The Gotsman–Linial conjecture for average sensitivity of polynomial threshold functions

Let f:{1,1}n{1,1}f:\{-1,1\}^n\to\{-1,1\} be a polynomial threshold function of degree at most dd, abbreviated an (n,d)(n,d)-PTF. For a Boolean function ff, let AS[f]\mathbf{AS}[f] denote its average sensitivity. Let pn,dp^*_{n,d} be the monic univariate polynomial of degree dd with non-repeated roots at the dd integers closest to 00 having parity opposite to nn, and define

fn,d(x1,,xn)=sgn(pn,d(i[n]xi)).f^*_{n,d}(x_1,\ldots,x_n)=\operatorname{sgn}\left(p^*_{n,d}\left(\sum_{i\in[n]}x_i\right)\right).

Gotsman–Linial conjecture. For every (n,d)(n,d)-PTF ff,

AS[f]AS[fn,d].\mathbf{AS}[f]\leq\mathbf{AS}[f^*_{n,d}].

Gotsman and Linial proved the claim for d=0d=0 and d=1d=1, but the paper shows that it is false for almost all relevant pairs (n,d)(n,d), with a multiplicative separation of 1+Θd(n1)1+\Theta_d(n^{-1}).

Sources & referencesView supporting material

Primary source

Brynmor Chapman, “The Gotsman-Linial Conjecture is False”, arXiv:2108.02288 (2021).

Additional references

4 papers in this index state this conjecture (2012–2021). The statement above is taken from the most recent of them; the others are arXiv:2105.10386, arXiv:1210.1283, arXiv:1204.0543.

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.