English

Jackson's inequality on the hypercube

Functional Analysis 2024-10-29 v1 Combinatorics Probability

Abstract

We investigate the best constant J(n,d)J(n,d) such that Jackson's inequality infdeg(g)dfgJ(n,d)s(f), \inf_{\mathrm{deg}(g) \leq d} \|f - g\|_{\infty} \leq J(n,d) \, s(f), holds for all functions ff on the hypercube {0,1}n\{0,1\}^n, where s(f)s(f) denotes the sensitivity of ff. We show that the quantity J(n,0.499n)J(n, 0.499n) is bounded below by an absolute positive constant, independent of nn. This complements Wagner's theorem, which establishes that J(n,d)1J(n,d)\leq 1 . As a first application we show that reverse Bernstein inequality fails in the tail space L0.499n1L^{1}_{\geq 0.499n} improving over previously known counterexamples in LCloglog(n)1L^{1}_{\geq C \log \log (n)}. As a second application, we show that there exists a function f:{0,1}n[1,1]f : \{0,1\}^n \to [-1,1] whose sensitivity s(f)s(f) remains constant, independent of nn, while the approximate degree grows linearly with nn. This result implies that the sensitivity theorem s(f)Ω(deg(f)C)s(f) \geq \Omega(\mathrm{deg}(f)^C) fails in the strongest sense for bounded real-valued functions even when deg(f)\mathrm{deg}(f) is relaxed to the approximate degree. We also show that in the regime d=(1δ)nd = (1 - \delta)n, the bound J(n,d)Cmin{δ,max{δ2,n2/3}} J(n,d) \leq C \min\{\delta, \max\{\delta^2, n^{-2/3}\}\} holds. Moreover, when restricted to symmetric real-valued functions, we obtain Jsymmetric(n,d)C/dJ_{\mathrm{symmetric}}(n,d) \leq C/d and the decay 1/d1/d is sharp. Finally, we present results for a subspace approximation problem: we show that there exists a subspace EE of dimension 2n12^{n-1} such that infgEfgs(f)/n\inf_{g \in E} \|f - g\|_{\infty} \leq s(f)/n holds for all ff.

Cite

@article{arxiv.2410.19949,
  title  = {Jackson's inequality on the hypercube},
  author = {Paata Ivanisvili and Roman Vershynin and Xinyuan Xie},
  journal= {arXiv preprint arXiv:2410.19949},
  year   = {2024}
}
R2 v1 2026-06-28T19:36:11.256Z