{"id":"3c3c0a30-20e9-4d56-825c-c664ac9f93f7","arxiv_id":"2411.16382","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Conformal mapping of the Borel-transformed matching kernel reduces the scale-variation error band for quark correlation function to PDF matching by up to 40%.","lead":"This paper applies a known mathematical technique, conformal mapping, to the series that connects lattice QCD computations to parton distribution functions. The method reduces the factorization-scale dependence of the N3LO matching kernel by up to 40% in a numerical test with CT18NNLO PDFs.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 40% RMS reduction is computed from a scale-band width that the paper itself says cannot fully describe convergence; robustness of the reduction to the chosen omega and scale windows is untested.","rationale":"The paper is a proof-of-principle application of a standard conformal-mapping/Borel resummation technique (Eqs. 6-11) to the QCF-to-PDF matching kernel. It uses a genuine N3LO kernel input [30] and CT18NNLO PDFs, and the singularity positions uUV=-1 and uIR=1/2 are supported by cited quasi-PDF renormalon studies, so the approach itself is not the problem. The stress point is the evidence for the headline number, not the method. Section IV computes sigma_RMS over an omega grid from scale-variation bands. The same section states that the band width has a zero at the intersection point and that the width alone cannot fully describe convergence. Since the quoted 40% reduction is solely a ratio of such widths, the result may be an artifact of where the zero falls within the chosen 0<omega<20 range and the wide 1.3-15 GeV scale window. The acknowledged NNLO-PDF/N3LO-kernel mismatch further means the scale band mixes two sources of scale dependence, and the paper does not decompose them. A focused recomputation of sigma_RMS for sub-windows and a narrower scale range would settle whether the reduction is robust or window-dependent. If it is window-dependent, the paper should be reported as a method illustration requiring further validation, not as a quantitative convergence claim. This supports keeping the reader's CONDITIONAL verdict.","tokens_in":9169,"tokens_out":13490,"duration_ms":139763,"concrete_test":"Recompute the N3LO u-quark sigma_RMS ratio (non-power series vs alpha_s series) with the same inputs but (i) split the omega grid around the zero-width node (e.g., omega in [2,8] and [10,20] separately) and (ii) use a physically motivated narrower scale range (1.5-5 GeV) in addition to 1.3-15 GeV. If the ratio does not remain close to 0.54 (the 40% reduction) in all windows, varying by more than about 20%, the headline claim is window-dependent and should be downgraded.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central numerical claim rests entirely on the RMS width sigma_RMS of the factorization-scale band defined in Eq. (15) and quoted in Section IV. The manuscript itself flags, in the paragraph following FIG.2, that the band has a zero-width node and that 'the width of the error band alone cannot fully describe the convergence behavior of the series.' Because sigma_RMS averages half-widths over a fixed omega grid, its value is dominated by how quickly the band widens away from that node and by the chosen omega range and scale range (1.3-15 GeV). The 40% reduction could therefore be an artifact of the node position moving under the conformal map, rather than evidence of genuinely improved convergence. In addition, the PDFs are NNLO while the kernel is N3LO, as acknowledged in Section IV, so the residual scale dependence mixes PDF-order mismatch with missing higher-order kernel terms; the conformal map acts only on the kernel series, and the RMS reduction is not decomposed between these two sources. No sensitivity test is provided for either the window choice or the scale range.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes using a conformal mapping of the Borel plane, Eq. (8), to turn the conventional power series of the QCF-to-PDF matching kernel into a non-power series, Eq. (10), with the aim of taming renormalon-driven factorial growth. The construction uses the assumed nearest singularities at u_UV = -1 and u_IR = 1/2 in the MS scheme, adopted from refs. [46,47]. The numerical section evaluates QCFs with CT18NNLO PDFs and the N3LO matching kernel of ref. [30], varying the factorization scale between 1.3 GeV and 15 GeV, and reports that the RMS width of the resulting error band, Eq. (15), is reduced by about 40% when the non-power series is used for both u- and d-quarks (from 1.31e-2 to 7.04e-3 for u; from 1.40e-2 to 8.72e-3 for d). The central claim of the paper is that conformal mapping improves the convergence and stability of the matching series.","tokens_in":9378,"tokens_out":4904,"duration_ms":49454,"significance":"If the claimed improvement is robust, the paper offers a computationally cheap method to reduce perturbative uncertainty in lattice-QCD-based PDF extractions, with no new loop calculations required. The paper has genuine strengths: no parameters are fitted to the final result, the conformal map is fixed by external renormalon positions, and the numerical inputs are standard. The implementation of the Borel transform and conformal map is straightforward and, in principle, reproducible. However, the evidence for convergence improvement rests entirely on a single scale-variation proxy, and the paper does not provide sensitivity tests of the two load-bearing assumptions: the singularity positions and the range/window choices entering the RMS metric. The order mismatch between NNLO PDFs and an N3LO kernel further complicates the interpretation of the scale-variation band.","major_comments":[{"comment":"This is a complete sentence.","section":"§IV, Eq. (15)"},{"comment":"This is a complete sentence.","section":"§III.B, Eq. (8)"},{"comment":"This is a complete sentence.","section":"§IV, order-mismatch paragraph"}],"minor_comments":[{"comment":"This is a complete sentence.","section":"§II, Eq. (1)-(2)"},{"comment":"This is a complete sentence.","section":"§IV, Eq. (15)"},{"comment":"This is a complete sentence.","section":"FIG. 2"},{"comment":"This is a complete sentence.","section":"§III.B"},{"comment":"This is a complete sentence.","section":"§III.A, Eq. (11)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a short methodological note whose central numerical claim needs additional robustness checks before it can be accepted. The lack of any sensitivity analysis of the singularity positions and of the RMS-window choices is the main concern. If the revised version includes such tests and a decomposition of the scale-dependence sources, it could be suitable for publication; in its current form the 40% improvement claim is not fully supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague, quick take on Zhang's paper.\n\nThe new thing is the numbers. For the N3LO matching kernel with CT18NNLO PDFs, rewriting the alpha_s series as a conformally mapped non-power series reduces the RMS width of the factorization-scale error band from 1.31e-2 to 7.04e-3 for the u quark and 1.40e-2 to 8.72e-3 for the d quark, about 40%. That specific application of conformal mapping to the QCF-to-PDF matching kernel is new, as far as I can tell. The method itself is well-established, so the novelty is the demonstration, not the technique.\n\nWhat the paper does well: the write-up is clear, the conformal mapping construction is standard and correctly explained, and the author is upfront about limitations. He acknowledges that the PDFs are only NNLO, so N3LO scale dependence is not fully cancellable, and he explicitly says after Fig. 2 that the error band width alone cannot fully describe convergence. That is the right kind of honesty.\n\nThe soft spots are real but not fatal. The whole quantitative claim hangs on sigma_RMS, a single diagnostic defined in Eq. (15). The zero-width node in the band means sigma_RMS can be influenced by where that node sits and by the chosen omega and scale windows. The stress-test concern is fair: the 40% reduction might partly come from the node shifting under the conformal map rather than from genuinely improved convergence. The author does not test sensitivity to the assumed renormalon positions (uUV = -1, uIR = 1/2), nor to the scale window (1.3-15 GeV) or omega grid, and no code or tabulated values are provided, so the results cannot be independently reproduced. The scale variation also mixes the NNLO PDF-order mismatch with the N3LO kernel resummation, and that decomposition is not attempted.\n\nProportionately, this is a proof-of-principle with suggestive evidence, not a settled conclusion. The idea is plausible and the execution is competent. To make the 40% robust, the revision should add sensitivity scans, separate the sources of scale dependence, and release the numerical implementation. The paper would benefit from an external benchmark, but that is not essential for publication.\n\nWho is this for? Lattice QCD practitioners working on PDF extraction and people interested in resummation in LaMET. It is not a breakthrough, but it is a useful data point.\n\nMy recommendation: send it to peer review. It deserves referee time, but the referee should ask for the sensitivity tests and the data or code before publication. If the 40% survives those tests, this becomes a citable result.","headline":"A modest but legitimate proof-of-principle that conformal mapping reduces scale dependence of the N3LO QCF matching kernel by ~40%; the number rests on one self-referential diagnostic and needs sensitivity tests before I'd trust it.","tokens_in":9853,"tokens_out":3182,"would_cite":false,"duration_ms":28137,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["12.38.Bx","12.38.Aw","12.38.Gc"],"model":"deepseek-v4-flash","headline":"Rewriting a QCD matching kernel as a conformally mapped non-power series cuts its factorization-scale error band by about 40 percent.","keywords":["conformal mapping","Borel summation","renormalon singularities","parton distribution functions","lattice QCD","matching kernel","non-power series","factorization scale uncertainty"],"falsifier":"Looking at the Borel transform of the N3LO matching kernel at high order and locating the nearest singularity on the positive real axis would settle it: if it is not at $u=\\frac12$ in the MS scheme, the map loses its justification. A cheaper check is to recompute the RMS band of Section IV with the factorization-scale range widened or with renormalization-scale variation included; if the non-power series then no longer beats the $\\alpha_s$ series by roughly a factor of two, the claimed improvement is specific to the band measure used in the paper.","tokens_in":8984,"feed_emoji":"📉","tokens_out":7749,"duration_ms":66138,"temperature":0.7,"pith_summary":"The paper proposes that conformal mapping in the Borel plane can tame the divergent high-order expansion of the matching kernel that converts quark correlation functions into parton distribution functions (PDFs). For u- and d-quark PDFs from CT18NNLO at N3LO in the MS scheme, replacing the ordinary $\\alpha_s$ power series by the mapped non-power series shrinks the RMS half-width of the factorization-scale error band from $1.31\\times10^{-2}$ to $7.04\\times10^{-3}$ for the u quark and from $1.40\\times10^{-2}$ to $8.72\\times10^{-3}$ for the d quark, an improvement of about 40%. A sympathetic reader would take away that lattice-based PDF extractions can gain accuracy at fixed loop order by resumming the matching kernel more cleverly, without waiting for another order of perturbation theory.","feed_headline":"Conformal mapping cuts PDF matching error by 40%","feed_subtitle":"Rewriting the strong-coupling matching series as a non-power series cuts the N3LO error band for u and d quarks.","key_machinery":"The object that carries the argument is the conformal map $\\tilde w(u)$ of Eq. (8), which reparametrizes the Borel variable $u$ so that the nearest singularities of the Borel-transformed matching kernel land on the unit circle $|w|=1$. The series is then re-expanded in powers of $w$, producing new coefficients $c'_k$, and the observable is rebuilt as $\\phi(\\alpha_s)=\\sum_i c'_i W_i(\\alpha_s)$ with principal-value integrals $W_i(\\alpha_s)$ defined in Eq. (11). This construction preserves the original asymptotic expansion while turning it into a convergent non-power series over the holomorphic domain of the Borel transform, which is what shrinks the factorization-scale error band in the numerical section.","core_discovery":"The central claim is that the non-power series built from the conformal map $\\tilde w(u) = \\frac{\\sqrt{1-u/u_{\\rm UV}} - \\sqrt{1-u/u_{\\rm IR}}}{\\sqrt{1-u/u_{\\rm UV}} + \\sqrt{1-u/u_{\\rm IR}}}$ makes the matching kernel for quark correlation functions more convergent in practice. With the nearest renormalon singularities at $u_{\\rm UV}=-1$ and $u_{\\rm IR}=\\frac12$ in the MS scheme, the map moves them onto the unit circle, so the Borel-transformed series becomes well-defined in a larger domain. At N3LO with CT18NNLO PDFs, the RMS half-width of the band from varying the factorization scale between 1.3 GeV and 15 GeV drops from $1.31\\times10^{-2}$ to $7.04\\times10^{-3}$ for the u quark and from $1.40\\times10^{-2}$ to $8.72\\times10^{-3}$ for the d quark. The paper also notes that higher-order QCFs in the non-power series stay closely aligned with lower-order results, indicating that the order-by-order instability typical of the $\\alpha_s$ series is mitigated.","pith_inferences":["The reported improvement is measured by the factorization-scale band alone; a natural next test is to vary the renormalization scale and the scheme to see whether the roughly 40% gain survives a more complete uncertainty budget.","The map's parameters come from the assumed singularity positions $u_{\\rm UV}=-1$ and $u_{\\rm IR}=\\frac12$; directly extracting those positions from the N3LO Borel transform would convert the assumption into a measured input and would show how sensitive the gain is to the map.","The conformal-mapping mechanism applies to any Borel-summable QCD series with known leading renormalons, so the same construction could be tested on DGLAP evolution kernels or on the coefficient functions of other factorization theorems without new loop calculations."],"forward_implications":["At fixed perturbative order, the non-power matching kernel gives narrower factorization-scale uncertainties, so PDF extractions from existing lattice correlators can be quoted with smaller theory errors.","Because the RMS band shrinks by about 40% for both u and d quarks, the same gain is expected in the valence-quark sector of the PDF, where lattice data are most precise.","If the $u=\\frac12$ renormalon cancels in ratio or hybrid renormalization schemes, the same conformal treatment of those matching kernels should yield even faster convergence than the MS-scheme result shown here.","The method extends beyond quark PDFs: the paper's outlook applies the same non-power matching kernel strategy to distribution amplitudes and generalized parton distributions."],"supporting_citations":[{"why":"supplies the N3LO hard kernel in the MS scheme that is the perturbative input being resummed.","marker":"[30]"},{"why":"supplies the CT18NNLO PDFs for u and d quarks used to compute the numerical QCFs and the RMS error bands.","marker":"[2]"},{"why":"fix the nearest Borel-plane singularities at $u_{\\rm UV}=-1$ and $u_{\\rm IR}=\\frac12$ for the MS-scheme QCF, which set the parameters of the conformal map.","marker":"[46, 47]"},{"why":"defines the factorization formula that the matching kernel enters, so the kernel is the object whose convergence is being improved.","marker":"[11]"},{"why":"provides the conformal-mapping and resurgence framework in which the non-power series and the principal-value integrals are defined.","marker":"[26]"},{"why":"supports the claim that ratio and hybrid renormalization cancel the $u=\\frac12$ renormalon, shifting the singularity structure to $u_{\\rm IR}=2$ and pointing to the projected improvement in those schemes.","marker":"[34]"}],"fun_headline_variants":["Conformal trick slashes PDF error by 40%","Borel-plane map tames renormalons, shrinks PDF errors","Non-power expansion beats alpha_s for PDFs","Renormalon-cured series cuts PDF error 40%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the matching kernel's Borel transform really has its nearest singularities at $u_{\\rm UV}=-1$ and $u_{\\rm IR}=\\frac12$ in the MS scheme, because the conformal map in Eq. (8) is built on exactly those points and the claimed convergence gain would not materialize if they sat elsewhere.","fun_headline_variants_meta":{"raw":{"variants":["Conformal trick slashes PDF error by 40%","Borel-plane map tames renormalons, shrinks PDF errors","Non-power expansion beats alpha_s for PDFs","Renormalon-cured series cuts PDF error 40%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001373,"raw_usage":{"total_tokens":5611,"prompt_tokens":1035,"completion_tokens":4576,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":651,"completion_tokens_details":{"reasoning_tokens":4507}},"tokens_in":651,"tokens_out":4576,"duration_ms":38449,"temperature":1.0,"reasoning_tokens":4507,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:11:03.699174+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Looking at the Borel transform of the N3LO matching kernel at high order and locating the nearest singularity on the positive real axis would settle it: if it is not at $u=\\frac12$ in the MS scheme, the map loses its justification. A cheaper check is to recompute the RMS band of Section IV with the factorization-scale range widened or with renormalization-scale variation included; if the non-power series then no longer beats the $\\alpha_s$ series by roughly a factor of two, the claimed improvement is specific to the band measure used in the paper.","supporting_citations":[{"cited_title":"Quark correlation functions at three-loop order and extraction of splitting functions","cited_arxiv_id":"2410.05141","evidence_quote":"supplies the N3LO hard kernel in the MS scheme that is the perturbative input being resummed."},{"cited_title":"Power corrections to the modified QCD perturbative series based on conformal mapping of the Borel plane","cited_arxiv_id":"2403.10844","evidence_quote":"provides the conformal-mapping and resurgence framework in which the non-power series and the principal-value integrals are defined."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supports the claim that ratio and hybrid renormalization cancel the $u=\\frac12$ renormalon, shifting the singularity structure to $u_{\\rm IR}=2$ and pointing to the projected improvement in those schemes."}],"review_version":1}