{"id":"b6f25726-7f38-4652-8e2e-20fe9ad10cac","arxiv_id":"2607.01458","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":8.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"Sharp inequality |A1+⋯+An| ≥ (∏|Ai|)^{1/p} holds with p = n log(m+1)/log(nm+1) for Ai ⊆ {0..m}^d, exponent optimal, obtained from a functional inequality on Z^d.","lead":"The authors prove a sharp lower bound on the size of sumsets A1+...+An for subsets Ai inside the hypercube {0,1,...,m}^d in Z^d. This resolves an explicit conjecture in additive combinatorics via a stronger sup-convolution inequality derived from a mixed-volume representation of a lattice-path norm.","discovery_kind":"unclear","skeptic_critique":{"model":"grok-4.3","headline":"Validity of transferring the mixed-volume bound on the lattice-path norm to discrete sumset cardinalities via the 1D inequality","rationale":"The reader's weakest_assumption correctly flags the novel mixed-volume step and the 1D inequality as the points whose correctness is least externally corroborated. No internal inconsistency or parameter-count issue is visible from the abstract and claim structure; the concern is simply that these two ingredients have not yet been independently stress-tested. Hence the UNVERDICTED status is retained.","tokens_in":1748,"tokens_out":365,"duration_ms":22039,"concrete_test":"Isolate the precise statement of the 1D functional inequality (presumably in §3 or §4) and the definition of the lattice-path norm; substitute the indicator functions of arithmetic-progression sets achieving the conjectured equality case for m=2, n=3 and recompute both sides of the resulting sup-convolution inequality to check whether equality holds exactly.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The argument obtains the sumset lower bound as a corollary of a sup-convolution inequality proved via a novel mixed-volume representation of the lattice-path norm plus a sharp 1D functional inequality on Z. The representation equates a certain norm to a mixed volume; the 1D inequality is then applied to marginals or slices. It is not immediate that this yields a strict lower bound on |A1+⋯+An| without an intermediate discretization step or approximation argument whose error term vanishes in the claimed regime. The abstract states the exponent is best possible, but the equality cases must arise exactly from the 1D construction; any mismatch in how the mixed volume interacts with the support constraints {0,…,m}^d would weaken the sharpness claim.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper proves sharp lower bounds on sumset cardinalities for subsets A_j of the hypercube {0,1,...,m}^d in Z^d: |A1+⋯+An| ≥ (|A1|⋯|An|)^{1/p} with p = n log(m+1)/log(nm+1), and shows the exponent is optimal. It also treats the case of distinct m_j. The result is obtained as a corollary of a stronger sup-convolution inequality for functions on Z^d, derived from a novel mixed-volume representation of the lattice-path norm combined with a sharp one-dimensional functional inequality on Z.","tokens_in":1910,"tokens_out":545,"duration_ms":13402,"significance":"If correct, the result resolves a folklore conjecture in additive combinatorics with the first sharp bounds for general m beyond the cases m=1 (all n) and m=2 (n=2). The mixed-volume approach to the lattice-path norm constitutes a methodological advance with potential for further applications in discrete geometry and inequalities.","major_comments":[{"comment":"The central claim rests on transferring the mixed-volume representation of the lattice-path norm to a strict lower bound on discrete sumset cardinalities via the 1D functional inequality. The manuscript must supply an explicit discretization or approximation argument (with error term shown to vanish in the relevant regime) to confirm that the inequality passes to the indicator functions of the sets A_j without loss of sharpness; this step is load-bearing for both the inequality and the optimality statement.","section":"Proof of the sup-convolution inequality (via mixed-volume representation and 1D inequality)"},{"comment":"Equality cases are asserted to arise exactly from the 1D construction. The interaction between the mixed-volume representation and the support constraints {0,…,m}^d must be verified to ensure that the extremal examples in the hypercube achieve equality without additional discretization error; otherwise the sharpness claim requires adjustment.","section":"Sharpness / equality cases"}],"minor_comments":[{"comment":"Clarify the precise statement of the one-dimensional functional inequality (including the class of functions to which it applies) and its proof, as this is invoked as a black box for the higher-dimensional result.","section":null},{"comment":"Add a short comparison table or paragraph contrasting the new exponent with the previously known sharp cases (m=1 and m=2, n=2) to highlight the improvement.","section":null}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the careful reading and positive assessment of the significance of the results. We address the major comments point by point below.","responses":[{"response":"The mixed-volume representation is developed directly for the lattice-path norm on the discrete group Z^d, and the sup-convolution inequality is proved for arbitrary nonnegative functions on Z^d. The one-dimensional functional inequality holds exactly on Z. Indicator functions of subsets A_j ⊆ {0,…,m}^d are admissible inputs, so the passage to the sumset cardinality bound is immediate and exact, with no approximation or discretization step required. We will insert a short clarifying paragraph after the statement of the functional inequality to make this direct applicability explicit.","revision_made":"partial","referee_comment":"[Proof of the sup-convolution inequality (via mixed-volume representation and 1D inequality)] The central claim rests on transferring the mixed-volume representation of the lattice-path norm to a strict lower bound on discrete sumset cardinalities via the 1D functional inequality. The manuscript must supply an explicit discretization or approximation argument (with error term shown to vanish in the relevant regime) to confirm that the inequality passes to the indicator functions of the sets A_j without loss of sharpness; this step is load-bearing for both the inequality and the optimality statement."},{"response":"The extremal configurations are obtained by taking identical one-dimensional extremal sets in each coordinate; these sets lie inside {0,…,m} and therefore inside the hypercube. Because both the mixed-volume identity and the one-dimensional inequality are exact equalities on Z, the resulting functions on Z^d attain equality in the sup-convolution inequality with no error term. We will add an explicit verification of equality attainment (including the coordinate-wise construction) in the revised manuscript.","revision_made":"yes","referee_comment":"[Sharpness / equality cases] Equality cases are asserted to arise exactly from the 1D construction. The interaction between the mixed-volume representation and the support constraints {0,…,m}^d must be verified to ensure that the extremal examples in the hypercube achieve equality without additional discretization error; otherwise the sharpness claim requires adjustment."}],"tokens_in":1468,"tokens_out":481,"duration_ms":25104,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"Gonçalves and Radchenko settle the conjecture on the sharp lower bound for |A1 + ⋯ + An| when each Aj lives in {0,…,m}^d. They obtain |sum| ≥ (∏|Aj|)^{1/p} with p = n log(m+1)/log(nm+1) and show the exponent is best possible. The only prior sharp cases were m=1 for any n and m=2 with n=2; this covers arbitrary m and n, plus the unequal-mj version.\n\nThe proof route is a stronger sup-convolution inequality derived from a mixed-volume representation of the lattice-path norm, followed by a sharp one-dimensional functional inequality on Z. That representation is new for this problem and avoids the usual parameter-fitting or circular reductions.\n\nThe argument looks direct: the mixed volume gives the norm bound, the 1D inequality is applied to slices or marginals, and the cardinality statement drops out as a corollary. Equality cases are expected to arise exactly from the 1D construction, which aligns with the sharpness claim.\n\nThe main point to check is whether the transfer from the mixed-volume representation to the exact discrete sumset cardinality is free of approximation or discretization error that could affect the constant. The abstract presents the step as exact, so if the 1D inequality holds with equality on the relevant functions and the support constraints are respected, the bound stands. No circularity or invented entities appear in the setup.\n\nThis is for people working on additive combinatorics in restricted domains or on discrete Brunn-Minkowski analogues. A reader who cares about optimal constants in sumset estimates will get the most from it.\n\nThe paper deserves a serious referee. It resolves an open conjecture with an optimal result and supplies a technique worth testing on related problems.","headline":"Gonçalves and Radchenko prove the general sharp sumset bound in hypercubes via mixed volume plus a 1D inequality, closing the conjecture.","tokens_in":2360,"tokens_out":446,"would_cite":true,"duration_ms":13531,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Sumsets of n subsets inside an m-hypercube obey a sharp lower bound |A1+⋯+An| ≥ (|A1|⋯|An|)^{1/p} with p = n log(m+1)/log(nm+1).","keywords":["sumsets","hypercubes","additive combinatorics","lower bounds","mixed volumes","lattice paths","sup-convolution"],"falsifier":"Explicit sets A_j inside {0,…,m}^d whose sumset has cardinality strictly smaller than the right-hand side for the stated p would falsify the claim.","tokens_in":2651,"feed_emoji":"","tokens_out":734,"duration_ms":15725,"temperature":0.7,"pith_summary":"The paper proves a lower bound on the size of the sumset formed by n subsets of the integer lattice each confined to the hypercube with coordinates running from 0 to m. The bound states that the cardinality of the sum is at least the product of the individual cardinalities raised to the power 1/p, where p is defined as n times log of (m+1) divided by log of (nm+1), and this exponent cannot be improved. The result resolves a conjecture that had been circulating in additive combinatorics and extends the only previously known sharp cases (binary hypercubes for any n, and ternary for n=2). The proof proceeds by establishing a stronger inequality for the sup-convolution of functions on Z^d and then specializing to indicator functions.","feed_headline":"Hypercube sumsets obey |A1+⋯+An| ≥ product of sizes to power 1/p","feed_subtitle":"The exponent p = n log(m+1)/log(nm+1) is optimal for subsets of {0..m}^d and resolves a long-standing conjecture.","key_machinery":"Mixed-volume representation of the lattice-path norm together with a sharp one-dimensional functional inequality on Z.","core_discovery":"For any sets A_j ⊆ {0,1,…,m}^d the inequality |A1+⋯+An| ≥ (|A1|⋯|An|)^{1/p} holds with p = n log(m+1)/log(nm+1), and the exponent is best possible. The same conclusion holds when the upper limits m_j are allowed to differ across the sets. This follows from a stronger functional inequality on sup-convolutions that is proved via a mixed-volume representation of the lattice-path norm together with a sharp one-dimensional inequality.","pith_inferences":["The mixed-volume approach may extend to other discrete Brunn-Minkowski-type statements on product sets.","Small explicit examples in low dimension and small m can be checked computationally to verify the bound saturates.","The functional form may yield new inequalities for the support size of convolutions of arbitrary nonnegative functions rather than just indicators."],"forward_implications":["The bound remains valid when each set A_j is confined to its own hypercube {0,…,m_j}^d.","The exponent p cannot be replaced by any smaller number while keeping the inequality true for all choices of sets.","The inequality specializes to the known sharp results when m=1 for any n and when m=2 and n=2."],"fun_headline_variants":["Hypercubes yield optimal sumset bound with exponent p","Sharp sumset bound |A1+...+An| >= sizes to power 1/p","Proved sharp lower bound for sumsets confined to hypercubes","Hypercube subsets satisfy best possible sumset inequality"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"A sharp one-dimensional functional inequality holds for the relevant functions on the integers.","fun_headline_variants_meta":{"raw":{"variants":["Hypercubes yield optimal sumset bound with exponent p","Sharp sumset bound |A1+...+An| >= sizes to power 1/p","Proved sharp lower bound for sumsets confined to hypercubes","Hypercube subsets satisfy best possible sumset inequality"]},"model":"grok-4.3","cost_usd":0.006851,"raw_usage":{"total_tokens":3219,"prompt_tokens":742,"num_sources_used":0,"completion_tokens":73,"cost_in_usd_ticks":68512000,"prompt_tokens_details":{"text_tokens":742,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2404,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":742,"tokens_out":73,"duration_ms":16921,"temperature":1.0,"reasoning_tokens":2404,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-03T19:52:55.429046+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"Explicit sets A_j inside {0,…,m}^d whose sumset has cardinality strictly smaller than the right-hand side for the stated p would falsify the claim.","supporting_citations":[],"review_version":1}