{"id":"963f3eb4-3a90-4145-810c-6a3d0dd4301c","arxiv_id":"2504.21345","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"The median hypersimplex Δ(2k,k) is Minkowski indecomposable, and the hemi-icosahedral Bier sphere admits a 12-vertex polytopal realization in R5.","lead":"The authors prove that the median hypersimplex, a standard family of 0/1 polytopes, cannot be decomposed as a Minkowski sum of two smaller polytopes; this makes it a ray in the deformation cone of the permutahedron. They also give an explicit 12-vertex polytope in 5 dimensions whose boundary is the Bier sphere of the minimal triangulation of the real projective plane.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The polytopality proof for Bier(I6) depends on unverified floating-point coordinates: the paper admits 5-decimal rounding destroys convexity, yet no exact-arithmetic or interval certificate is supplied.","rationale":"The reader's conditional verdict is appropriate, and I agree that the manuscript needs revision before the claims are fully rigorous. The reader identified two weaknesses: the undefined condition (5.1) behind Lemma 3.12 and the floating-point verification of Theorem 4.2. I regard the second as the single most load-bearing concern because Theorem 4.2 is an existence proof whose only evidence is a delicate numerical computation; the admitted fragility under rounding makes an exact certificate essential. The Lemma 3.12 issue is real but appears repairable: the proof sketch for the specific complex K = [2k]≤k−1 is largely self-contained, using Proposition 3.4 to establish the needed facet containment and adjacency, and the missing condition (5.1) is likely a missing cross-reference rather than a hidden assumption. Therefore I would keep the verdict at CONDITIONAL (equivalently, UNCHANGED relative to the reader), with the required revision being an exact-arithmetic certification of the coordinates in Theorem 4.2 and a proper statement or reference for condition (5.1) plus a fuller proof of Lemma 3.12.","tokens_in":11149,"tokens_out":16656,"duration_ms":163808,"concrete_test":"Interpret the twelve rows of matrix (4.3) as exact rational vectors in Q^5 (each entry is an integer multiple of 10^-7). In a computer algebra system with exact rational arithmetic (e.g., SageMath or Polymake with rational arithmetic), compute the convex hull of these twelve points. Verify that the resulting face lattice is isomorphic to Bier(I6), e.g., by matching the facet list (4.4) and edge list (4.5) under the relabeling sigma. For each of the 60 five-vertex facets, compute the exact affine hyperplane through its vertices and check that all other seven vertices lie strictly on the same open halfspace, or equivalently extract the 60 facet-defining inequalities with rational coefficients and confirm all vertices have strictly positive slack.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of Theorem 4.2 is that the twelve rows of matrix (4.3) are vertices of a 5-dimensional polytope whose boundary is the Bier sphere Bier(I6). The proof is a report of a Polymake computation in floating-point arithmetic: 'Applying Polymake to the data above... we obtain the face lattice.' No code, no exact rational arithmetic, and no interval-arithmetic certificate are provided. The paper itself notes that rounding the same coordinates to 5 decimal places destroys convexity and changes the face lattice. This means the realization sits very close to the boundary of the polytopality region, so a tiny numerical perturbation could change the combinatorial type. Since the coordinates are printed to 7 decimals, they are rational numbers with denominator 10^7; one can check them exactly. Without such an exact check, the theorem that a polytopal realization exists is not rigorously established—it rests on the reliability of floating-point output. I do not regard the undefined 'condition (5.1)' in Lemma 3.12 as the primary risk: the proof sketch for K = [2k]≤k−1 is largely self-contained, and the missing reference appears to be an expositional defect rather than a mathematical gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies deformation cones of hypersimplices and polytopality of Bier spheres. Its first main result, Theorem 3.10, claims that the median hypersimplex Δ_{2k,k} is Minkowski indecomposable, equivalently that its essential deformation cone is one-dimensional, and hence that Δ_{2k,k} spans a ray in the submodular cone of the permutahedron. The proof uses the wall-crossing relations of Proposition 3.1 applied to a Bier-sphere refinement of the normal fan of Δ_{2k,k}, relying on the identification of the polar of the diplo-simplex with Δ_{2k,k} from previous work. The second main result, Theorem 4.2, claims that the Bier sphere of the hemi-icosahedron I_6 is polytopal, realized as the boundary of a five-dimensional polytope whose twelve vertices are given by the rows of the numerical matrix (4.3). The proof consists of a Polymake computation of the face lattice of the convex hull of these rows and a comparison with the facet list of Bier(I_6).","tokens_in":11228,"tokens_out":8710,"duration_ms":93577,"significance":"If both theorems are rigorously established, the paper makes a meaningful contribution: it provides a new infinite family of rays in the submodular cone, a step toward the longstanding problem initiated by Edmonds, and it supplies a new non-threshold example of a polytopal Bier sphere, the 12-vertex hemi-icosahedral Bier sphere. The wall-crossing framework and the explicit computational data in matrix (4.3) are concrete and checkable, which is a strength of the paper's overall approach. The caveat is that the current manuscript leaves two load-bearing points under-proved, so the significance of the results is real but the rigor of the present write-up is not yet at the level needed for publication.","major_comments":[{"comment":"Lemma 3.12 is the load-bearing step that turns the Bier-fan refinement into the wall-crossing equalities x_S = y_T for every balanced partition S⊔T of [2k]. As stated, the lemma invokes 'condition (5.1)', but no equation (5.1) appears anywhere in Section 5 or elsewhere in the manuscript; the only numbered conditions in Section 5 are (5.2), (5.3), and (5.6). The proof is also only a short sketch. Since the conclusion that LinDef(Δ_{2k,k}) is cut out by (3.13) depends on every balanced S being covered, the manuscript must either state condition (5.1) explicitly, give a complete proof of Lemma 3.12, or provide a precise reference to the condition in [20] or [40].","section":"3.2, Lemma 3.12"},{"comment":"The proof that the boundary of the convex hull Q of the rows of matrix (4.3) is the Bier sphere Bier(I_6) is a report of a Polymake computation in floating-point arithmetic, and the paper itself states that rounding the same coordinates to 5 decimal places destroys convexity and changes the face lattice. No exact-arithmetic or interval-arithmetic certificate is supplied. Because every entry in (4.3) is a decimal fraction with denominator 10^7, an exact rational verification is feasible: one can scale the matrix by 10^7 and recompute the face lattice in rational arithmetic, or provide an independent exact certificate. Without such a check, Theorem 4.2 is not rigorously established.","section":"4.2, Theorem 4.2"},{"comment":"The final inference from Lemma 3.15 to one-dimensionality is too compressed. After imposing the normalizing conditions x_i = y_i, the equations in (3.13) become 2 x_S = x_[n] for every k-subset S of [2k]. To conclude that the essential deformation cone is one-dimensional, the proof must show that these equations force all x_i to be equal, leaving only x_[n] as a parameter. This follows by comparing the equations for two k-subsets that differ in one element, but that argument is not written down; the sentence 'x[n] = y[n] is the only variable parameter' is asserted rather than proved.","section":"3.2, Lemma 3.15 and proof of Theorem 3.10"}],"minor_comments":[{"comment":"The section title contains a typo: 'submoduar cone' should be 'submodular cone'.","section":"2.1"},{"comment":"In equation (3.11), the dummy index k in the sum ∑_{k∈T} y_k clashes with the fixed parameter k in Δ_{2k,k}; using another letter, such as j, would improve readability.","section":"3.2, Eq. (3.11)"},{"comment":"The abstract refers to a 'twelve vertex, 4-dimensional polytopal realization' while Theorem 4.2 speaks of a 'five dimensional convex polytope' whose boundary is the Bier sphere; these are consistent because the boundary is a 4-dimensional sphere, but the wording should be made unambiguous.","section":"Abstract and Theorem 4.2"},{"comment":"The facet list in table (4.4) uses barred labels ¯1,...,¯6, but the relabeling map σ is described only in words and the correspondence between the barred labels and the rows of matrix (4.3) is not spelled out; an explicit list would help reproducibility.","section":"4.2, Table (4.4)"}],"recommendation":"major_revision","confidential_remarks":"The two main theorems are potentially correct and interesting, but the manuscript currently has two load-bearing gaps: the undefined condition (5.1) in Lemma 3.12 and the unverified floating-point certificate in Theorem 4.2. I would not reject the paper, because both gaps appear fixable within the scope of the manuscript—the first by completing the proof of Lemma 3.12 and the second by an exact rational recomputation of the face lattice of the 12-point set defined by (4.3). I would ask the authors to address both points in revision and also to make the computation reproducible by supplying the exact arithmetic input and output."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know about this paper. First, the indecomposability theorem for the median hypersimplex is a real result and the proof is essentially sound. Second, the polytopality result for the 12-vertex Bier sphere is plausible but the proof depends on floating-point output; it needs a certificate before I'd call it rigorous.\n\nThe main theorem, that Δ_{2k,k} is Minkowski indecomposable and hence a ray in the submodular cone, is a genuine addition to a problem that has been open since Edmonds. The strategy is clean: identify the normal fan of the median hypersimplex with the radial fan of the diplo-simplex, refine by a Bier fan, and use wall-crossing equalities to force all deformation coordinates to be equal after translation. I checked the linear algebra; the essential cone really is one-dimensional. The only annoyance is Lemma 3.12, which refers to a 'condition (5.1)' that does not exist in the appendix, and uses notation like (S1,T;{i1}) without definition. For the concrete choice K = [2k]_{≤k−1} the proof sketch is convincing, so this reads as an expositional defect rather than a mathematical gap. They need to fix the undefined reference and spell out the facet/ridge argument.\n\nThe Bier sphere part is more concerning. Theorem 4.2 reports a polytopal realization of Bier(I6) via twelve 7-decimal coordinates, verified with Polymake in floating point. The paper honestly notes that rounding to 5 decimals destroys convexity, which means the point lies close to a face boundary. That is not disqualifying—the coordinates are rational, so exact arithmetic can settle it—but the manuscript gives no exact certificate, no code, and no interval-arithmetic check. The facet and edge lists are printed, so a referee can re-run the verification, but the burden is on the authors to supply the exact check. I would want that before accepting.\n\nThe citation pattern looks fine; the reliance on the same group's earlier paper [20] for the diplo-simplex identification is legitimate since that result is published. The paper deserves a serious referee. My recommendation: send it to review, and require a revision that fixes the missing condition in Lemma 3.12 and provides an exact or interval-certified verification of the coordinates in Theorem 4.2.","headline":"A solid new indecomposability theorem for the median hypersimplex, plus a plausible but under-certified computational polytopality claim that needs an exact verification before it is fully convincing.","tokens_in":11965,"tokens_out":5605,"would_cite":true,"duration_ms":57887,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52B12","52B70","05E45"],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper proves that the median hypersimplex cannot be split as a Minkowski sum, placing it as a ray in the permutahedron's deformation cone, and that the Bier sphere of the hemi-icosahedron is an actual polytope with explicit…","keywords":["median hypersimplex","Minkowski indecomposability","deformation cone","submodular cone","permutahedron","Bier sphere","hemi-icosahedron","polytopal sphere"],"falsifier":"For the first theorem, enumerate the wall-crossing equalities in the Bier fan of $K$, the complex of all subsets of size at most $k-1$: if any balanced $S \\sqcup T$ fails to produce $x_S = y_T$, the linear span in (3.13) grows and the essential deformation cone has dimension greater than one. For the second, recompute the convex hull of the rows of matrix (4.3) in exact rational arithmetic and compare its facets with table (4.4); any discrepancy shows the floating-point coordinates do not establish polytopality.","tokens_in":10812,"feed_emoji":"🧊","tokens_out":12458,"duration_ms":113955,"temperature":0.7,"pith_summary":"This paper establishes two results in the geometry of polytopes. First, the median hypersimplex — the polytope whose vertices are the 0/1 vectors with exactly k ones among 2k coordinates — cannot be expressed as a non-trivial Minkowski sum, meaning it cannot be built by adding together two genuinely different polytopes pointwise. Because the median hypersimplex is a deformed permutahedron, this indecomposability means its essential deformation cone is a single ray, adding a symmetric ray to the submodular cone of the permutahedron, a structure whose rays have been sought since the 1970s. Second, the Bier sphere built from the hemi-icosahedron, the smallest triangulation of the real projective plane, is polytopal: it is combinatorially equivalent to the boundary of a 5-dimensional polytope with 12 vertices, and the paper gives explicit coordinates for those vertices. A sympathetic reader should care because both results answer concrete 'can it be split?' and 'can it be realized?' questions with objects that are central in the study of polytopal spheres and deformation cones.","feed_headline":"Median hypersimplex cannot be split; 12-vertex sphere is polytopal","feed_subtitle":"A one-dimensional deformation cone supplies a ray; seven-decimal coordinates give a real 5-D polytope.","key_machinery":"The argument runs through three linked objects. The median hypersimplex $\\Delta_{2k,k}$ is the convex hull of all 0/1 vectors of length $2k$ with coordinate sum $k$. Its polar dual is, up to affine equivalence, the diplo-simplex $\\Omega_{2k}$, the convex hull of a simplex and its negative; Corollary 3.9 turns the normal fan of the hypersimplex into the radial fan of this diplo-simplex. The Bier sphere $\\mathrm{Bier}(K)$ of a simplicial complex $K$ is the deleted join of $K$ with its Alexander dual $K^\\circ$, the family of complements of non-faces of $K$; Bier-sphere fans provide simplicial refinements of the radial fan. The load-bearing mechanism is Lemma 3.12: for $K$ the complex of all subsets of size at most $k-1$, each balanced partition $S \\sqcup T$ of $[2k]$ forces a wall-crossing equality $x_S = y_T$ among deformation parameters, and these equalities leave only one free parameter in the essential deformation cone. For Theorem 4.2, the machinery is a previously developed algorithm that produces polytopal Bier-sphere realizations by successive local re-triangulations from a canonical threshold realization, combined with a computer search that supplies the 12-by-5 coordinate matrix (4.3).","core_discovery":"On the paper's own terms, the core discovery is Theorem 3.10 and Theorem 4.2. Theorem 3.10 states that the hypersimplex $\\Delta_{2k,k}$ is Minkowski indecomposable, equivalently its essential deformation cone $\\mathrm{Def}_{\\mathrm{ess}}(\\Delta_{2k,k})$ is one-dimensional; consequently $\\Delta_{2k,k}$ represents a symmetric ray in the deformation cone of the permutahedron, i.e. in the submodular cone. The proof identifies the normal fan of $\\Delta_{2k,k}$ with the radial fan of the polar diplo-simplex $\\Omega_{2k}$, refines that fan by a Bier-sphere fan, and uses the wall-crossing equalities of Proposition 3.1 to show that any deformation vector is determined up to scaling and translation. Theorem 4.2 states that the Bier sphere $\\mathrm{Bier}(I_6)$ of the hemi-icosahedron — the minimal 6-vertex triangulation of the real projective plane — is polytopal; the rows of matrix (4.3) are coordinates in $\\mathbb{R}^5$ whose convex hull has exactly the Bier sphere's face lattice.","pith_inferences":["Beyond the paper, the same wall-crossing method could test whether non-median hypersimplices, which appear as summands in the permutahedron decomposition, are decomposable; if so, the median hypersimplices would be exactly the ray-generating members of the family.","Beyond the paper, the near-degeneracy of the 12-vertex coordinates suggests seeking an exact rational realization, whose existence would strengthen the polytopality result to a certified construction.","Beyond the paper, applying the incremental algorithm to Bier spheres of other minimal triangulations, such as higher-dimensional projective-plane analogs, could locate the first genuinely non-polytopal Bier sphere.","Beyond the paper, the wall-crossing equalities derived from extremal-volume Bier spheres may point to additional previously unknown rays in the submodular cone, not just the median-hypersimplex family."],"forward_implications":["Because the essential deformation cone of $\\Delta_{2k,k}$ is one-dimensional, every admissible deformation of the median hypersimplex is a translate and a dilate of it.","The submodular cone of the permutahedron contains the symmetric rays spanned by the median hypersimplices $\\Delta_{2k,k}$ for every $k$.","The twelve-vertex Bier sphere of the hemi-icosahedron has a concrete polytopal realization in $\\mathbb{R}^5$, extending the known polytopality of all Bier spheres with up to eleven vertices.","The face lattice of the 12-vertex convex hull of matrix (4.3) coincides with the facets and edges of $\\mathrm{Bier}(I_6)$ listed in tables (4.4) and (4.5), giving an explicit checkable instance of a non-threshold Bier sphere that is polytopal."],"supporting_citations":[{"why":"Supplies Proposition 3.1, the wall-crossing equalities and inequalities that determine the deformation cone from a simplicial refinement of the normal fan.","marker":"[32]"},{"why":"Proves Theorem 3.7 and Corollary 5.4: the polar of the diplo-simplex is the median hypersimplex and maximal-volume Bier spheres share the same convex body, giving the normal-fan identification.","marker":"[20]"},{"why":"Defines Bier spheres as deleted joins of a complex and its Alexander dual, the combinatorial objects at the center of both theorems.","marker":"[27]"},{"why":"Introduces generalized permutahedra and the Minkowski decomposition of the permutahedron into hypersimplices, making $\\Delta_{2k,k}$ a deformed permutahedron and linking it to the submodular cone.","marker":"[34]"},{"why":"Introduced polymatroids as the deformation cone of the permutahedron and posed the problem of determining its rays, the context in which the new ray is identified.","marker":"[14]"},{"why":"Proves Theorem 5.5 that threshold Bier spheres are polytopal as convex hulls of two simplices, the canonical starting point for the incremental realization algorithm.","marker":"[21]"},{"why":"Provides the successive-modification algorithm and the floating-point search that produces the twelve-vertex coordinates realizing $\\mathrm{Bier}(I_6)$.","marker":"[40]"},{"why":"The computational polytope tool used to compute the convex hull's face lattice and verify that it matches the Bier sphere facets.","marker":"[17]"}],"fun_headline_variants":["Indecomposable median hypersimplex; polytopal Bier sphere","Median hypersimplex cannot split; 12-vertex Bier sphere is polytopal","Hypersimplex indecomposable, Bier sphere polytopal","Ray in deformation cone; 12-vertex sphere polytopal"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The first theorem stands on the unstated condition (5.1) in Lemma 3.12, which must guarantee that every balanced partition $S \\sqcup T$ of $[2k]$ yields the wall-crossing equality $x_S = y_T$; the second stands on the exactness of the seven-decimal coordinates in (4.3), since rounding them to five decimals destroys convexity.","fun_headline_variants_meta":{"raw":{"variants":["Indecomposable median hypersimplex; polytopal Bier sphere","Median hypersimplex cannot split; 12-vertex Bier sphere is polytopal","Hypersimplex indecomposable, Bier sphere polytopal","Ray in deformation cone; 12-vertex sphere polytopal"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000246,"raw_usage":{"total_tokens":1537,"prompt_tokens":938,"completion_tokens":599,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":554,"completion_tokens_details":{"reasoning_tokens":516}},"tokens_in":554,"tokens_out":599,"duration_ms":5575,"temperature":1.0,"reasoning_tokens":516,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:06:53.695612+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For the first theorem, enumerate the wall-crossing equalities in the Bier fan of $K$, the complex of all subsets of size at most $k-1$: if any balanced $S \\sqcup T$ fails to produce $x_S = y_T$, the linear span in (3.13) grows and the essential deformation cone has dimension greater than one. For the second, recompute the convex hull of the rows of matrix (4.3) in exact rational arithmetic and compare its facets with table (4.4); any discrepancy shows the floating-point coordinates do not establish polytopality.","supporting_citations":[{"cited_title":"and Poullot, G.,Deformation cones of hypergraph polytopes","cited_arxiv_id":null,"evidence_quote":"Supplies Proposition 3.1, the wall-crossing equalities and inequalities that determine the deformation cone from a simplicial refinement of the normal fan."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proves Theorem 3.7 and Corollary 5.4: the polar of the diplo-simplex is the median hypersimplex and maximal-volume Bier spheres share the same convex body, giving the normal-fan identification."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines Bier spheres as deleted joins of a complex and its Alexander dual, the combinatorial objects at the center of both theorems."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces generalized permutahedra and the Minkowski decomposition of the permutahedron into hypersimplices, making $\\Delta_{2k,k}$ a deformed permutahedron and linking it to the submodular cone."},{"cited_title":"Combinatorial Structures and their Applications","cited_arxiv_id":null,"evidence_quote":"Introduced polymatroids as the deformation cone of the permutahedron and posed the problem of determining its rays, the context in which the new ray is identified."},{"cited_title":"and Joswig, M.,polymake: a framework for analyzing convex polytopes","cited_arxiv_id":null,"evidence_quote":"The computational polytope tool used to compute the convex hull's face lattice and verify that it matches the Bier sphere facets."}],"review_version":1}