{"id":"90c9aa31-b14f-4f08-98cb-077c3e126443","arxiv_id":"2607.11145","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Gradient-descent inversion of redshifted BBH mass distributions recovers the PBH mass function and, via regularized Press-Schechter, a candidate O(10^{-2}) bump in the small-scale primordial power spectrum.","lead":"The paper builds a pipeline that turns redshifted black-hole masses from gravitational-wave catalogs into a reconstructed primordial power spectrum on small scales. It shows the method can recover a candidate bump of size ~0.01 near k ~ 5.7e5 Mpc^-1 when applied to LVK events under simple mass cuts, and argues the same pipeline will work cleanly with high-redshift detections from next-generation detectors.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"The reported PPS bump is load-bearing on treating the mass-cut LVK sample as pure early-universe PBH binaries under the Sec. III assumptions; contamination or late-time effects would artifact the feature.","rationale":"The reader correctly isolates the pure-PBH sample assumption as the weakest link for the concrete LVK result while recognizing that the end-to-end inverse pipeline (redshifted-mass → f(m) via GD → β → σ^{2} → regularized P_R) is carefully derived, uniqueness is argued for the discretized problem, regularization diagnostics are shown, and the LVK application is explicitly labeled illustrative. No deeper internal inconsistency (e.g., in the nonlinear kernel or the Press-Schechter inversion) was found that would overturn the methodological core or force a stronger verdict change. The concern therefore leaves the CONDITIONAL assessment and the feasibility claim for next-generation high-z detectors intact; it only underscores that the reported O(10^{-2}) bump must be read as a proof-of-concept feature under the stated assumptions.","tokens_in":29168,"tokens_out":638,"duration_ms":44196,"concrete_test":"Generate 100 mock catalogs that mix the real 174 m>15 events with a free astrophysical fraction f_astro drawn from a standard LVK power-law+peak stellar BH mass function (same SNR and pastro cuts); re-run the full pipeline (GD mass-function reconstruction + f_PBH normalization via Eq. (42) + Tikhonov PPS inversion for the same λ grid). If the recovered β_p falls below ~10^{-3} or the feature loses regularization stability for any f_astro ≳ 0.3, the candidate bump is an artifact of the pure-PBH assumption.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest claim’s concrete numerical result (regularization-stable O(10^{-2}) bump at k_peak ≃ 5.7e5 Mpc^{-1}) is obtained by feeding the m>15 M_⊙ (174 events) and m>50 M_⊙ (6 events) LVK subsets into the gradient-descent inversion of Eq. (20) under the three Sec. III assumptions (Poisson spatial distribution, common t_dec, negligible late-time mass evolution) plus the circular-orbit SNR>8 window of Eqs. (24)–(35). The paper itself flags that these events are not known to be PBHs and that accretion, disruption, and eccentricity are omitted. If a non-negligible astrophysical fraction remains after the cuts, or if the binary formation/redshift distribution deviates from the early-universe Poisson model, both the reconstructed f(m) peak near ~20–30 M_⊙ and the subsequent Press-Schechter + Tikhonov PPS bump become selection/model artifacts rather than a primordial signal. The high-z feasibility argument is unaffected, but the LVK-derived candidate feature is not.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript develops an inverse pipeline that reconstructs the PBH mass function f(m) from the observed redshifted-mass distribution of high-redshift BBHs via gradient descent on Eq. (20), then maps the resulting abundance through the Press–Schechter relation and a Tikhonov-regularized inverse convolution to the small-scale primordial power spectrum P_R(k). The forward map from f(m) to P(m_1^z, m_2^z) is derived under three early-universe assumptions (Poisson spatial distribution, common decoupling time, negligible late-time mass evolution), with a circular-orbit SNR>8 detector window. Uniqueness of the discretized mass-function inverse is argued in Appendix A. As an illustration, mass-selected LVK events (m>15 M_⊙, 174 events; m>50 M_⊙, 6 events) are treated as PBH candidates, yielding a lognormal-like f(m) peaking near ~20–35 M_⊙, f_PBH≃1.08×10^{-3}, and a regularization-stable candidate O(10^{-2}) bump in P_R(k) near k_peak≃5.7×10^5 Mpc^{-1}. Weak non-Gaussian (skewness) corrections are also explored. The stated goal is to demonstrate feasibility for next-generation detectors at z≳20.","tokens_in":29534,"tokens_out":1688,"duration_ms":30991,"significance":"If the pipeline is robust, it offers a concrete, data-driven route from high-redshift GW catalogs to the small-scale curvature spectrum—scales inaccessible to CMB/LSS—using next-generation detectors (ET, DECIGO, LISA, TianQin/Taiji). Strengths include a carefully derived forward relation (Eq. 19), an explicit uniqueness argument for the discretized inverse under positivity (Appendix A), transparent Tikhonov regularization with L-curve guidance (Appendix B), and public data availability. The high-z feasibility argument is the scientifically durable contribution; the LVK numbers are secondary and correctly flagged by the authors as illustrative rather than a definitive PBH detection.","major_comments":[{"comment":"Abstract and §VII present a “regularization-stable candidate bump-like enhancement of order O(10^{-2}) … centered around k_peak≃5.7×10^5 Mpc^{-1}” as a concrete finding, while §I and §V state that LVK events are not known to be PBHs and that the exercise is only for illustration. The numerical feature is load-bearing for how the paper will be cited. Please either (i) demote the LVK bump to a clearly labeled mock/illustrative result in the abstract and conclusions, or (ii) add a quantitative contamination test: inject a controlled astrophysical fraction into the m>15 M_⊙ sample and show how the reconstructed f(m) peak and P_R bump degrade. Without this, the abstract overstates what the LVK application can support.","section":null},{"comment":"§III assumptions 1–3 (Poisson spatial distribution, common t_dec, negligible late-time mass evolution) and the circular-orbit window of Eqs. (24)–(35) are used both for the method and for the LVK illustration. The text correctly defers accretion, disruption, three-body capture, and eccentricity to future work, but these effects enter the kernel that is inverted. For the feasibility claim aimed at next-generation detectors, a short robustness subsection (or appendix) is needed: e.g., how a 10–30% late-time mass shift or a non-Poisson clustering term biases the recovered f(m) and the subsequent P_R peak. Without it, the mapping from high-z catalogs to P_R remains unquantified under realistic deviations from the three assumptions.","section":null},{"comment":"§VI, after Eq. (50): the reconstruction identifies the collapse fraction β(M) with the abundance in each mass bin, “valid for sharply peaked spectra,” with a more accurate differential/excursion-set treatment left for future work. The reconstructed f(m) is lognormal-like with σ_MF~0.5–0.8 (Table I), not infinitely sharp. This approximation directly sources σ^2(R) and therefore the location and amplitude of the reported P_R bump. Please quantify the bias (e.g., by comparing the bin-identification prescription to dβ/dlnM or a simple excursion-set estimate on the same f(m)), or restrict the PPS claim to the sharply-peaked limit and state the systematic floor on k_peak and β_p.","section":null},{"comment":"§V, Eq. (42): f_PBH is fixed by requiring N_PBH(f_PBH)=174 for the m>15 M_⊙ sample. The shape of f(m) comes from the redshifted-mass inversion, but the overall amplitude that enters β(m) and thus P_R is set by this match. Please report the sensitivity of the reconstructed P_R amplitude to a factor-of-few variation in the adopted merger-rate formula (Eq. 40) and to the choice of which events are counted as PBH candidates, so that the O(10^{-2}) height is not read as uniquely determined by the data.","section":null}],"minor_comments":[{"comment":"Fig. 3 and Table I: the power-law fit α_MF=1.56 is quoted for the combined sample, but the right panel of Fig. 3 shows it only over a limited range; clarify the mass interval used for each fit and whether the reduced χ^2_ν accounts for the full covariance of the 100 reconstruction draws.","section":null},{"comment":"Eq. (19) vs. Eq. (20): the switch from P to P_O / f_p is clear in the text but the notation for the physical vs. assumed mass function (f_p vs. f̃) is easy to lose in later sections; a short notation table or consistent subscripts would help.","section":null},{"comment":"§V: pastro≥0.9 is used as a selection cut; pastro does not distinguish stellar vs. primordial origin. A one-sentence reminder of this limitation next to the cut definition would avoid misreading.","section":null},{"comment":"Fig. 4: the IR non-decaying tail is correctly caveated in the text; adding a vertical band marking the k-range that is actually constrained by the reconstructed mass interval would make the figure self-contained.","section":null},{"comment":"Appendix A: the uniqueness argument is for the equal-mass slice m_1^z=m_2^z and for a fixed regularization operator L and λ. A brief remark that the full two-dimensional (m_1^z,m_2^z) inverse inherits uniqueness under the same positivity and fixed-(L,λ) conditions would close the loop with the gradient-descent procedure of §IV.","section":null},{"comment":"References: GWTC-4/5 arXiv numbers and the 2025–2026 LVK papers are cited; ensure final published versions are updated at proof stage if available.","section":null}],"recommendation":"major_revision","confidential_remarks":"The methodological core (forward map + gradient-descent inverse + regularized Press–Schechter) is solid and appropriate for the journal. The main risk is that the abstract’s LVK bump will be quoted as a primordial detection; requiring the authors to either demote it or add a contamination/robustness test is the right bar. I would not reject: the high-z feasibility claim stands independently of the LVK numbers. No concerns about citation pattern or scope fit."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The real contribution here is the joint pipeline: gradient-descent inversion of the two-dimensional redshifted-mass distribution (their Eq. 19–23) for the PBH mass function, followed by Tikhonov-regularized Press-Schechter inversion for P_R(k). That combination is a genuine step past the earlier reconstruction papers they cite. The forward map is carefully derived, the discretized inverse is shown unique under positivity (App. A), and the L-curve diagnostics plus public data deposit make the method usable by others.\n\nWhat they do well is keep the LVK application explicitly labeled as illustration. Dual mass cuts (m>15 and m>50) give a consistency check in the overlap region, the abundance is fixed only to match the selected event count, and they flag that accretion, disruption, and eccentricity are omitted. The reported O(10^{-2}) bump near 5.7e5 Mpc^{-1} is therefore a proof-of-concept feature under the Sec. III assumptions, not a high-confidence cosmological measurement. The high-z feasibility argument for ET/DECIGO/LISA stands independently of that number.\n\nSoft spots are real but proportionate. The circular-orbit SNR>8 window and the three early-universe assumptions are simplifications; if a non-negligible astrophysical fraction survives the cuts, both the mass-function peak and the PPS bump become artifacts. Mild circularity exists in fitting overall f_PBH, and free parameters (λ, δ_c, γ_m, selection thresholds) are present, but they are transparent rather than hidden. Non-Gaussianity is treated perturbatively and does not change the qualitative picture. Math and citation pattern look solid; no load-bearing contradictions.\n\nThis is for people building high-z PBH pipelines or small-scale inflation probes. It deserves a serious referee. I would engage with the method and cite the framework; I would not cite the specific LVK bump as a result.","headline":"Solid end-to-end inverse pipeline from redshifted BBH masses to small-scale PPS; the LVK bump is only a labeled illustration and should not be over-read.","tokens_in":30190,"tokens_out":492,"would_cite":true,"duration_ms":8160,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"High-redshift binary black holes can be inverted into a candidate bump in the small-scale primordial power spectrum.","keywords":["primordial black holes","primordial power spectrum","gravitational waves","binary black holes","gradient descent","Press-Schechter formalism","Tikhonov regularization","high-redshift BBHs"],"falsifier":"A large, clean sample of binary black holes at z ≳ 20 whose reconstructed mass function shows no peak near tens of solar masses (or whose regularized power spectrum shows no corresponding O(10^{-2}) bump near 5 \times 10^5 Mpc^{-1}) would falsify the candidate feature under the same pipeline.","tokens_in":30014,"feed_emoji":"🔭","tokens_out":973,"duration_ms":9930,"temperature":0.7,"pith_summary":"The paper develops a two-stage reconstruction that turns the redshifted mass pairs of high-redshift binary black holes into the underlying primordial black hole mass function and then into the small-scale primordial power spectrum. Gradient descent first recovers the mass function from the observed redshifted distribution after folding in detector sensitivity and binary redshift distribution; a regularized Press–Schechter inversion then converts that mass function into the curvature power spectrum. Applied as a benchmark to LVK events selected by simple mass cuts, the pipeline produces a lognormal-like mass peak near tens of solar masses and a regularization-stable candidate power-spectrum enhancement of order 10^{-2} centered near 5.7 \times 10^5 Mpc^{-1}. The authors present the LVK exercise as an illustration rather than a definitive detection, arguing that the same chain will become far cleaner once next-generation detectors deliver large samples of binaries at z ≳ 20 where stellar-origin contamination is strongly suppressed. If the method works as claimed, gravitational-wave catalogs will supply a new observational window on primordial fluctuations on scales far smaller than those accessible to the cosmic microwave background.","feed_headline":"GW binaries inverted into a small-scale primordial power bump","feed_subtitle":"Gradient descent plus regularized Press–Schechter recovers an O(10^{-2}) candidate feature near 5.7e5 Mpc^{-1}","key_machinery":"The integral map (Eq. 19/20) that expresses the observed redshifted mass distribution P_O(m_1^z, m_2^z) as a double convolution of the PBH mass function f(m) with the binary-formation weight η, detector window W, and redshift distribution p(z); this map is inverted by gradient descent on a discretized mass vector, after which the Press–Schechter relation plus Tikhonov regularization recovers P_R(k).","core_discovery":"A gradient-descent inversion of the observed redshifted mass distribution of high-redshift binary black holes, followed by Tikhonov-regularized Press–Schechter inversion, yields a regularization-stable candidate bump-like enhancement of order O(10^{-2}) in the primordial power spectrum centered at k_peak ≃ 5.7 \times 10^5 Mpc^{-1} under the adopted assumptions, demonstrating that the reconstruction chain is feasible for next-generation detectors.","pith_inferences":["Once Einstein Telescope or DECIGO catalogs exist, the identical inversion can be run in redshift slices to test whether the candidate bump is redshift-independent, as expected for a primordial feature.","The method supplies a practical route to place joint constraints on both the amplitude and the width of any small-scale power-spectrum feature that produces solar-mass-scale PBHs.","If late-time accretion or eccentricity systematically alter the observed mass distribution, the recovered peak position will drift; monitoring that drift with improved templates would diagnose the size of the neglected effects."],"forward_implications":["Next-generation detectors that accumulate high-redshift binaries can map the small-scale primordial power spectrum without relying on cosmic-microwave-background or large-scale-structure data.","The same pipeline supplies an independent estimate of the total PBH dark-matter fraction once the merger abundance is matched.","Weak non-Gaussian corrections can be folded into the Press–Schechter step, shifting the reconstructed amplitude in a controlled, testable way.","Consistency between independent mass cuts (e.g., m > 15 M_⊙ versus m > 50 M_⊙) becomes a built-in internal check on the reconstructed mass function."],"fun_headline_variants":["Gradient descent inverts high-z BBH masses into small-scale PPS bump","Redshifted BBH distribution yields O(10^{-2}) PPS feature near 5.7e5 Mpc^{-1}","Regularized Press-Schechter recovers candidate PPS enhancement from PBHs","High-redshift binaries map to primordial power spectrum via mass inversion","LVK high-z BBHs reconstruct small-scale PPS bump under PBH selection"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"That the black-hole events kept after simple mass cuts can be treated as a pure primordial sample whose binaries form under the paper’s three simplifying assumptions and whose detection is captured by a circular-orbit signal-to-noise window.","fun_headline_variants_meta":{"raw":{"variants":["Gradient descent inverts high-z BBH masses into small-scale PPS bump","Redshifted BBH distribution yields O(10^{-2}) PPS feature near 5.7e5 Mpc^{-1}","Regularized Press-Schechter recovers candidate PPS enhancement from PBHs","High-redshift binaries map to primordial power spectrum via mass inversion","LVK high-z BBHs reconstruct small-scale PPS bump under PBH selection"]},"model":"grok-4.5","effort":"low","cost_usd":0.006404,"raw_usage":{"total_tokens":1658,"prompt_tokens":790,"num_sources_used":0,"completion_tokens":111,"cost_in_usd_ticks":64040000,"prompt_tokens_details":{"text_tokens":790,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":757,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":790,"tokens_out":111,"duration_ms":7067,"temperature":1.0,"reasoning_tokens":757,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-14T06:44:04.301330+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A large, clean sample of binary black holes at z ≳ 20 whose reconstructed mass function shows no peak near tens of solar masses (or whose regularized power spectrum shows no corresponding O(10^{-2}) bump near 5 \times 10^5 Mpc^{-1}) would falsify the candidate feature under the same pipeline.","supporting_citations":[],"review_version":1}