet closure and test whether the resulting self-energy shifts by more than the claimed one-tenth of a percent.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the exact Ising solution that identifies the $A_1$ energy channel and the $B_1$, $B_2$ stress-tensor channels."},{"cited_title":"Berges, N","cited_arxiv_id":null,"evidence_quote":"Provides the cluster algorithm used to generate the Monte Carlo data."},{"cited_title":"Gaenko, A","cited_arxiv_id":null,"evidence_quote":"Gives the functional-derivation form of the parquet equations that collapses the three channels into a single momentum-transfer iteration."},{"cited_title":"Updated Core Libraries of the ALPS Project","cited_arxiv_id":"1811.08331","evidence_quote":"Fixes the Schwinger–Dyson sunset diagram through which the vertex enters the self-energy."},{"cited_title":"Gunnarsson, G","cited_arxiv_id":null,"evidence_quote":"Supplies the rank-one sign-flip stabilization used to reach the repulsive fixed point at the parquet convergence wall."}],"review_version":2}