
Distributionfree Junta Testing
We study the problem of testing whether an unknown nvariable Boolean fu...
02/13/2018 ∙ by Xi Chen, et al. ∙ 0 ∙ shareread it

Almost Optimal Distributionfree Junta Testing
We consider the problem of testing whether an unknown nvariable Boolean...
01/01/2019 ∙ by Nader H. Bshouty, et al. ∙ 0 ∙ shareread it

Is your function lowdimensional?
We study the problem of testing if a function depends on a small number ...
06/26/2018 ∙ by Anindya De, et al. ∙ 0 ∙ shareread it

Approximating the Distance to Monotonicity of Boolean Functions
We design a nonadaptive algorithm that, given a Boolean function f{0,1}^...
11/16/2019 ∙ by Ramesh Krishnan S. Pallavoor, et al. ∙ 0 ∙ shareread it

A Predictive Model using the Markov Property
Given a data set of numerical values which are sampled from some unknown...
01/08/2016 ∙ by Robert A. Murphy, et al. ∙ 0 ∙ shareread it

DistributionFree OnePass Learning
In many largescale machine learning applications, data are accumulated ...
06/08/2017 ∙ by Peng Zhao, et al. ∙ 0 ∙ shareread it

Testing noisy linear functions for sparsity
We consider the following basic inference problem: there is an unknown h...
11/03/2019 ∙ by Xue Chen, et al. ∙ 0 ∙ shareread it
DistributionFree Testing of Linear Functions on R^n
We study the problem of testing whether a function f:R^n>R is linear (i.e., both additive and homogeneous) in the distributionfree property testing model, where the distance between functions is measured with respect to an unknown probability distribution over R. We show that, given query access to f, sampling access to the unknown distribution as well as the standard Gaussian, and eps>0, we can distinguish additive functions from functions that are epsfar from additive functions with O((1/eps)log(1/eps)) queries, independent of n. Furthermore, under the assumption that f is a continuous function, the additivity tester can be extended to a distributionfree tester for linearity using the same number of queries. On the other hand, we show that if we are only allowed to get values of f on sampled points, then any distributionfree tester requires Omega(n) samples, even if the underlying distribution is the standard Gaussian.
READ FULL TEXT