Reconstruction under outliers for Fourier-sparse functions
Reconstruction under outliers for Fourier-sparse functions
复制标题
傅里叶稀疏函数异常值下的重建
DOI:
10.1137/1.9781611975994.124
复制
发表时间:
2020
期刊:
影响因子:
--
通讯作者:
De, Anindya
中科院分区:
文献类型:
--
作者:
Chen, Xue;De, Anindya
We consider the problem of learning an unknownfwith a sparse Fourier spectrum in the presence of outlier noise. In particular, the algorithm has access to a noisy oracle for (an unknown)fsuch that (i) the Fourier spectrum offisk-sparse; (ii) at any query pointx, the oracle returnsysuch that with probability 1 –ρ, |y–f(x)| ≤ε. However, with probability p, the errory–f(x) can be arbitrarily large.We study Fourier sparse functions over both the discrete cube {0, 1}nand the torus [0, 1) and for both these domains, we design efficient algorithms which can tolerate anyρ< 1/2 fraction of outliers. We note that the analogous problem for low-degree polynomials has recently been studied in several works [AK03, GZ16, KKP17] and similar algorithmic guarantees are known in that setting.While our main results pertain to the case where the location of the outliers, i.e.,xsuch that |y–f(x)| >εis randomly distributed, we also study the case where the outliers are adversarially located. In particular, we show that over the torus, assuming that the Fourier transform satisfies a certaingranularitycondition, there is a sample efficient algorithm to tolerateρ= Ω(1) fraction of outliers and further, that this is not possible without such a granularity condition. Finally, while not the principal thrust, our techniques also allow us non-trivially improve on learning low-degree functionsfon the hypercube in the presence of adversarial outlier noise.Our techniques combine a diverse array of tools from compressive sensing, sparse Fourier transform, chaining arguments and complex analysis.
登录
查看更多内容
DOI:
10.4230/oasics.sosa.2019.19
发表时间:
2018-09
期刊:
--
影响因子:
--
作者:
Sushrut Karmalkar;Eric Price
通讯作者:
Sushrut Karmalkar;Eric Price
DOI:
10.1145/3313276.3316363
发表时间:
2018
期刊:
Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing
影响因子:
--
作者:
I. Čižnár;A. Hoštacká;C. González;K. Krovacek
通讯作者:
K. Krovacek
DOI:
--
发表时间:
1997
期刊:
影响因子:
--
作者:
P. Borwein;T. Erdélyi
通讯作者:
T. Erdélyi
DOI:
--
发表时间:
--
期刊:
影响因子:
--
作者:
通讯作者:
--
DOI:
--
发表时间:
1990
期刊:
影响因子:
--
作者:
M. Talagrand
通讯作者:
M. Talagrand