English

Fourier-sparse interpolation without a frequency gap

Data Structures and Algorithms 2016-09-07 v1

Abstract

We consider the problem of estimating a Fourier-sparse signal from noisy samples, where the sampling is done over some interval [0,T][0, T] and the frequencies can be "off-grid". Previous methods for this problem required the gap between frequencies to be above 1/T, the threshold required to robustly identify individual frequencies. We show the frequency gap is not necessary to estimate the signal as a whole: for arbitrary kk-Fourier-sparse signals under 2\ell_2 bounded noise, we show how to estimate the signal with a constant factor growth of the noise and sample complexity polynomial in kk and logarithmic in the bandwidth and signal-to-noise ratio. As a special case, we get an algorithm to interpolate degree dd polynomials from noisy measurements, using O(d)O(d) samples and increasing the noise by a constant factor in 2\ell_2.

Keywords

Cite

@article{arxiv.1609.01361,
  title  = {Fourier-sparse interpolation without a frequency gap},
  author = {Xue Chen and Daniel M. Kane and Eric Price and Zhao Song},
  journal= {arXiv preprint arXiv:1609.01361},
  year   = {2016}
}

Comments

FOCS 2016

R2 v1 2026-06-22T15:40:41.792Z