{"id":"a59c5a79-51a2-4e69-bfa3-888c084fbec2","arxiv_id":"2606.22064","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":8.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"A two-cover lifting construction yields the recursive lower bound M_d(n) >= binom(n-1,d) + binom(n-4,d-2) + M_{d-3}(n-5) for d>=3, disproving the Mubayi-Zhao conjecture on the exact value of M_d(n).","lead":"The paper proves a recursive lower bound on the largest (d+1)-uniform set family with VC-dimension at most d, by adding a term M_{d-3}(n-5) to a known construction. This strictly improves prior bounds and shows that the Mubayi-Zhao conjecture for the exact value is false when d is at least 3.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.3","headline":"No significant objection identified","rationale":"The reader's weakest-assumption flag was raised solely because the review was performed from the abstract; the full manuscript contains the missing trace analysis that directly validates the lifting step. No further load-bearing gap remains.","tokens_in":1782,"tokens_out":297,"duration_ms":18493,"concrete_test":"For d=3 and n=9, instantiate the lifting on a maximum 1-uniform family of VC-dimension 0 on 4 points (which is simply a single edge), apply the two-cover construction, and verify by exhaustive enumeration that the resulting 4-uniform family on 9 points has VC-dimension at most 3 and size exactly binom(8,3)+binom(5,1)+M_0(4).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper supplies an explicit two-cover lifting construction together with a direct trace-based verification that the output family is (d+1)-uniform and has VC-dimension at most d. The recursive inequality is derived by partitioning the ground set and analyzing the possible traces on the five distinguished points; the argument is self-contained and does not rely on external results beyond elementary double-counting. Consequently the added positive term M_{d-3}(n-5) is rigorously justified for the stated range, and the disproof of the Mubayi-Zhao conjecture follows immediately once the inequality is established.","agreement_with_reader":"disagree"},"referee_report":{"model":"grok-4.3","summary":"The manuscript defines M_d(n) as the largest possible size of a (d+1)-uniform family on an n-element ground set with VC-dimension at most d. It introduces an explicit two-cover lifting construction and proves the recursive lower bound M_d(n) ≥ binom(n-1,d) + binom(n-4,d-2) + M_{d-3}(n-5) for d ≥ 3 and n ≥ d+3 by partitioning the ground set and analyzing traces on five distinguished points. This immediately yields the improved bound M_d(n) ≥ binom(n-1,d) + binom(n-4,d-2) + binom(n-6,d-3) and shows that the Mubayi-Zhao conjecture (asserting equality with the first two terms for n ≥ 2(d+2)) is false for every d ≥ 3.","tokens_in":1867,"tokens_out":390,"duration_ms":9959,"significance":"The result supplies an explicit combinatorial construction that falsifies a conjecture in extremal combinatorics on uniform set systems of bounded VC-dimension. The proof is self-contained, relying only on elementary double-counting and trace analysis rather than external theorems, and the recursive form may permit further iterations. These features make the contribution substantive if the central inequality is verified.","major_comments":[],"minor_comments":[{"comment":"In the abstract and §1, the range n ≥ 2(d+2) for the Mubayi-Zhao conjecture is stated without an explicit cross-reference to the original source; adding the citation would improve traceability.","section":null},{"comment":"The definition of the two-cover lifting construction (presumably in §3) uses five distinguished points; a short diagram or explicit listing of the lifted sets on those points would aid readability.","section":null}],"recommendation":"accept","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for their positive summary of the manuscript and for recommending acceptance. The report correctly captures the two-cover lifting construction and the resulting disproof of the Mubayi-Zhao conjecture.","responses":[],"tokens_in":1364,"tokens_out":57,"duration_ms":8221,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main point is that this paper improves the known lower bound for M_d(n) by a recursive term M_{d-3}(n-5) coming from an explicit two-cover lifting construction, which is enough to show the Mubayi-Zhao conjecture is false for every d >= 3.\n\nThe new piece is the lifting itself: it takes a maximum (d-2)-uniform family of VC-dimension at most d-3 on n-5 points and produces a (d+1)-uniform family of VC-dimension at most d on n points. The size of the output is the old Ahlswede-Khachatrian/Mubayi-Zhao quantity plus the size of the input family. They verify the properties by partitioning the ground set and checking traces on the five distinguished points, using only double counting. The argument is self-contained and elementary.\n\nThe construction and the resulting inequality look solid. The stress-test note confirms the lifting preserves uniformity and the VC-dimension bound without external results, so the added term is justified and the disproof follows directly. There is no circularity; the recursion is the standard reduction to smaller parameters. Minor soft spots are limited to checking the boundary cases for small d and n, but the stated ranges (d >= 3, n >= d+3) are handled.\n\nThis is for people working on extremal problems in VC-dimension and uniform set systems. A reader who cares about exact values or constructions in this niche will find the explicit improvement useful. It deserves a serious referee because the central claim rests on a direct, checkable combinatorial argument rather than heavy machinery.","headline":"The paper gives a recursive lower bound via two-cover lifting that adds a positive term and disproves the Mubayi-Zhao conjecture for d >= 3.","tokens_in":2342,"tokens_out":405,"would_cite":false,"duration_ms":24265,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"A recursive construction gives a larger lower bound and disproves the Mubayi-Zhao conjecture on the size of uniform families with VC-dimension at most d for all d at least 3.","keywords":["uniform families","VC-dimension","recursive lower bounds","Mubayi-Zhao conjecture","extremal set theory","lifting construction","traces"],"falsifier":"An explicit (d+1)-uniform family on n points with VC-dimension at most d whose size exceeds binom(n-1,d) + binom(n-4,d-2) + binom(n-6,d-3) for some d >= 3 and n >= 2(d+2), or a direct verification that the lifting step fails to preserve the VC-dimension bound.","tokens_in":2660,"feed_emoji":"","tokens_out":803,"duration_ms":26796,"temperature":0.7,"pith_summary":"M_d(n) measures the largest number of (d+1)-subsets of an n-set whose trace on any d+1 points misses at least one pattern. Earlier work supplied the lower bound binom(n-1,d) plus binom(n-4,d-2). The paper adds a new recursive term M_{d-3}(n-5) obtained by lifting a smaller family. This produces a strictly stronger explicit lower bound and shows the conjecture claiming the earlier expression is already optimal must be false. A reader cares because the maximum size of these restricted families is now known to be larger than previously believed.","feed_headline":"Recursive bound falsifies Mubayi-Zhao conjecture for d>=3","feed_subtitle":"The maximum size of (d+1)-uniform families with VC-dimension at most d exceeds the earlier conjectured value by at least M_{d-3}(n-5).","key_machinery":"The two-cover lifting construction that takes a maximum (d-2)-uniform family of VC-dimension at most d-3 on n-5 points and produces a (d+1)-uniform family of VC-dimension at most d on n points.","core_discovery":"The authors prove that M_d(n) >= binom(n-1,d) + binom(n-4,d-2) + M_{d-3}(n-5) for d >= 3 and n >= d+3 via an explicit two-cover lifting construction. Applying the inequality once more yields the concrete improvement M_d(n) >= binom(n-1,d) + binom(n-4,d-2) + binom(n-6,d-3). This exceeds the value conjectured to be exact for n >= 2(d+2) and thereby falsifies the Mubayi-Zhao conjecture for every d >= 3.","pith_inferences":["The same lifting idea may extend to other pairs of uniformity and dimension.","The correct asymptotic growth rate of M_d(n) is now known to be strictly larger than the conjectured one.","Small-case computation for d=3 and moderate n could test whether the new bound is tight."],"forward_implications":["The Mubayi-Zhao conjecture is false for every d >= 3.","M_d(n) is at least as large as the previous conjectured value plus binom(n-6,d-3).","The exact value of M_d(n) remains undetermined but exceeds the earlier expression.","Repeated application of the recursion produces still larger lower bounds."],"fun_headline_variants":["Recursive bound disproves Mubayi-Zhao conjecture for d>=3","Two-cover construction exceeds Mubayi-Zhao value for d>=3","Improved M_d(n) lower bound falsifies VC-dimension conjecture","Explicit recursive inequality refutes Mubayi-Zhao bound"],"cache_read_input_tokens":2112,"weakest_assumption_plain":"The two-cover lifting construction applied to a maximum (d-2)-uniform family of VC-dimension at most d-3 on n-5 points produces a valid (d+1)-uniform family of VC-dimension at most d.","fun_headline_variants_meta":{"raw":{"variants":["Recursive bound disproves Mubayi-Zhao conjecture for d>=3","Two-cover construction exceeds Mubayi-Zhao value for d>=3","Improved M_d(n) lower bound falsifies VC-dimension conjecture","Explicit recursive inequality refutes Mubayi-Zhao bound"]},"model":"grok-4.3","cost_usd":0.00678,"raw_usage":{"total_tokens":3203,"prompt_tokens":767,"num_sources_used":0,"completion_tokens":63,"cost_in_usd_ticks":67799500,"prompt_tokens_details":{"text_tokens":767,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":2373,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":767,"tokens_out":63,"duration_ms":19622,"temperature":1.0,"reasoning_tokens":2373,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-06-26T11:56:19.234034+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"An explicit (d+1)-uniform family on n points with VC-dimension at most d whose size exceeds binom(n-1,d) + binom(n-4,d-2) + binom(n-6,d-3) for some d >= 3 and n >= 2(d+2), or a direct verification that the lifting step fails to preserve the VC-dimension bound.","supporting_citations":[],"review_version":1}