Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Statement
Setting (p. 1). A -sunflower is a family of sets whose pairwise intersections are identical.
Theorem 1 (p. 1, quoted). "There is a universal constant such that every family of more than sets of size must contain a -sunflower."
The paper declares on p. 3 that all its logarithms are to base 2; in Theorem 1 the base only rescales . The constant is not made explicit. The introduction (p. 1) presents the result as a simpler proof of a bound similar to that of Alweiss, Lovett, Wu and Zhang, who showed that sets of size force a -sunflower; for comparison it recalls Erdős and Rado's bound and their family of sets of size with no -sunflower.
Source. Anup Rao, Coding for sunflowers, Discrete Analysis 2020:2, 8 pp., doi:10.19086/da.11887 (arXiv:1909.04774v2). Theorem 1 is on p. 1, its deduction from Lemma 2 on p. 2. Card: Rao 2020.
Read depth. Claims checked: the statement and the definition of a -sunflower were read clause by clause on the printed page. The deduction on p. 2 was read for structure only.
Proof pointer
Page 2, by induction on , with . For a family of more than distinct singletons contains of them, which form a -sunflower. For , if the family is not -spread, some nonempty lies in more than of its sets; removing from those sets and applying the induction hypothesis, using that does not decrease in , gives a -sunflower. If the family is -spread, Lemma 2 gives pairwise disjoint sets, which form a -sunflower.
Dependencies
Lemma 2 (p. 2), itself deduced from Lemma 4.
Bears on
- Problem 20: in the problem's notation, with the size of the sets and the size of the sunflower, Theorem 1 gives . The base grows with , so the bound is not of the form the problem asks for, and the theorem does not answer it.