Inducibility upper bound 1/e for non-trivial graphs
Conjecture 1.2 (The Large Inducibility Conjecture) · arXiv:1805.06848
Status solved high confidence
The Large Inducibility Conjecture is settled as a corollary of the Edge-statistics Conjecture (Conjecture 1.1 of arXiv:1805.06848), which the source paper explicitly states implies it. The Edge-statistics Conjecture was fully proved by the combination of Kwan, Sudakov, and Tran (handling the intermediate range) and Fox and Sauermann (arXiv:1809.01352, Advances in Combinatorics 2020, handling the remaining cases 1 ≤ ℓ ≤ C·k). The 2024 paper arXiv:2411.17362 further characterizes which non-trivial graphs achieve inducibility close to 1/e, strengthening the structural picture.
Cited literature (2)
-
Proves the full Edge-statistics Conjecture (Conjecture 1.1) by handling the cases 1 ≤ ℓ ≤ C·k; since the source paper explicitly states the Large Inducibility Conjecture is implied by the Edge-statistics Conjecture, this settles Conjecture 1.2 as a corollary.
-
Characterizes which graphs achieve inducibility close to 1/e: establishes ind(H) ≤ 1/e + o_k(1) for all H ∉ {K_k, ĀK_k} and shows that graphs other than stars and their complements have inducibility bounded away from 1/e by an absolute constant.
Reviewer notes. The conjecture is implicitly resolved via the proved Edge-statistics Conjecture: the source paper states Conjecture 1.2 is implied by Conjecture 1.1, and the latter is now fully proved (Fox-Sauermann 2020 + Kwan-Sudakov-Tran, arXiv:1809.02576, not independently WebFetched). Fox-Sauermann do not explicitly name or cite Conjecture 1.2, so the resolution is by logical implication rather than an explicit statement in the follow-up papers. Author names for arXiv:2411.17362 were not extracted from the WebFetch response.
Context
This conjecture for graph inducibilities is presented as an analogue of the Edge-statistics Conjecture (Conjecture 1.1), and would be implied by it. It asserts that no graph other than a complete or empty graph can have inducibility exceeding $1/e$ in the limit over large host graphs.
Source paper
Edge-statistics on large graphs
Noga Alon, Dan Hefetz, Michael Krivelevich, Mykhaylo Tyomkyn · 2019-11-01
https://arxiv.org/abs/1805.06848
PDF source