A Small PRG for Polynomial Threshold Functions of Gaussians

  • Authors:
  • Daniel M. Kane

  • Affiliations:
  • -

  • Venue:
  • FOCS '11 Proceedings of the 2011 IEEE 52nd Annual Symposium on Foundations of Computer Science
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

We discuss a psuedorandom generator to $\epsilon$-fool degree-d polynomial threshold functions with respect to the Gaussian distribution with seed length $2^O_c(d)\log(n) \epsilon^{-4-c}$. Our analysis involves several new ideas including what we term the "noisy derivative" of a function and a stronger version of standard anticoncentration results.