{"id":"bed8936e-8a3e-4071-8011-b0cd5d732886","arxiv_id":"2507.07014","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A hard-soft split combined with a finite-temperature Loop-Tree Duality method computes the resummed 4d thermal effective potential without high-temperature expansions, demonstrated in a scalar-Yukawa model.","lead":"Scientists calculating early-universe phase transitions usually rely on high-temperature approximations that fail in the strongest cases. This paper introduces a numerical method that keeps the full temperature dependence and tests it in a simple model at two-loop order plus one three-loop integral.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Test the δL=0 truncation at large y by recomputing the three-loop Mercedes counterterm with the first higher-dimensional (φ^6/dimension-6) operator included; if unmatched UV/IR or order-one shifts appear, full-mass-range validity fails.","rationale":"The reader's weakest assumption is the δL=0 truncation, and I agree that it is the single load-bearing point. The central claim is full-mass-range validity, so any check must target whether the omitted higher-dimensional operators are actually suppressed in the strong-transition regime they highlight. The paper offers only an assertion of natural alignment, not an argument or a test. Importantly, the three-loop example in Eq. (24)-(30) does not test the truncation: it only exercises the M3 and G3 counterterms, and leaves d_UV undetermined, so the paper's own three-loop completeness claim is not yet independently checkable. The two-loop calculation is coherent, the hotLTD results are archived (Zenodo), and the comparison with existing high-T results at small y provides real support, so the correct response is conditional acceptance rather than rejection. An honest but concrete test of the leading higher-dimensional operator is the minimal check that would settle the scope of the central claim.","tokens_in":38796,"tokens_out":3188,"duration_ms":31013,"concrete_test":"Recompute the three-loop Mercedes example of Section V after restoring a single representative higher-dimensional operator in Eq. (13), e.g. the dimension-6 interaction c_6(φ) φ^6/T^2 obtained by integrating out the heavy fermion at one loop without a high-T expansion. Keep δL ≠ 0, match its coefficient, and re-evaluate the IR counterterm of Eq. (29)-(30); if the counterterm changes by more than a few percent of the retained δG3|_M term at y ≳ 0.8 (benchmark Eq. (31)), or if a new uncancelled UV pole in ε appears, then the δL=0 truncation is inconsistent at the claimed order.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim is that Eq. (2) plus hotLTD computes the resummed 4d thermal effective potential across all mass regimes, including strong transitions with field-dependent masses of order πT. The most load-bearing assumption is the reader-identified δL=0 truncation of Eq. (13): the paper asserts that with fully massive sum-integrals, higher-dimensional operators align with the coupling expansion, but it gives no proof or check. The only three-loop example uses the EFT only through M3 and δG3 counterterms and omits the whole higher-dimensional tower. In the strong-transition regime Mψ ∼ yφ ∼ πT, a dimension-6 operator such as a fermion-induced φ^6/(πT)^2 correction is not obviously suppressed relative to the retained G3φ^3/3! and λ3φ^4/4! terms. If those operators contribute at the same order, V_res,soft is incomplete and the missed IR counterterms are not subtracted from ΔV_hard, so Eq. (30) is not the full three-loop answer even though the individual sum-integrals may be evaluated correctly. The Fig. 3 benchmark results at large y therefore rest on an unverified truncation precisely in the regime where the high-T expansion fails.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a new framework for computing the equilibrium thermal effective potential without recourse to the high-temperature expansion. The organizing identity is Eq. (2), which rewrites the resummed potential as a naive hard contribution minus a naive soft contribution plus a resummed soft contribution. The hard contribution is evaluated with a finite-temperature generalization of Loop-Tree Duality (hotLTD), while the soft contribution is generated from a dimensionally-reduced 3d effective theory. As a proof of principle in a scalar-Yukawa model, the paper presents a complete one- and two-loop calculation (Eqs. (21) and (23)), a partially completed three-loop example involving the Mercedes-type sum-integral (Eqs. (24)-(30)), and numerical results for Tc and latent heat in Fig. 3. The central claims are that the framework works across all mass regimes, that the two-loop result is complete, and that the three-loop example represents the first fully massive three-loop sum-integral computation without high-temperature expansions.","tokens_in":39146,"tokens_out":4805,"duration_ms":60012,"significance":"If the framework is correct, it would be a substantial advance for strong first-order phase transitions, where field-dependent masses can be of order pi*T and the high-T expansion is unreliable. The paper's strengths include an explicit and plausible two-loop construction, local cancellation of IR divergences at the integrand level, use of the R-operation for UV subtraction, and public numerical data for the thermal functions in Zenodo. The two-loop thermal functions bT and fT are also cross-checked against Ref. [76], which is an important independent validation. The broader conceptual claim that Eq. (2) plus hotLTD can systematically push order-by-order is attractive and worth developing. However, the load-bearing assumption that the 3d EFT can be truncated at super-renormalizable order (delta L = 0) is asserted without proof or test, and the three-loop result is only partial, so the strongest claims in the abstract and Section V must be qualified or defended.","major_comments":[{"comment":"The statement that with fully massive sum-integrals 'the impact of higher-dimensional operators naturally aligns with the coupling expansion' and that it suffices to set delta L = 0 is a load-bearing assumption, but it is not proved or tested. In the strong-transition regime where M_psi ~ y*phi ~ pi*T, a dimension-6 operator such as a fermion-induced phi^6/(pi*T)^2 correction is not obviously suppressed relative to the retained G3*phi^3/3! and lambda3*phi^4/4! terms. Since V_res,soft and V_naive,soft are both constructed from this truncated EFT, the IR counterterms subtracted from Delta V_hard inherit the truncation. If higher-dimensional operators contribute at the same order, Eq. (2) with Eq. (13) is not exact over the claimed full mass range, and the large-y results in Fig. 3 rest on an unverified premise. The authors should provide an explicit power-counting estimate of the leading higher-dimensional operator, or include the first such operator in the three-loop counterterm calculation and demonstrate that its effects are numerically negligible in the benchmark of Eq. (31).","section":"Sec. III, Eq. (13)"},{"comment":"The three-loop result is presented as a 'novel three-loop extension' and as the 'first fully massive three-loop sum-integral computation', but Eq. (30) is not a complete three-loop contribution. The text explicitly states that determining the constant piece of d_UV requires two-loop thermal contributions to linear order in epsilon and is 'a formidable task lying beyond the scope of this work,' and that the remaining renormalized thermal contributions are to be determined later. Thus, only the finite thermal function d_T is actually computed, while d_UV is left undetermined. The paper should clearly state in the abstract and in Section V that the three-loop calculation is partial: it demonstrates the viability of the hotLTD machinery on one non-trivial topology, but it does not yet provide a fully renormalized three-loop effective potential.","section":"Sec. V, Eq. (30)"},{"comment":"The numerical evaluation of the three-loop thermal function d_T in Eq. (30) relies entirely on the hotLTD algorithm, which is described only in an in-preparation reference [99], and the paper's Appendix C works out a two-loop sunset example rather than the three-loop diagram. Consequently, the flagship three-loop numerical result has no fully independent verification: the agreement with Ref. [76] covers only the two-loop functions b_T and f_T, not the three-loop d_T. The authors should either release the hotLTD implementation, provide a detailed three-loop subtraction and Matsubara-summation description for the Mercedes diagram, or give an independent numerical cross-check for d_T before the claim of a completed three-loop computation is made.","section":"Sec. II and Appendix C"}],"minor_comments":[{"comment":"The text states that there are 29 distinct three-loop sum-integrals, of which 15 are non-factorized, but only the single Mercedes-type integral M is evaluated. The paper should explicitly state that the other 14 non-factorized three-loop structures are not computed in this work, so that readers do not infer a complete three-loop effective potential.","section":"Sec. V"},{"comment":"The notation h(q_phi) - (1/2) h(2 q_psi) + 4 h(q_psi) would benefit from an explicit definition of the argument scaling for the fermionic function; the current text explains the rescaling verbally but the reader must reconstruct the exact relation from the h defined in Eq. (A5).","section":"Eq. (21)"},{"comment":"The caption says that the bands depict renormalization-scale variation, but the text in Section VI explains that the range is Lambda in [0.5 pi T_c^LO, 2 pi T_c^LO]. It would be clearer to state the range directly in the caption.","section":"Fig. 3"},{"comment":"The hotLTD algorithm is cited as an in-preparation paper, and the Zenodo data [126] is from the same group. Please provide a stable DOI or version identifier for the data and, if possible, a preprint number for [99], so that the numerical claims can be verified independently.","section":"References [99] and [126]"}],"recommendation":"major_revision","confidential_remarks":"The main concern is the delta L = 0 truncation in Eq. (13), which is exactly the assumption that must hold for the claimed full-mass-range validity. This is not a circularity or a disagreement with consensus; it is an unverified power-counting assumption in precisely the regime where the high-T expansion fails. I believe it is fixable by adding a dedicated estimate or a numerical test with the first higher-dimensional operator, but until then the strongest claims should be softened. The three-loop result should also be labelled as partial. I recommend major revision rather than rejection: the two-loop construction and the two-loop validation against Ref. [76] are solid and valuable, and the framework is worth developing further."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper does something real: it pushes the no-high-T-expansion program forward by computing the full two-loop effective potential in a scalar-Yukawa model and demonstrating that the hard-soft split from QCD pressure works. The two-loop expression Eq. (23) is assembled from explicit sum-integrals, the scale logarithms cancel, and the thermal functions bT and fT match Ref. [76] where they overlap. The hotLTD evaluation of the three-loop Mercedes sum-integral in Eqs. (24)-(30) is a genuine technical advance, and the Zenodo data makes the numerics reproducible. The discussion of imaginary parts is thoughtful and the thermodynamic comparison in Fig. 3 is clear. The paper also cites Laine-Losada and Laine-Meyer-Nardini honestly.\n\nThe soft spots are real but not lethal. The three-loop result is partial: d_UV is left undetermined, and only one of fifteen non-factorized sum-integrals is evaluated. That doesn't invalidate the method, but it means the abstract's 'three-loop computation' should be read as 'one three-loop sum-integral.' The hotLTD code is still in preparation, so the key numerical engine is currently a black box, though the Zenodo data partially mitigates that.\n\nThe bigger issue is the δL=0 truncation in Eq. (13). The paper asserts that fully massive sum-integrals align higher-dimensional operators with the coupling expansion, but it gives no proof. In the large-y region where the zero-mode mass is of order πT, a dimension-6 operator like φ^6/(πT)^2 is not obviously negligible relative to the retained G3 φ^3 and λ3 φ^4 terms. If those operators matter, V_res,soft is incomplete and the IR counterterms are missing, so Eq. (30) isn't the full three-loop answer even though the individual sum-integral is correct. The stress-test suggestion to recompute with the first dimension-6 operator included is the right check.\n\nOverall: this is a genuine advance with an honest caveat list, and the central two-loop results hold up. The δL truncation deserves scrutiny, but it's an addressable assumption, not a fatal flaw. I'd send this to a serious referee, and I'd probably cite it for the two-loop framework once I trust the hotLTD details.","headline":"A genuine two-loop no-high-T effective potential with a promising but incomplete three-loop extension; the δL=0 truncation is the main thing to test.","tokens_in":39609,"tokens_out":3632,"would_cite":true,"duration_ms":42255,"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":"A hard-soft split of the resummed 4d thermal effective potential, evaluated with finite-temperature Loop-Tree Duality, yields full-mass two-loop and three-loop phase-transition thermodynamics without high-temperature expansions.","keywords":["cosmological phase transitions","thermal effective potential","high-temperature expansion","dimensional reduction","loop-tree duality","thermal sum-integrals","scalar-Yukawa model","gravitational waves"],"falsifier":"Compute the scalar-Yukawa model at the benchmark point of Eq. (31) with y ≈ 0.9 using a full 4d lattice Monte Carlo simulation and compare T_c and latent heat with the paper's full one- and two-loop results; a disagreement well outside the quoted renormalization-scale band would refute the claim of full-mass-range validity. Alternatively, add the leading higher-dimensional operator in the 3d EFT and check whether its effect at strong transitions is suppressed relative to the super-renormalizable terms.","tokens_in":38625,"feed_emoji":"⚛️","tokens_out":8612,"duration_ms":87410,"temperature":0.7,"pith_summary":"The paper claims that cosmological phase-transition thermodynamics can be computed perturbatively to high loop orders without the standard high-temperature expansion. The key step is an exact reorganization of the resummed four-dimensional thermal effective potential into a naively computed hard part and a resummed soft part, V_res = (V_naive − V_naive,soft) + V_res,soft, with all masses kept exact. The hard part is evaluated numerically using a finite-temperature generalization of Loop-Tree Duality, while the soft part comes from a three-dimensional effective theory with fully massive matching. As proof of principle, the authors present a complete two-loop effective potential and a three-loop extension—the first fully massive three-loop thermal sum-integral computation without high-temperature expansions—for a scalar-Yukawa model. If correct, this removes a key obstacle to reliable predictions for strong first-order phase transitions, where high-temperature approximations break down.","feed_headline":"Three-loop thermal potential computed without high-T expansions","feed_subtitle":"Strong first-order transitions, where standard approximations fail, become tractable at full mass.","key_machinery":"The central object is the splitting identity V_res = (V_naive − V_naive,soft) + V_res,soft (Eq. 2), together with the 'hotLTD' algorithm that evaluates the hard combination ΔV_hard = V_naive − V_naive,soft. The splitting isolates the physics of hard modes (momenta of order πT) from soft static zero modes; the soft part is computed in a 3d effective theory with parameters matched to the full theory at fully massive one-loop order, while the hard part is evaluated by analytic Matsubara summation, local UV subtraction via Bogoliubov's R-operation, and Monte Carlo integration of the remaining 3d momentum integrals. The matching corrections $δM_3^{2}$ and δG_3 enter as IR counterterms that cancel the soft singularities of the naive sum-integrals locally at the integrand level.","core_discovery":"The central claim is that the identity V_res = (V_naive − V_naive,soft) + V_res,soft gives a valid order-by-order reorganization of the resummed 4d thermal effective potential across all mass regimes. V_naive is the ordinary unresummed loop expansion in the full theory; V_res,soft resums the static soft zero modes within dimensional reduction, with matching parameters computed without high-temperature expansions; V_naive,soft is the expansion of that soft expression in the matching corrections, acting as local thermal counterterms that cancel infrared divergences in V_naive at the integrand level. In the scalar-Yukawa model this yields a complete two-loop resummed potential and a three-loop contribution from a Mercedes-type diagram with a fermion loop, stated as the first fully massive three-loop sum-integral evaluation without high-temperature expansions. The paper further claims that the resulting full one- and two-loop thermodynamics remain under control at large Yukawa couplings, giving significantly stronger transitions than the high-temperature 3d EFT and avoiding its breakdown near y ≈ 0.97.","pith_inferences":["Beyond the paper: if the framework's predictions survive comparison with full 4d lattice simulations, the standard no-go result that single-step electroweak transitions cannot produce LISA-visible gravitational waves would need to be revisited, since that bound was derived within the high-temperature 3d EFT.","Beyond the paper: the same integrand-level hard-soft decomposition could be applied to the effective action and bubble nucleation rate, where the high-temperature expansion is also questionable for strong transitions; the paper lists this as future work but gives no demonstration.","Beyond the paper: a direct test of the paper's claim about imaginary parts is to compute the bubble profile and nucleation rate in the tachyonic region and verify that the imaginary component of the potential does not alter the physical decay rate; the paper only checks the minima in its own numerical examples.","Beyond the paper: combining the full-mass potential with a field-dependent renormalization scale (which the paper does not implement) would extend the benchmark to physically stable zero-temperature vacuum at large y and could change the quantitative predictions for strong transitions."],"forward_implications":["For strong first-order transitions in the scalar-Yukawa benchmark, the full one- and two-loop results predict smaller critical temperatures and larger latent heat than the high-temperature 3d EFT, and they remain well-behaved at Yukawa couplings where the 3d EFT (at NLO in the high-T expansion) ceases to have a transition.","Because the UV and IR subtractions are performed locally at the integrand level, the framework is automatable and extends beyond two loops, as shown by the three-loop Mercedes example.","The same hard-soft split applies in gauge-Higgs theories, where the high-temperature expansion for the transition-inducing fields generically breaks down in strong first-order transitions, so the framework can be used to scrutinize current 3d EFT and lattice predictions.","The resummed soft part in the small-mass regime reduces to the standard daisy resummation, but with the thermal mass computed without high-temperature expansions, giving a consistent resummation at higher orders."],"supporting_citations":[{"why":"Establishes the effective-field-theory treatment of high-temperature thermodynamics and the unit-operator matching coefficient, the conceptual template for separating hard and soft scales.","marker":"[13]"},{"why":"Introduces the (naive minus soft) plus resummed-soft reorganization of the pressure in dense QCD, the direct antecedent of Eq. (2).","marker":"[25]"},{"why":"Provides the generic dimensional-reduction matching rules for building the 3d soft effective theory and its parameters.","marker":"[20]"},{"why":"Supplies the finite-temperature Loop-Tree Duality algorithm (hotLTD) used for the numerical evaluation of massive multiloop sum-integrals.","marker":"[99]"},{"why":"The first finite-density extension of Loop-Tree Duality with integrand-level subtractions, on which the temperature generalization builds.","marker":"[110]"},{"why":"Demonstrates automatable local UV subtraction via the R-operation in a finite-temperature four-loop computation, supporting the subtraction strategy.","marker":"[111]"},{"why":"Earlier two-loop calculation of the effective potential without temperature expansions, providing the massive two-loop sum-integrals the present computation extends.","marker":"[18]"},{"why":"Previous full two-loop thermal effective potential for strong transitions, supplying the numerical thermal functions and the scale-dependence baseline the paper improves on.","marker":"[76]"},{"why":"Establishes the scalar-Yukawa model power counting, the high-temperature 3d EFT results, and the benchmark used for the numerical comparison.","marker":"[117]"}],"fun_headline_variants":["No high-T expansions: full 4d resummed potential to three loops","Three-loop thermal potential without high-T approximations","Massive three-loop sum-integrals computed via Loop-Tree Duality","Full 4d resummation for cosmological transitions without high-T","First fully massive three-loop thermal sum-integral evaluation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The scheme assumes that once all masses are kept exact, the infinite tower of higher-dimensional operators in the 3d effective theory is suppressed at the perturbative order considered, so that truncating the EFT at super-renormalizable order (δL = 0) misses nothing at that order in the strong-transition regime.","fun_headline_variants_meta":{"raw":{"variants":["No high-T expansions: full 4d resummed potential to three loops","Three-loop thermal potential without high-T approximations","Massive three-loop sum-integrals computed via Loop-Tree Duality","Full 4d resummation for cosmological transitions without high-T","First fully massive three-loop thermal sum-integral evaluation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001325,"raw_usage":{"total_tokens":5403,"prompt_tokens":962,"completion_tokens":4441,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":578,"completion_tokens_details":{"reasoning_tokens":4352}},"tokens_in":578,"tokens_out":4441,"duration_ms":33398,"temperature":1.0,"reasoning_tokens":4352,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:49:52.057805+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the scalar-Yukawa model at the benchmark point of Eq. (31) with y ≈ 0.9 using a full 4d lattice Monte Carlo simulation and compare T_c and latent heat with the paper's full one- and two-loop results; a disagreement well outside the quoted renormalization-scale band would refute the claim of full-mass-range validity. Alternatively, add the leading higher-dimensional operator in the 3d EFT and check whether its effect at strong transitions is suppressed relative to the super-renormalizable terms.","supporting_citations":[{"cited_title":"A renormalization group improvement for thermally resummed effective potential","cited_arxiv_id":"2307.02153","evidence_quote":"Supplies the finite-temperature Loop-Tree Duality algorithm (hotLTD) used for the numerical evaluation of massive multiloop sum-integrals."},{"cited_title":"Braaten and A","cited_arxiv_id":null,"evidence_quote":"Demonstrates automatable local UV subtraction via the R-operation in a finite-temperature four-loop computation, supporting the subtraction strategy."}],"review_version":1}