{"id":"a583842c-9feb-461e-ac58-0db54b9d7817","arxiv_id":"2508.14698","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":8.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"For d>=3, typical homogeneous self-similar measures have positive Fourier dimension: all but a zero-dimensional set of contraction ratios yield power Fourier decay under spanning or affine irreducibility assumptions.","lead":"This mathematics paper proves that self-similar measures generated by contracting similarities in three or more dimensions typically have power-law Fourier decay, which implies they are absolutely continuous in a super-critical parameter range. It extends prior low-dimensional results to d>=3 with general orthogonal rotations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Uniformity of the transversality estimate over all O and p is the load-bearing assumption; the abstract does not establish it.","rationale":"The reader's weakest_assumption identifies the same potential gap: the proof of typicality over λ must rely on a transversality or non-concentration estimate that is uniform for the digit set D and the rotation O, and the abstract does not state the needed separation or non-degeneracy conditions. This is exactly the point on which the central claim could fail. Because the full text is not available, the appropriate verdict remains UNVERDICTED; a concrete check is to inspect the transversality lemma and test it on the extremal case O=I with D={0,e1,...,ed}. I do not see grounds to reject or accept on the abstract alone, so the reader's UNVERDICTED verdict is unchanged.","tokens_in":886,"tokens_out":11970,"duration_ms":151918,"concrete_test":"Locate the manuscript's key transversality lemma (the one controlling products of the digit Fourier sums over a scale-invariant set of frequencies). Check whether the hypotheses include 'O has no invariant subspace' or 'D is not contained in a proper affine subspace' in the first theorem. If they do, apply the lemma to the legitimate case O=I, D={0,e1,...,ed}, p uniform: if the hypotheses fail, the theorem is overstated as written. If the lemma is stated for all spanning D and all O, verify the proof by re-deriving the known one-dimensional Shmerkin-type bound in the coordinate-axis marginal; the polynomial decay should follow with a zero-Hausdorff-dimension exceptional set. If that marginal derivation breaks under the one-hot digit structure, the uniform claim collapses.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The first theorem quantifies over every orthogonal O and every positive probability vector p, and claims power Fourier decay for all λ outside a zero-Hausdorff-dimension set. The proof must therefore contain a non-concentration/transversality estimate for the exponential sums Σ_j p_j e^{2πi⟨a_j, O^n ξ⟩} that is uniform in O and p and valid whenever D spans R^d. The weakest spot is precisely this uniformity. If the estimate requires O to be irreducible, to have infinite order, or the digit set to be in 'general position' beyond spanning, then the theorem as stated overreaches. In particular, O=I and D={0,e1,...,ed} with arbitrary p is a legitimate instance of the hypotheses, and the Fourier transform is a product over scales of φ(λ^n ξ) with a one-digit-per-coordinate step. The claimed decay for all but a zero-Hausdorff set of λ is a strong assertion about this product and is not justified by the abstract. Since the full proof is unavailable, this hidden uniformity condition is the most load-bearing concern.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper (arXiv:2508.14698) studies homogeneous self-similar measures in R^d, d >= 3, generated by IFS maps f_j(x) = lambda O x + a_j with a_0 = 0, O orthogonal, and digits D = {a_0,...,a_m}. For a positive probability vector p, the associated self-similar measure is denoted mu(lambda O, D, p). Two theorems are claimed. Theorem 1: if D spans R^d, then for every fixed orthogonal O and probability vector p, the measure has power Fourier decay (positive Fourier dimension) for all lambda outside a zero-Hausdorff-dimension subset of (0,1). Theorem 2: for even d >= 4, under only the necessary affine irreducibility condition, power Fourier decay holds for almost all homogeneous self-similar measures. The abstract also states that these results, combined with work of Corso and Shmerkin, imply absolute continuity in the super-critical region. The review is based on the abstract only, as the full text was not available.","tokens_in":1119,"tokens_out":3613,"duration_ms":48669,"significance":"If the theorems are correct, they represent substantial progress in the Fourier-analytic theory of self-similar measures. In particular, the first theorem would show that power Fourier decay is typical in lambda for a very general class of higher-dimensional homogeneous IFSs, with no separation or overlap assumptions beyond spanning of the digit set. The second theorem is notable for relaxing digit-set assumptions under affine irreducibility. The stated connection to absolute continuity is also significant. The paper appears to contain no fitted parameters or ad hoc numerical ingredients; the claims are precise mathematical statements. However, because the proof is not available for inspection, the soundness of the central claims cannot currently be verified.","major_comments":[{"comment":"The theorem quantifies over every orthogonal O and every probability vector p, with only D spanning R^d. The reader's stress-test concern about uniformity over O and p does not land exactly as stated, because O and p are fixed before the exceptional set of lambda is chosen. However, the more precise concern remains: the proof must handle degenerate rotations such as O = I. With O = I and D = {0, e_1, ..., e_d}, the Fourier transform becomes a product over scales, and the claimed power decay for all but a zero-Hausdorff-dimension set of lambda is a strong assertion. The abstract does not state the transversality or non-concentration lemma that must control such cases, nor any separation/overlap condition. Please include the relevant lemma and explicitly verify that the hypotheses cover O = I and arbitrary p.","section":"Abstract, Theorem 1"},{"comment":"The second theorem is stated too loosely: 'power Fourier decay for almost all homogeneous self-similar measures' under affine irreducibility. It is not specified which parameters are random (lambda? O? p? D?) nor the underlying probability measure on parameter space. The quantifier structure is load-bearing for both the mathematical claim and the subsequent absolute-continuity application. Please state the theorem with precise hypotheses on the probability space of parameters and the meaning of 'almost all.'","section":"Abstract, Theorem 2"},{"comment":"The abstract states that the results imply absolute continuity in the super-critical parameter region when combined with Corso and Shmerkin. The 'super-critical parameter region' is not defined, and it is not clear whether the parameter space includes lambda, O, D, and p jointly or only lambda. Since this consequence is one of the headline applications, the precise region and the mechanism by which the cited work converts Fourier decay into absolute continuity should be stated in the introduction or abstract.","section":"Abstract, absolute-continuity consequence"}],"minor_comments":[{"comment":"Please define 'affine irreducibility' explicitly and state why it is necessary, especially for readers coming from the one-dimensional self-similar measures literature.","section":"Abstract, notation"},{"comment":"The phrase 'zero-Hausdorff dimension set of lambda' is potentially ambiguous; specify that it means a set of lambda in (0,1) with Hausdorff dimension zero.","section":"Abstract, terminology"},{"comment":"The abstract says d >= 3 but the second theorem only covers even d >= 4. The parity restriction should be explained at least in the introduction, even if the proof requires it.","section":"General"},{"comment":"The reference to Corso and Shmerkin [arXiv:2409.04608] should include the full bibliographic information and the precise statement of the result being combined with the present theorems.","section":"References"},{"comment":"Since the paper is about homogeneous self-similar measures, clarify whether the standard open-set condition or weak separation condition is ever assumed; the abstract suggests not, which is a strength, but this should be explicit.","section":"General"}],"recommendation":"uncertain","confidential_remarks":"The manuscript appears potentially significant, but the available material is only the abstract. I cannot verify the proofs or the precise hypotheses. I recommend that the editor obtain the full manuscript; in particular, the O = I test case and the exact quantifier structure of Theorem 2 should be checked by a specialist. There is no indication of circularity or fitted parameters."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this looks like the real next step in the typical Fourier decay story for self-similar measures, moving to d>=3 and allowing arbitrary rotations. The two theorems are precisely stated, and the connection to absolute continuity via Corso-Shmerkin is a nice payoff. The main thing to check in the proof is whether the transversality estimate is genuinely uniform over all orthogonal O and all probability vectors p, as the first theorem claims.\n\nWhat's new: for fixed O and p, if D spans R^d, the measure has power Fourier decay for all λ outside a zero-Hausdorff-dimension set. That's strong and, as far as I know, new for d>=3. The second result relaxes the spanning condition to affine irreducibility but only gets almost all (Lebesgue) and even d>=4. That's an odd restriction but not a defect. No fitted parameters, no circularity; it's a theorem about typical parameters.\n\nThe soft spot: the abstract doesn't show the proof, so we can't verify the uniformity over O and p. The stress-test example O=I with D={0,e_1,...,e_d} is a legitimate test case; the theorem says decay for all but a zero-Hausdorff set of λ, so the proof needs a non-concentration estimate that holds with no Diophantine condition on λ beyond typicality. If that estimate exists, the result is solid. I don't see an obvious counterexample, and Solomyak knows the transversality toolbox, so my prior is that the theorem is correct. But this is exactly the place a referee should push.\n\nThe even-dimension restriction in the second theorem is unexplained in the abstract; likely a technical constraint in the argument. It doesn't undermine the first theorem.\n\nVerdict: worth a serious referee, and worth discussing in a reading group once the proof is available. I'd send it to review; the claims are specific enough that a competent referee can check the uniformity estimate in a few hours.\n\nWho this is for: people working on Fourier dimension, self-similar measures, and absolute continuity of self-similar measures. It's a meaningful advance, though not a full resolution of the absolute continuity conjecture.","headline":"A likely solid extension of typical Fourier decay to d>=3; the main thing to verify is uniformity over rotations and weights in the first theorem.","tokens_in":1457,"tokens_out":3075,"would_cite":true,"duration_ms":31904,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["28A80","42B10","37C45"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that, for d≥3, typical homogeneous self-similar measures in R^d have power Fourier decay, with an exceptional set of contraction ratios of Hausdorff dimension zero when the digit set spans R^d.","keywords":["Fourier decay","self-similar measures","homogeneous iterated function systems","Hausdorff dimension","absolute continuity","affine irreducibility","positive Fourier dimension","super-critical region"],"falsifier":"Fix a spanning digit set, an orthogonal rotation, and positive weights; for contraction ratios drawn from a Cantor set of positive Hausdorff dimension, compute the Fourier transform of the invariant measure along a growing sequence of frequencies. If the values fail to decay polynomially for all such ratios, the zero-Hausdorff-dimension exceptional-set claim is refuted.","tokens_in":840,"feed_emoji":"📐","tokens_out":7819,"duration_ms":99295,"temperature":0.7,"pith_summary":"This paper studies measures generated by affine iterated function systems on R^d for d≥3, where each step applies the same contraction and a fixed orthogonal rotation, then adds one of finitely many digit vectors with a fixed probability. The first result shows that whenever the digit set spans R^d, and the rotation and probabilities are held fixed, the invariant measure has power Fourier decay for all contraction ratios except a set of Hausdorff dimension zero. The second result relaxes the spanning condition to affine irreducibility for even d≥4, still yielding power Fourier decay for almost all parameters. Such decay means the measure has positive Fourier dimension, and the paper concludes, with a recent companion result, that these typical measures are absolutely continuous in the super-critical parameter region.","feed_headline":"Typical self-similar measures in R^d get positive Fourier dimension","feed_subtitle":"A zero-Hausdorff-dimension set of contraction ratios is the only exception; even with overlaps, affine irreducibility suffices in even d≥4.","key_machinery":"The central object is the homogeneous self-similar measure μ(λO,D,p), defined as the unique probability measure satisfying μ = Σ p_j μ ∘ (λO x + a_j)^{-1}. The proof analyses the Fourier transform of this measure along scales set by the contraction; polynomial decay arises from cancellations in the exponential sums over the digit vectors. Typicality is measured by the Hausdorff dimension of the exceptional parameter set, which is how the paper makes precise that non-decay is negligible.","core_discovery":"The central claim is that the obstructions to Fourier decay for homogeneous self-similar measures in high dimensions are extremely rare in parameter space. Theorem 1 states: fix any digit set that spans R^d, any orthogonal matrix, and any positive probability vector; the set of contraction ratios λ for which the self-similar measure lacks power Fourier decay has Hausdorff dimension zero. Theorem 2 replaces the spanning condition by the weaker affine irreducibility, for even d≥4, and concludes power Fourier decay for almost all homogeneous self-similar measures. Combined with a known reduction, these results imply absolute continuity of such measures in the super-critical region.","pith_inferences":["The paper leaves open whether the even-dimension restriction in the second theorem is a technical artifact; if the parity-dependent estimates can be replaced by a symmetric argument, affine irreducibility alone would also settle odd d≥3.","The zero-Hausdorff-dimension exceptional set may be uncountable, so the theorem permits non-decaying examples on a Cantor-like scale; deciding whether such sets are dense in the parameter space would sharpen the typicality statement.","A concrete extension would be to instantiate a small spanning digit set in R^3 and numerically estimate the Fourier transform for contraction ratios drawn from a positive-Hausdorff-dimension Cantor set; observing a positive-dimensional block of non-decay would refute the first theorem."],"forward_implications":["For any fixed spanning digit set, rotation, and probability weights, the contraction ratios that fail to produce positive Fourier dimension form a set of Hausdorff dimension zero, so non-decay is negligible in a strong dimensional sense.","In even d≥4, overlapping digit configurations that are affine irreducible still produce typical positive Fourier dimension, meaning overlaps alone do not destroy decay.","In the super-critical region, the typical self-similar measures are absolutely continuous with respect to d-dimensional Lebesgue measure, not singular fractal distributions.","Power Fourier decay gives quantitative control of equidistribution and convolution, so these measures inherit smooth statistical behaviour at large scales."],"supporting_citations":[],"fun_headline_variants":["Most self-similar measures in R^d have positive Fourier dimension","Fourier decay for typical self-similar measures in high dimensions","Rare exceptions: Fourier decay holds for almost all contraction ratios","Self-similar measures: power Fourier decay except on a zero-Hausdorff set","Even with overlaps, affine irreducibility yields Fourier decay in even d"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The digit set must be genuinely d-dimensional—spanning in the first theorem, or affine irreducible in the second—because if the digits lie in a proper affine subspace the measure is trapped there and cannot have full-dimensional Fourier decay in R^d.","fun_headline_variants_meta":{"raw":{"variants":["Most self-similar measures in R^d have positive Fourier dimension","Fourier decay for typical self-similar measures in high dimensions","Rare exceptions: Fourier decay holds for almost all contraction ratios","Self-similar measures: power Fourier decay except on a zero-Hausdorff set","Even with overlaps, affine irreducibility yields Fourier decay in even d"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000133,"raw_usage":{"total_tokens":1037,"prompt_tokens":874,"completion_tokens":163,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":618,"completion_tokens_details":{"reasoning_tokens":68}},"tokens_in":618,"tokens_out":163,"duration_ms":2985,"temperature":1.0,"reasoning_tokens":68,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T18:18:34.788442+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Fix a spanning digit set, an orthogonal rotation, and positive weights; for contraction ratios drawn from a Cantor set of positive Hausdorff dimension, compute the Fourier transform of the invariant measure along a growing sequence of frequencies. If the values fail to decay polynomially for all such ratios, the zero-Hausdorff-dimension exceptional-set claim is refuted.","supporting_citations":[],"review_version":1}