{"id":"d9cbc7ae-038a-418d-8b46-13c9be25c38a","arxiv_id":"2607.10971","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"For q≥4 the lattice-point error in Cygan-Koranyi balls on the Heisenberg group satisfies |E_q(t)| ≲ t^{2q-1+241/753}, improving Gath's exponent 1/3 and recovering his q=3 bound up to a log factor.","lead":"The paper improves the error bound for counting lattice points inside Cygan-Koranyi balls on Heisenberg groups of dimension 2q+1 (q≥4) from t to the power 2q-1+1/3 down to 2q-1+241/753. This is the first advance on Gath's conjecture that the true order is exactly 2q-1, obtained by a simpler analytic method that reduces the problem to exponential sums controlled by van der Corput tests.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The reader correctly isolates the derivative bounds as the only technical point that requires care, yet those bounds follow by elementary calculus from an explicit phase and are handled by the (5,6)-pair exactly as claimed. The reduction via Landau’s formula (Lemma 2.1) and the subsequent Stečkin conversion are standard and free of gaps. Consequently the improvement |E_q(t)|≪t^{2q-1+241/753} for q≥4 is rigorously established, the q=3 recovery up to a logarithm is correctly noted, and no adjustment of the ACCEPT verdict is warranted.","tokens_in":17237,"tokens_out":472,"duration_ms":4295,"concrete_test":"Independently recompute the sixth-derivative lower bound on eU_0^{(1)} and eU_0^{(3)} (the intervals where the fifth derivative vanishes) and re-balance W=L^{257/753}x^{128/753} in the Stečkin estimate; if the resulting exponent for the Ψ-sum is still 497/1506, the claimed bound for q≥4 is confirmed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Theorem 1.2 rests on a transparent reduction (Landau’s formula + slicing) to Ψ-sums that are then converted by Stečkin’s inequality into exponential sums controlled by the 5th/6th van der Corput tests. The only potentially delicate step is the uniform lower bounds on |F_m^{(5)}| and |F_m^{(6)}| on the dyadic pieces U_j and the three subintervals of U_0 (after (2.16) and Figure 1). These bounds are obtained by direct differentiation of the explicit dual phase F_m(s)=x^{1/2}(m_L^2+s^2)^{1/2} and are standard; the case distinctions precisely cover the zeros of the fifth derivative by switching to the sixth. No hidden assumption, free parameter or circularity appears, and the resulting exponent 241/753 is correctly computed from the balancing that produces (2.23). The argument is therefore self-contained and correct as written.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The paper studies the lattice-point error E_q(t) for Cygan–Korányi balls on the Heisenberg group H^q. After slicing and applying Landau’s formula for the Euclidean ball discrepancy in dimension 2q, the authors reduce E_q(t) to a family of 1-periodic Ψ-sums of the form ∑ Ψ(L^{-1}√(x-n^{2})+R). Stečkin’s inequality converts these into exponential sums; the B-process produces dual phases F_m whose fifth and sixth derivatives are controlled on carefully chosen dyadic pieces (including three subintervals of the critical piece U_0). Van der Corput’s 5th/6th derivative tests then yield the bound |E_q(t)| ≲ t^{2q-1+241/753} for every integer q≥4 (t≥10), while for q=3 the same argument recovers Gath’s exponent 16/3 up to a logarithmic factor. The resulting exponent is obtained by explicit balancing of the terms that arise from the derivative tests.","tokens_in":17419,"tokens_out":770,"duration_ms":6055,"significance":"Gath’s conjecture asserts that the optimal order of E_q is 2q-1 for q≥3. The only previous upper bound for q≥3 was Gath’s own O(t^{2q-1+1/3}). The present work supplies the first improvement of that exponent (to 2q-1+241/753≈2q-1+0.320) for all q≥4, and does so by a comparatively elementary method that relies only on classical tools (Landau, Stečkin, van der Corput). The reduction also makes transparent the intimate link with the Gauss circle problem, suggesting that further progress on the latter (e.g., via the Bombieri–Iwaniec method) would immediately improve the Heisenberg exponent. The argument is fully explicit, free of free parameters, and self-contained.","major_comments":[],"minor_comments":[{"comment":"In the abstract and Theorem 1.2 the two equivalent writings of the exponent (2q-2+994/753 and 2q-1+241/753) are both correct, but a single consistent form should be chosen throughout the paper to avoid momentary confusion.","section":null},{"comment":"Figure 1 is helpful, yet the caption and the surrounding text never state the precise range of the horizontal axis (s/m_L). Adding that information would make the zero of the fifth derivative immediately visible.","section":null},{"comment":"After (2.11) the authors discard the case 2^{j/2}m L^{-1}<1 by a trivial bound; a one-line remark that the same bound is absorbed into the error term of the B-process would make the logic slightly cleaner.","section":null},{"comment":"Several arXiv preprints of the first author are cited as 2026; if any have since appeared in print, the published references should be substituted.","section":null},{"comment":"Typographical: “Derivation Test” appears once in Remark 1.3(1) instead of “Derivative Test”; “forgoing claim” on p. 9 should be “foregoing claim”.","section":null}],"recommendation":"accept","confidential_remarks":"The manuscript is short, cleanly written, and constitutes genuine progress on a conjecture of Gath. I see no reason to delay acceptance; the minor points listed above can be handled at the proof stage."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This paper gives a genuine but small advance on the Cygan–Korányi lattice-point problem for q≥4. The new exponent is 2q-1+241/753 instead of Gath's 2q-1+1/3; for q=3 they recover the same power with an extra log. That is the whole result.\n\nWhat is new is the route. They slice differently, feed the Euclidean ball error into Landau's classical formula, and reduce everything to a family of Ψ-sums that look like the sawtooth sums in the Gauss circle problem. Stečkin turns those into exponential sums; the dual phases after the B-process are handled by the 5th and 6th van der Corput tests on carefully chosen dyadic pieces so that the zeros of the fifth derivative are covered by the sixth. The calculation is fully written out, the balancing that produces 497/1506 (and then 241/753) is transparent, and the link to the circle problem is made explicit. The method is simpler than Gath's restricted-slicing-plus-weighted-annuli argument.\n\nThe soft spots are minor and already flagged by the authors. The derivative lower bounds are standard but require case distinctions near the zeros (Figure 1); nothing is hidden. For q=3 the outer sum loses a log, and for q=2 the method does not improve existing bounds. The new exponent is still far from the conjectured 2q-1, and the paper does not claim otherwise. Citations are clean; the only self-references are to earlier related work by the first author.\n\nThis is for people who already care about lattice-point discrepancies on nilpotent groups or about classical exponential-sum techniques applied to non-Euclidean gauges. It is not a breakthrough, but it is a correctly executed, self-contained improvement that advances an explicit conjecture. I would send it to a referee without hesitation.","headline":"Solid, modest improvement on Gath's exponent for q≥4 via a cleaner reduction to classical exponential sums; first real progress on the conjecture, nothing more.","tokens_in":18051,"tokens_out":509,"would_cite":true,"duration_ms":4765,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11P21","43A80","22E40","11L03"],"pacs":[],"model":"grok-4.5","headline":"The lattice-point error for Cygan–Korányi balls on Heisenberg groups drops below the previous 1/3-exponent for every dimension q ≥ 4.","keywords":["lattice points","Heisenberg group","Cygan–Korányi ball","Gauss circle problem","van der Corput estimates","Landau formula","exponential sums"],"falsifier":"An explicit numerical computation of E_q(t) for a fixed q ≥ 4 and a sequence of large t that grows faster than t^{2q−1 + 241/753} would immediately disprove the claimed upper bound.","tokens_in":18099,"feed_emoji":"🔢","tokens_out":732,"duration_ms":5779,"temperature":0.7,"pith_summary":"Counting integer points inside Cygan–Korányi balls on the Heisenberg group is the natural non-commutative analogue of the classical sphere problem. Earlier work reduced the error to size roughly t^{2q−1 + 1/3}. This paper improves the exponent for every dimension q ≥ 4 by combining a slicing formula with Landau’s classical lattice-point expansion and van der Corput’s fifth- and sixth-derivative tests. The new error is O(t^{2q−1 + 241/753}), which is the first concrete advance toward the conjectured optimal order 2q−1. For the transitional case q = 3 the same method recovers the previous bound up to a single logarithm. The argument also makes plain that the problem is essentially as hard as the two-dimensional Gauss circle problem, so further progress is expected to follow any improvement of the circle problem.","feed_headline":"Heisenberg lattice-point error drops below the 1/3 barrier","feed_subtitle":"First progress toward the conjectured optimal order 2q−1 for Cygan–Korányi balls","key_machinery":"A new exact formula for E_q(t) obtained by slicing the Heisenberg ball into Euclidean balls and inserting Landau’s formula for the Euclidean lattice-point error; the resulting family of 1-periodic Ψ-sums is converted by Stečkin’s inequality into exponential sums that are estimated by the simultaneous fifth- and sixth-derivative tests of van der Corput.","core_discovery":"For every integer q ≥ 4 and all t ≥ 10 the lattice-point discrepancy E_q(t) of a Cygan–Korányi ball of radius t satisfies |E_q(t)| ≤ C_q t^{2q−2 + 994/753} (equivalently t^{2q−1 + 241/753}). When q = 3 the same method yields the slightly weaker bound O(t^{16/3} log t). Both estimates improve, or match up to a logarithm, the best previous results and constitute the first progress toward Gath’s conjecture that the true order is 2q−1.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Heisenberg lattice error drops to t^{2q-1+241/753} for q≥4","Cygan-Koranyi balls: lattice discrepancy below 1/3 barrier","First progress toward Gath 2q-1 conjecture on Heisenberg groups","Landau and van der Corput push Heisenberg lattice bound past 1/3","Improved E_q(t) ≪ t^{2q-1+241/753} for Cygan-Koranyi balls"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The fifth and sixth derivatives of the dual phase must stay uniformly bounded away from zero on each dyadic piece of the summation range; near the zeros of the fifth derivative this lower bound is delicate and must be checked by hand.","fun_headline_variants_meta":{"raw":{"variants":["Heisenberg lattice error drops to t^{2q-1+241/753} for q≥4","Cygan-Koranyi balls: lattice discrepancy below 1/3 barrier","First progress toward Gath 2q-1 conjecture on Heisenberg groups","Landau and van der Corput push Heisenberg lattice bound past 1/3","Improved E_q(t) ≪ t^{2q-1+241/753} for Cygan-Koranyi balls"]},"model":"grok-4.5","effort":"low","cost_usd":0.01331,"raw_usage":{"total_tokens":2921,"prompt_tokens":958,"num_sources_used":0,"completion_tokens":121,"cost_in_usd_ticks":133100000,"prompt_tokens_details":{"text_tokens":958,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1842,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":958,"tokens_out":121,"duration_ms":13768,"temperature":1.0,"reasoning_tokens":1842,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-14T07:56:46.410644+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"An explicit numerical computation of E_q(t) for a fixed q ≥ 4 and a sequence of large t that grows faster than t^{2q−1 + 241/753} would immediately disprove the claimed upper bound.","supporting_citations":[],"review_version":1}