{"id":"304e2f0a-2bde-47f7-882f-ddaeb18e4e5d","arxiv_id":"2411.18605","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Families of sets in R^d whose homological shatter function is bounded by a slowly growing function have the fractional Helly property, proved using new graded Radon and Helly numbers.","lead":"Families of sets in high-dimensional space can satisfy a fractional Helly theorem even when their Radon number is unbounded, if the Betti numbers of their intersections grow slowly. The proof introduces graded Radon and Helly numbers, a new tool for understanding intersection patterns.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 3.1's key step is outsourced to [2, Theorem 1.2], whose hypotheses are nowhere stated; the proof gives no check that a family with φ^⌈d/2⌉ ≤ Ψ_{d,b} satisfies them.","rationale":"The reader's weakest_assumption correctly identifies the central gap: the proof of Theorem 3.1 delegates the only conversion from many intersecting (d+1)-tuples to many intersecting m0-tuples to [2, Theorem 1.2], but the paper neither states that theorem nor verifies its hypotheses. If [2, Theorem 1.2] requires hypotheses beyond φ^{ceil(d/2)}(m0) ≤ b0, the proof does not establish the fractional Helly theorem. I additionally note the internal irregularity of Ψ_{d,b}: its two branches overlap at t = r(b,d) and the function can decrease, which makes the statement of Theorem 3.1 ambiguous and potentially vacuous for nondecreasing φ. Both issues are concrete, but the external bridge is load-bearing because it is the step that connects the shatter-function bound to the graded colorful-Helly framework. Since my main concern matches the reader's and the existing conditional verdict already records the need for verification, I do not move the verdict.","tokens_in":7479,"tokens_out":13701,"duration_ms":134575,"concrete_test":"Obtain [2, Theorem 1.2] from arXiv:2103.09286 and verify hypothesis-by-hypothesis against the application in the proof of Theorem 3.1: with X = F', d fixed, m = m0, and the bound φ^{ceil(d/2)}_F(m0) ≤ b0, check whether all assumptions (which homological shatter functions are bounded, any forbidden-homological-minor condition, any size threshold on |F'|) hold solely from φ^{ceil(d/2)}_F ≤ Ψ_{d,b}. If they do not, either add the missing hypothesis to Theorem 3.1 or supply an independent proof of the claimed (d+1)-to-m0 bridge; this settles whether the proof's critical step lands.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 3.1, immediately after fixing n0, uses [2, Theorem 1.2] as the only bridge from α|F'| intersecting (d+1)-tuples to α'|F'| intersecting m0-tuples, using solely φ^{ceil(d/2)}_F(m0) ≤ b0. The paper does not state [2, Theorem 1.2] or its hypotheses, so it is impossible to verify that the hypotheses are met by every family satisfying φ^{ceil(d/2)}_F ≤ Ψ_{d,b}; for example, the theorem might require control of all Betti numbers up to dimension d, or a forbidden-homological-minor assumption, or a different relation between the shatter bound and m0. If any hypothesis is missing, the central claim is unproved as written. Separately, the statement of Theorem 3.1 is not well-formed: Ψ_{d,b} is defined by overlapping intervals at t = r(b,d) (taking values b-1 and b), and Ψ can decrease from b to S(t) immediately after m(r(b,d))r(b,d) despite φ being nondecreasing; the theorem's hypothesis therefore needs repair before its applicability can be assessed. The external bridge, however, is the load-bearing step.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper introduces graded analogues of the Radon, Helly, and colorful Helly numbers of a set system, and relates them to the homological shatter function. The main result (Theorem 3.1) claims that if the homological shatter function of order ceil(d/2) is bounded above by an explicitly defined slowly growing function Psi_{d,b}, then the family satisfies a fractional Helly theorem: a positive fraction of intersecting (d+1)-tuples forces a positive fraction of members with a common point. The proof combines inequalities of Patakova and of Holmsen-Lee with an external theorem [2, Theorem 1.2] that is used as a black box. The paper concludes with constructions of set systems with prescribed homological shatter functions and with logarithmically growing graded Radon numbers.","tokens_in":7674,"tokens_out":8892,"duration_ms":78140,"significance":"If the main theorem is correct, it extends fractional Helly theorems to families whose Radon number may be unbounded, giving a concrete step toward the Kalai-Meshulam conjecture. The graded-parameter framework is natural, Proposition 2.2 provides a clean general bound on graded Radon numbers, and Lemma 4.1 gives an explicit realization of arbitrary nondecreasing homological shatter functions. These are genuine contributions. However, the proof of the central theorem currently depends on an unstated external result and on an ill-defined bounding function, so the significance is conditional on repairing the gaps described below.","major_comments":[{"comment":"The only bridge from a positive fraction of intersecting (d+1)-tuples to a positive fraction of intersecting m0-tuples is the invocation of [2, Theorem 1.2], but that theorem is neither stated nor paraphrased. Consequently, the hypotheses that the family F must satisfy are never checked; for example, [2, Theorem 1.2] might require bounds on all Betti numbers up to dimension d, a forbidden homological minor, or a different relation between m0 and the shatter function. Since this transfer is the load-bearing step, the proof is incomplete as written. Please state [2, Theorem 1.2] and verify its hypotheses from the assumption phi^{ceil(d/2)}_F <= Psi_{d,b}.","section":"Section 3, proof of Theorem 3.1 (after fixing n0)"},{"comment":"The piecewise definition of Psi_{d,b} has overlapping intervals at t = r(b,d), where the first branch assigns the value b-1 and the second assigns b, so Psi_{d,b} is not a well-defined function at that point. Moreover, unless one proves that m(r(b,d))r(b,d) >= r(b+1,d), the third branch can assign a value S(t) < b immediately after m(r(b,d))r(b,d); since phi is nondecreasing, no nondecreasing function can be bounded above by Psi_{d,b} in that regime, which would make the hypothesis of Theorem 3.1 vacuous for all families. Please repair the definition with disjoint intervals and either prove that Psi_{d,b} is nondecreasing or add this as an explicit condition.","section":"Section 3, definition of Psi_{d,b}"},{"comment":"The conclusion of Lemma 2.3, as stated in the manuscript, is that every m0 sets of G intersect, not that every r0 sets of G intersect. The deduction that all elements of G intersect requires first passing from m0-wise intersection to r0-wise intersection, which needs m0 >= r0; no such inequality is established. Also, the argument that h(F) <= r0 turns r0-wise intersection into total intersection is only valid if |G| >= r0, so the choice of n0 should explicitly ensure that beta(alpha',r0,m0) n0 >= r0. Please either prove these inequalities or reformulate the application so that it matches the stated hypotheses and conclusion of Lemma 2.3.","section":"Section 3, proof of Theorem 3.1 (application of Lemma 2.3)"},{"comment":"The proof uses the notation b0 and sets r0 = r(b0,d), while the theorem is stated with a fixed parameter b; this mismatch makes it impossible to see which value of b is being used in the final beta(b,d,alpha). Please unify the notation throughout the statement and proof.","section":"Section 3, Theorem 3.1 statement and proof"}],"minor_comments":[{"comment":"There are several typos, including 'growi ng', 'bounde d', 'speci c', and 'untrue'; a careful proofreading pass is needed.","section":"Abstract and Introduction"},{"comment":"The proof refers to 'Theorem 2.3' when the cited statement is 'Lemma 2.3'; please correct the cross-reference.","section":"Section 3, proof of Theorem 3.1"},{"comment":"In the final display, beta(alpha',r0,m) should presumably be beta(alpha',r0,m0); please clarify the argument.","section":"Section 3, proof of Theorem 3.1"},{"comment":"The text attributes the conjecture to 'Kalai and Meshulam in [7]', but reference [7] is listed as a single-author paper by Kalai; either correct the attribution or add the appropriate joint reference.","section":"Introduction and References"},{"comment":"The phrase 'for i < d-h, we can adapt the construction' is vague; please specify the adaptation or remove the clause.","section":"Section 4, Lemma 4.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is very short and leans heavily on an external arXiv reference for the main transfer step. The editor may wish to confirm the status of [2] and whether its Theorem 1.2 indeed has the power attributed to it. The main theorem's hypothesis also needs to be made well-defined before the result can be evaluated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper does something genuinely useful: it introduces graded versions of the Radon, Helly, and colorful Helly numbers for set systems, and shows how growth of the homological shatter function controls these graded parameters. That framework is reusable and is the right kind of idea. The paper also includes a nice characterization result (Lemma 4.1: every nondecreasing function is the homological shatter function of some family) and an example with logarithmically growing graded Radon numbers that shows why the non-stationary case is worth separating from the bounded-homological-complexity setting. The writing is clear and the citations look right.\n\nThe main theorem (Theorem 3.1) is plausible, but I can't call it proven as written. Two problems, one small and one load-bearing.\n\nSmall: the function Ψ_{d,b} is not well-defined at t = r(b,d), where the first and second cases both apply and give b−1 and b. Worse, it can drop from b to S(t) immediately after m(r(b,d))·r(b,d), even though the homological shatter function it is meant to bound is nondecreasing in t. This is fixable by redefining the intervals (e.g., half-open) and by making Ψ nondecreasing, but the statement needs that repair before the hypothesis makes sense.\n\nLoad-bearing: the proof's only bridge from a positive fraction of intersecting (d+1)-tuples to a positive fraction of intersecting m0-tuples is [2, Theorem 1.2]. The paper never states that theorem or its hypotheses. The proof just says that φ^⌈d/2⌉_F(m0) ≤ b0 lets you apply it. But the theorem might require control of all Betti numbers up to dimension d, or a forbidden-homological-minor condition, or something else entirely. Without seeing the theorem, I can't verify that every family with φ^⌈d/2⌉ bounded by Ψ_{d,b} satisfies its hypotheses. This is the kind of gap that can be filled by quoting the theorem and checking the condition, but until then the central claim is unproved.\n\nThe examples in Section 4 are honest: the authors admit their construction does not need the full strength of the theorem and that only the ⌈d/2⌉-th shatter function is controlled. That is fine.\n\nWho is this for? People working in topological combinatorics and fractional Helly theory. They will want the graded-parameter definitions regardless of the fate of Theorem 3.1. The paper deserves a serious referee, but the referee should require the author to state [2, Theorem 1.2] and verify its hypotheses, and to clean up Ψ. It is close to being a good paper, but not there yet.","headline":"A short, honest note: the graded parameter framework is a real contribution, but the main theorem as stated is not yet proven because the proof outsources its central step to an unstated external theorem and the bounding function has a fixable but real defect.","tokens_in":8253,"tokens_out":1624,"would_cite":true,"duration_ms":15671,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["52A35","52A01"],"pacs":[],"model":"deepseek-v4-flash","headline":"For families of sets in R^d, a slowly growing homological shatter function is enough for the fractional Helly theorem, even when the Radon number is unbounded.","keywords":["topological combinatorics","fractional Helly theorem","Radon number","homological shatter function","graded Helly number","homological VC dimension","Betti numbers"],"falsifier":"Exhibit a family $\\mathcal{F}$ of sets in $\\mathbb{R}^d$ with $\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}$ bounded by $\\Psi_{d,b}$ in which a positive fraction of $(d+1)$-tuples intersect but no positive fraction of $m_0$-tuples intersect, where $m_0 = m(r(b,d))$; such a family would refute the cited threshold theorem on which the proof depends. A concrete construction would be the minimal example showing the bridge from $(d+1)$-wise to $m_0$-wise intersections fails under only the low-degree Betti bound.","tokens_in":7214,"feed_emoji":"🧩","tokens_out":15403,"duration_ms":117761,"temperature":0.7,"pith_summary":"This paper proves a fractional Helly theorem for families of sets in $\\mathbb{R}^d$ whose intersection homology is controlled by a slowly growing 'homological shatter function', a hypothesis that does not force the classical Radon number to be finite. The main result, Theorem 3.1, says that for every $b,d \\ge 0$ and $\\alpha \\in (0,1)$ there are $\\beta > 0$ and $n_0$ such that if $\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}$ is bounded above by the paper's constructed function $\\Psi_{d,b}$, then any finite subfamily with a positive fraction of intersecting $(d+1)$-tuples contains a positive fraction of members with a common point. The novelty is that this conclusion is reached through new graded analogues of the Radon, Helly, and colorful Helly numbers, whose growth is shown to be controlled by the homological shatter function. The proof chains two conversion steps: a cited threshold theorem converts many intersecting $(d+1)$-tuples into many intersecting $m_0$-tuples, and a clique-number lemma converts those into a large intersecting subfamily. If the theorem is correct, it is a concrete step toward the conjecture that polynomial growth of the homological shatter function implies a fractional Helly theorem.","feed_headline":"Fractional Helly theorem holds even without bounded Radon number","feed_subtitle":"Graded Radon and Helly numbers rescue the classical theorem when the usual Radon number is infinite.","key_machinery":"The load-bearing objects are the graded parameters $r^{(t)}(\\mathcal{F})$, $h^{(t)}(\\mathcal{F})$, $ch^{(t)}(\\mathcal{F})$ — the suprema of the ordinary Radon, Helly, and colorful Helly numbers over all subfamilies of size at most $t$ — together with the homological shatter function $\\phi_{\\mathcal{F}}^{(h)}(t)$, which records the largest reduced Betti number in dimensions $0,\\dots,h$ among all intersections of at most $t$ sets of $\\mathcal{F}$. The proof's mechanism is the chain of inequalities (3) and (4), which bound these graded parameters by functions of $\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}(t)$, and the specifically constructed function $\\Psi_{d,b}$, built by inverting the Radon bound $R_d(b') = r(b'+1,d)$, which is chosen so that the bounds satisfy the quantitative hypotheses of the conversion lemmas. In effect, $\\Psi_{d,b}$ is a threshold: any family whose low-degree intersection homology grows no faster than this function is guaranteed to have bounded Helly number and bounded graded colorful Helly number at the needed scale.","core_discovery":"The central claim is Theorem 3.1. For every $b,d \\ge 0$ and every $\\alpha \\in (0,1)$, there exist $\\beta > 0$ and $n_0$ such that every family $\\mathcal{F}$ of sets in $\\mathbb{R}^d$ with $\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}$ bounded from above by $\\Psi_{d,b}$ has the fractional Helly property: every finite subfamily $\\mathcal{F}'$ with $|\\mathcal{F}'| \\ge n_0$ and at least $\\alpha \\binom{|\\mathcal{F}'|}{d+1}$ intersecting $(d+1)$-tuples contains at least $\\beta |\\mathcal{F}'|$ members with a common point. The proof introduces the graded parameters $r^{(t)}$, $h^{(t)}$, $ch^{(t)}$ — the suprema of the ordinary Radon, Helly, and colorful Helly numbers over subfamilies of size at most $t$ — and shows, via the Radon bound $r(\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}(t), d)$, that these graded numbers cannot grow too fast. The function $\\Psi_{d,b}$ is chosen precisely so that this growth satisfies inequalities (5) and (6), which put the family inside the hypotheses of the two conversion lemmas. The conclusion that all members of the final subfamily intersect uses the bounded Helly number $h(\\mathcal{F}) \\le r_0$. Because $\\Psi_{d,b}$ is unbounded, the family may have infinite Radon number.","pith_inferences":["The proof depends on the unstated threshold theorem; re-proving that theorem under the low-degree Betti bound $\\phi^{(\\lceil d/2 \\rceil)}$ alone would make Theorem 3.1 self-contained and likely extend to other ambient spaces.","The graded-parameter framework is not specific to $\\mathbb{R}^d$: any space with a Radon-type bound $r(\\cdot,\\cdot)$ would inherit the same fractional Helly theorem from the same argument.","Section 4's examples have very simple nerves (disjoint unions of cliques), so they do not stress the new theorem; a natural test case is a family where $\\phi^{(h)}$ is bounded for some $h < \\lceil d/2 \\rceil$ but not at $\\lceil d/2 \\rceil$, to see whether the dimension threshold in the theorem is sharp.","The gap between the very slow $\\Psi_{d,b}$ and the conjectured polynomial threshold suggests that a genuinely different bridge, not a refinement of the present conversion lemmas, will be needed to settle the conjecture."],"forward_implications":["Any family of sets in $\\mathbb{R}^d$ with $\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}$ bounded by $\\Psi_{d,b}$ satisfies the fractional Helly theorem, even when its Radon number is infinite.","The constants in the theorem depend only on $b$, $d$, and $\\alpha$, so the conclusion is uniform across all families with this shatter-function bound.","The graded-parameter method extends the fractional Helly theorem from families with finite homological complexity to families with slowly growing homological complexity.","The construction of $\\Psi_{d,b}$ shows the current proof reaches only very slow growth — roughly the iterated logarithm — so the polynomial-growth conjecture remains strictly out of reach of this technique."],"supporting_citations":[{"why":"Supplies the unstated theorem that a positive fraction of intersecting $(d+1)$-tuples implies a positive fraction of intersecting $m_0$-tuples under $\\phi^{(\\lceil d/2 \\rceil)}(m_0) \\le b_0$; the proof of Theorem 3.1 invokes it directly.","marker":"[2]"},{"why":"Supplies the Radon-type bound $r(\\cdot,\\cdot)$ that controls the graded Radon number in terms of the homological shatter function, used in inequalities (3) and (4).","marker":"[11]"},{"why":"Supplies the function $m(\\cdot)$ bounding the colorful Helly number by the Radon number, used to control the graded colorful Helly numbers.","marker":"[6]"},{"why":"Supplies the quantitative clique lemma (Lemma 2.3) that produces a large subfamily whose every $r_0$ members intersect from a positive fraction of intersecting $m_0$-tuples.","marker":"[5]"}],"fun_headline_variants":["Graded Radon numbers save fractional Helly theorem","Fractional Helly works even if Radon number explodes","New proof: fractional Helly without bounded Radon","Graded Radon/Helly numbers rescue fractional Helly theorem"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole argument rests on a cited but unstated theorem that many intersecting $(d+1)$-tuples imply many intersecting $m_0$-tuples whenever $\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}(m_0)$ is bounded; if that theorem actually needs control of all Betti numbers up to dimension $d$, the proof of Theorem 3.1 does not go through.","fun_headline_variants_meta":{"raw":{"variants":["Graded Radon numbers save fractional Helly theorem","Fractional Helly works even if Radon number explodes","New proof: fractional Helly without bounded Radon","Graded Radon/Helly numbers rescue fractional Helly theorem"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000157,"raw_usage":{"total_tokens":1225,"prompt_tokens":954,"completion_tokens":271,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":570,"completion_tokens_details":{"reasoning_tokens":203}},"tokens_in":570,"tokens_out":271,"duration_ms":2541,"temperature":1.0,"reasoning_tokens":203,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:02:22.499183+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Exhibit a family $\\mathcal{F}$ of sets in $\\mathbb{R}^d$ with $\\phi_{\\mathcal{F}}^{(\\lceil d/2 \\rceil)}$ bounded by $\\Psi_{d,b}$ in which a positive fraction of $(d+1)$-tuples intersect but no positive fraction of $m_0$-tuples intersect, where $m_0 = m(r(b,d))$; such a family would refute the cited threshold theorem on which the proof depends. A concrete construction would be the minimal example showing the bridge from $(d+1)$-wise to $m_0$-wise intersections fails under only the low-degree Betti bound.","supporting_citations":[{"cited_title":"Intersection patterns in spaces with a forbidden homological minor","cited_arxiv_id":"2103.09286","evidence_quote":"Supplies the unstated theorem that a positive fraction of intersecting $(d+1)$-tuples implies a positive fraction of intersecting $m_0$-tuples under $\\phi^{(\\lceil d/2 \\rceil)}(m_0) \\le b_0$; the proof of Theorem 3.1 invokes it directly."},{"cited_title":"Pat´ akov´ a","cited_arxiv_id":null,"evidence_quote":"Supplies the Radon-type bound $r(\\cdot,\\cdot)$ that controls the graded Radon number in terms of the homological shatter function, used in inequalities (3) and (4)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the function $m(\\cdot)$ bounding the colorful Helly number by the Radon number, used to control the graded colorful Helly numbers."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the quantitative clique lemma (Lemma 2.3) that produces a large subfamily whose every $r_0$ members intersect from a positive fraction of intersecting $m_0$-tuples."}],"review_version":1}