{"id":"90f631aa-ce85-49bb-b27b-c968498b6da8","arxiv_id":"2608.01040","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Every centered convex body in R^n with only the origin as an interior lattice point and volume (n+1)^n/n! is unimodularly equivalent to the centered standard simplex.","lead":"A new proof shows that if a convex body in n dimensions has its barycenter at the origin, contains no other lattice point in its interior, and has exactly the maximal possible volume, then it must be a unimodular image of the standard simplex. This settles the equality case of Ehrhart's volume conjecture, completing a proof that was announced for the inequality side in August 2026.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The arithmetic half rests on a false theorem: determinant-one lattices avoiding the open unit cube need not be upper unitriangular (e.g., Λ=Z(2,0)⊕Z(1,1/2) in R^2), so Theorem 7.2's proof collapses and Theorem 1.3 is not established by this paper.","rationale":"The reader's weakest assumption targeted the imported analytic lemmas from [OAI26], but the clearest load-bearing failure is in the arithmetic half, which is the paper's own contribution and is essential for the final unimodularity conclusion. The false invocation of Hajós's theorem is not a matter of unverified externality: the statement as used has an explicit two-dimensional counterexample. Even granting every analytic lemma, Proposition 6.4 only yields an invertible linear map A with |det A|=1 and K=AS_n; the conclusion A∈GL_n(Z) depends entirely on Theorem 7.2, whose proof is invalid. The central theorem of the paper may still be true, and the analytic half may be repairable or even sound, but as written the proof does not establish the equality classification. I therefore recommend REJECT rather than CONDITIONAL: the gap is concrete and central, not a missing regularity hypothesis or an unproved external lemma that might be supplied later. Credit is due for the substantial analytic structure, the sharp jet-sum theorem, and the simplex reduction, but the arithmetic endgame is a fatal flaw in the submitted argument.","tokens_in":34019,"tokens_out":11866,"duration_ms":170537,"concrete_test":"Verify the quoted form of [BGM+22, Theorem 4.2] against Λ={ (2a+b, b/2) : a,b∈Z } ⊂ R^2: det Λ=1, Λ∩(−1,1)^2={0}, yet Λ is not any permutation of an upper triangular lattice with all diagonal entries 1. Then check whether Theorem 7.2's induction can be run without that bridge; if no alternative argument supplies a unitriangular basis, the arithmetic half is unsupported.","verdict_should_be":"REJECT","load_bearing_attack":"The most load-bearing defect is in Theorem 7.2. Lemma 7.1 only proves that the lattice Λ avoids the open unit cube (−1,1)^n. The proof then invokes [BGM+22, Theorem 4.2] in the form: any covolume-one lattice avoiding (−1,1)^n is, after a permutation, UZ^n with U upper triangular and all diagonal entries one. This statement is false. In R^2 take Λ = Z(2,0) ⊕ Z(1,1/2), i.e. points (2a+b, b/2) with a,b∈Z; its generator matrix has determinant 2·(1/2)−1·0=1, and one checks Λ∩(−1,1)^2={0}: a point with both coordinates strictly inside would require b∈{−1,1} and 2a+b to be −1,0, or 1, forcing the first coordinate to be ±1 or 0 with the second coordinate ±1/2, never both strictly inside. Yet Λ is not upper unitriangular up to permutation: its second coordinate must be a half-integer, so the lattice cannot equal UZ^2 with an upper triangular U whose diagonal entries are all 1. Since Theorem 7.2 is the only step converting |det A|=1 in Proposition 6.4 into A∈GL_n(Z), the main theorem is not proved. A repair would require a correct cube-avoidance characterization or a direct argument from the stronger hypothesis 'avoids int(S_n)' to a unitriangular basis, but none appears in the text.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper claims to prove the equality case of Ehrhart's volume conjecture: if K ⊂ R^n is a full-dimensional compact convex body with barycenter 0, int(K)∩Z^n = {0}, and vol(K) = (n+1)^n/n!, then K is a unimodular image of the centered standard simplex (n+1)Δ_n − (1,…,1). The proof has an analytic half, which develops a jet-sum theorem, lattice-width bounds, toric limit rays, and a pyramid-to-simplex argument, and an arithmetic half, which is a critical-lattice theorem (Theorem 7.2) asserting that Z^n is the only determinant-one lattice avoiding the interior of S_n. The manuscript explicitly relies on two lemmas from the unpublished report [OAI26] for the analytic framework and on a theorem attributed to [BGM+22, Theorem 4.2] for the arithmetic endgame.","tokens_in":34435,"tokens_out":27050,"duration_ms":343902,"significance":"If the proof were correct, the result would resolve a long-standing equality classification and would also provide a new critical-lattice uniqueness theorem for the centered simplex. The analytic strategy is original and well structured: Theorem 3.8, the hypersurface exclusion of Theorem 4.2, the width saturation of Corollary 4.5, and the pyramid theorem (Theorem 5.7) are substantial intermediate results that could be of independent interest. The paper also deserves credit for making its logical dependencies explicit, including the admission that it uses [OAI26] rather than reproving it. However, the central claim is not established as written because a load-bearing step in the arithmetic half relies on a false statement, and the analytic half depends on non-peer-reviewed external lemmas that are not reproduced.","major_comments":[{"comment":"The proof of Theorem 7.2 collapses at the invocation of [BGM+22, Theorem 4.2]. The statement as quoted — any covolume-one lattice avoiding the open unit cube (−1,1)^n is, after a permutation, UZ^n with upper triangular U and all diagonal entries one — is false. In R^2, the lattice Λ = Z(2,0) ⊕ Z(1,1/2) has determinant 1 and contains no nonzero point of (−1,1)^2, but it is not upper unitriangular up to permutation: its points have second coordinate in (1/2)Z, whereas every lattice of the form UZ^2 with U upper unitriangular has one coordinate equal to an integer. Lemma 7.1 supplies only cube avoidance, so the false theorem is the only bridge from the hypothesis int(S_n)∩Λ={0} to a unitriangular basis. Since Theorem 7.2 is the sole step converting |det A|=1 in Proposition 6.4 into A∈GL_n(Z), Theorem 1.3 is not established by the manuscript. The author needs either a correct characterization of determinant-one lattices avoiding int(S_n), proved directly, or a substantially different arithmetic argument.","section":"Section 7, Theorem 7.2"},{"comment":"The analytic half is load-bearing on two results imported verbatim from [OAI26, Chapter 8, Lemmas 2.1 and 2.2], which are not proved in the manuscript and are not available in a peer-reviewed source. Lemma 2.4 supplies the monomial orthogonality basis indexed by int(kK), and Lemma 2.5 supplies the local uniform convergence of (1/k) log B_k and total-variation convergence of the associated measures. These lemmas underlie Theorem 3.8 and therefore all subsequent sharp statements, including the hypersurface exclusion and the toric limit rays. The paper states in Section 2.2 that it uses these results directly rather than reproving them, but for a journal proof of the main theorem this is a serious verifiability gap: either the lemmas must be reproduced with complete hypotheses and proofs, or the paper must be revised to depend only on published, independently checkable sources.","section":"Section 2.2, Lemmas 2.4 and 2.5"},{"comment":"The sketch in Section 1 describes the arithmetic half as: sign-flip argument, then Hajós's theorem, then an explicit shear-point construction. This is not merely a presentation issue: the sign-flip argument (Lemma 7.1) gives only avoidance of the open unit cube, while Hajós's theorem is a theorem about cube tilings and does not imply that every cube-avoiding determinant-one lattice has a unitriangular basis, as the counterexample in the first major comment shows. The stronger hypothesis int(S_n)∩Λ={0} must therefore be used directly in the arithmetic argument; it is not used anywhere after Lemma 7.1 in the current text.","section":"Section 1, sketch of the proof and Section 7"}],"minor_comments":[{"comment":"The title contains a typo: “Ehrhar T’s” should read “Ehrhart's”; the same typo appears in the abstract heading.","section":"Title and abstract"},{"comment":"The measure dν is normalized with angular Haar measure of total mass one in Section 2.1, and Lemma 3.4 writes dν = π^{−n} ∏ |z_i|^{-2} dλ(z). This is consistent, but the normalization should be restated at the point of use to avoid confusion.","section":"Section 2.2, Eq. (2.1) and Lemma 3.4"},{"comment":"The paper repeatedly cites [OAI26] and the AI-system references [Liu+26, Ju+26]; these are not standard peer-reviewed publications. Even if they are publicly available, the manuscript should state precisely which claims are proved in the paper and which are imported, and it should flag the imported claims as unverified external dependencies.","section":"General"},{"comment":"Remark 7.3 says that the value 1 of the critical determinant is classical and cites [Ehr79]; the uniqueness assertion is claimed to be new. Given the problems with the proof, the remark should be phrased as conditional on a correct theorem.","section":"Section 7, Remark 7.3"}],"recommendation":"major_revision","confidential_remarks":"The paper's central claim may well be true, and the analytic part contains substantial and plausible ideas, but the arithmetic half as written rests on a demonstrably false citation, and the analytic half is conditional on an unpublished report. I would not recommend acceptance until the arithmetic step is replaced by a correct proof and the external analytic lemmas are either proved or replaced by published references. The author's disclosure of AI assistance is transparent and is not itself a reason for rejection; the mathematical gaps are."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a serious, long proof attempt at the equality case of Ehrhart's volume conjecture. The analytic half rests on two lemmas imported verbatim from the OpenAI report [OAI26]; if those are correct, the machinery works. The arithmetic half is short and plausible. The stress-test note's alleged counterexample to Theorem 7.2 does not hold up: the lattice Z(2,0)⊕Z(1,1/2) avoids the open unit cube, but after swapping coordinates it is exactly [[1,1/2],[0,1]]Z^2, which is upper unitriangular with diagonal ones. So a permutation of the axes does make it upper unitriangular. The arithmetic claim survives that attack.\n\nWhat is actually new: the equality classification for general centered bodies was open. The paper introduces a radial degeneration of point-jet filtrations to monomial filtrations, uses the equality case of Prékopa–Leindler to force a pyramid structure, and proves a critical-lattice uniqueness theorem for the centered simplex. The proof is structured and internally coherent: jet-sum theorem, width saturation, dense apices forcing a simplex, then the lattice endgame. The author is also honest about external inputs and about the AI provenance; that is not a flaw.\n\nSoft spots: the two [OAI26] lemmas (monomial basis and Bergman-kernel convergence) are load-bearing and not reproduced here. They are standard in spirit, but a referee will want either proofs or a reliable version. The application of Berndtsson's convexity to a nonsmooth envelope is not fully regularized and needs a check. The claim of an internal proof-checking pipeline is not backed by a machine-checked artifact. None of these is a known error; they are a verification burden.\n\nThis paper is for convex geometers and geometry-of-numbers specialists. It deserves a serious referee: send it to peer review with a request to prove or replace the [OAI26] lemmas and to expand the Berndtsson step. My own verdict is conditional, matching the reader's, but the stress-test note is not the reason to be cautious.","headline":"A serious, conditional proof of the equality case of Ehrhart's volume conjecture; the stress-test counterexample to the arithmetic half does not hold up.","tokens_in":34870,"tokens_out":21810,"would_cite":true,"duration_ms":236376,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["11H06","52A40","52B20","32A25"],"pacs":[],"model":"deepseek-v4-flash","headline":"The centered simplex is the only maximizer of Ehrhart's volume bound","keywords":["Ehrhart volume conjecture","equality case","centered simplex","convex bodies","lattice points","critical lattices","unimodular equivalence","Bergman kernels"],"falsifier":"Find a full-dimensional compact convex body in $\\mathbb{R}^3$ with barycenter at the origin, origin as the unique interior lattice point, and volume $32/3$ that is not unimodularly equivalent to the centered simplex $4\\Delta_3-(1,1,1)$; any such body refutes Theorem 1.3. A less global check: for any candidate equality body, compute the normalized jet sums $A_k$ of Theorem 3.8 at one base point; if they have a subsequential limit other than $n$, the proof's first step is false.","tokens_in":33760,"feed_emoji":"📐","tokens_out":9874,"duration_ms":115051,"temperature":0.7,"pith_summary":"This paper proves the equality case of Ehrhart's volume conjecture: a full-dimensional compact convex body in $\\mathbb{R}^n$ whose barycenter is the origin, whose only interior lattice point is the origin, and whose volume equals $(n+1)^n/n!$, must be a unimodular image of the centered standard simplex $(n+1)\\Delta_n-(1,\\ldots,1)$. Together with the inequality proved in [OAI26], this completes the 1964 conjecture. If the proof is correct, the only shapes that saturate the sharp volume bound are lattice-preserving affine copies of one explicit simplex, in every dimension. The interest is that the sharp bound now carries a rigid equality structure, not a family of near-extremizers.","feed_headline":"The centered simplex is the only maximizer of Ehrhart's volume bound","feed_subtitle":"Equality case of the 1964 conjecture: every centered body with one interior lattice point and maximal volume is a unimodular simplex.","key_machinery":"The load-bearing device is the sharp jet-sum theorem (Theorem 3.8): for every base point $p$ of the complex torus, the normalized sum over the vanishing-order filtration, $$A_k=\\frac{1}{k d_k}\\sum_{j=1}^{(n+1)k}\\dim F^j_{k,p}$$ with $d_k=\\#(\\mathbb{Z}^n\\cap\\operatorname{int}(kK))$, is forced by the volume hypothesis to converge to $n$. This single saturation statement controls all point filtrations simultaneously and yields the hypersurface-exclusion and lattice-width bounds. The second mechanical pillar is the critical-lattice theorem (Theorem 7.2): $\\mathbb{Z}^n$ is the unique determinant-one lattice whose nonzero points all avoid $\\operatorname{int}(S_n)$, proved through the cube-tiling theorem of [Haj41] after a sign-flip reduction.","core_discovery":"On the paper's own terms, the central discovery is a two-part rigidity theorem. The analytic half shows that any body $K$ satisfying the three hypotheses is a simplex: the volume hypothesis forces the normalized vanishing-order sums $A_k$ of the weighted $L^2$ spaces $H_k$ to tend to $n$ at every point; degenerating the base point radially along a generic direction converts the point-jet filtrations into monomial filtrations with integer weights; the equality case of the Prékopa--Leindler inequality forces the resulting toric rays to be translation rays; a translation ray makes $K$ a pyramid whose cap volumes are exactly $s^n/n!$; and since the good directions are dense, the pyramid apices force $K$ to be a simplex. The arithmetic half proves that $\\mathbb{Z}^n$ is the only determinant-one lattice avoiding the interior of the centered simplex $S_n=(n+1)\\Delta_n-(1,\\ldots,1)$, by a sign-flip reduction to the cube-tiling theorem of [Haj41] and an explicit shear-point argument. Combining the two halves gives $K=A S_n$ with $A\\in GL_n(\\mathbb{Z})$.","pith_inferences":["Beyond the paper: the same 'volume forces saturation at every point' mechanism might classify equality bodies for weighted versions of the conjecture, where the barycenter condition is replaced by a specified moment condition; the paper does not address those.","Beyond the paper: the critical-lattice uniqueness for $S_n$ suggests a broader principle: reflexive simplices whose critical lattice is unique are exactly those whose interior lattice point is the barycenter; testing this on other reflexive polytopes is a concrete next step.","Beyond the paper: the proof's dictionary with projective-space characterizations hints that any equality result on the convex side will be proved by a rigidity statement about lattice widths plus a lattice-uniqueness statement, rather than by copying the algebraic-geometry argument."],"forward_implications":["Every equality body in Ehrhart's volume conjecture is a unimodular image of $(n+1)\\Delta_n-(1,\\dots,1)$; there are no exotic maximizers in any dimension.","Any equality body has every lattice width at least $n+1$, and the centered standard simplex attains this bound, so width saturation is a necessary feature of maximal volume.","The discrete version says that for large $k$, no nonzero real polynomial of total degree at most $(n+1-\\varepsilon)k$ can vanish at all lattice points of $\\operatorname{int}(kK)$.","The determinant-one lattice avoiding the interior of the centered simplex is unique, namely $\\mathbb{Z}^n$.","Combined with the inequality from [OAI26], the 1964 volume conjecture is settled in full: bound and equality case."],"supporting_citations":[{"why":"states the original volume conjecture whose equality case is the paper's target.","marker":"[Ehr64]"},{"why":"supplies the modern formulation of the equality case conjecture and the prior equality classification for bodies contained in the polar of a lattice polytope.","marker":"[NP14]"},{"why":"supplies the inequality half of the conjecture and, more load-bearing, the two analytic lemmas (monomial bases and kernel convergence) that the present proof imports directly.","marker":"[OAI26]"},{"why":"provides the real Monge-Ampère transport potential $\\varphi$ whose pushforward is the uniform measure on $K$ and which plays the role of the Kähler-Einstein metric.","marker":"[BB13]"},{"why":"gives the convexity of the logarithm of the slice Bergman kernel used to make the envelope partition functions convex.","marker":"[Ber06]"},{"why":"provides the filtration and vanishing-order strategy, and the Fano-side equality template that the analytic half adapts.","marker":"[Fuj18]"},{"why":"supplies the equality case of the Prékopa-Leindler inequality that forces toric limit rays to be translation rays.","marker":"[Dub77]"},{"why":"resolves Minkowski's cube-tiling conjecture and is used to put the critical lattice in upper unitriangular form.","marker":"[Haj41]"},{"why":"gives the jet-interpolation identity connecting codimensions of vanishing-order filtrations to polynomial restriction ranks.","marker":"[Per00]"}],"fun_headline_variants":["Ehrhart equality: only centered simplices hit the volume bound","Maximal-volume centered bodies are all unimodular simplices","Proof: equality in Ehrhart's conjecture forces simplex shape","Centered lattice-point bodies with peak volume must be simplices","Ehrhart's conjecture sharp: maximizers are unimodular simplices"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification stands or falls on two analytic lemmas imported from [OAI26] and explicitly not reproved: that the monomials indexed by interior lattice points of $kK$ form an orthogonal basis of the weighted space $H_k$, and that the normalized kernels converge with total-variation convergence of the associated measures; if either needs hypotheses beyond the paper's stated ones, the equality classification does not follow.","fun_headline_variants_meta":{"raw":{"variants":["Ehrhart equality: only centered simplices hit the volume bound","Maximal-volume centered bodies are all unimodular simplices","Proof: equality in Ehrhart's conjecture forces simplex shape","Centered lattice-point bodies with peak volume must be simplices","Ehrhart's conjecture sharp: maximizers are unimodular simplices"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000216,"raw_usage":{"total_tokens":1399,"prompt_tokens":882,"completion_tokens":517,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":498,"completion_tokens_details":{"reasoning_tokens":428}},"tokens_in":498,"tokens_out":517,"duration_ms":5728,"temperature":1.0,"reasoning_tokens":428,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T01:01:55.309334+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a full-dimensional compact convex body in $\\mathbb{R}^3$ with barycenter at the origin, origin as the unique interior lattice point, and volume $32/3$ that is not unimodularly equivalent to the centered simplex $4\\Delta_3-(1,1,1)$; any such body refutes Theorem 1.3. A less global check: for any candidate equality body, compute the normalized jet sums $A_k$ of Theorem 3.8 at one base point; if they have a subsequential limit other than $n$, the proof's first step is false.","supporting_citations":[],"review_version":2}