The Power and Limitations of Uniform Samples in Testing Properties of Figures

Piotr Berman, Meiram Murzabulatov, Sofya Raskhodnikova

Research output: Contribution to journalArticle

Abstract

We investigate testing of properties of 2-dimensional figures that consist of a black object on a white background. Given a parameter ϵ∈ (0 , 1 / 2) , a tester for a specified property has to accept with probability at least 2/3 if the input figure satisfies the property and reject with probability at least 2/3 if it is ϵ-far from satisfying the property. In general, property testers can query the color of any point in the input figure. We study the power of testers that get access only to uniform samples from the input figure. We show that for the property of being a half-plane, the uniform testers are as powerful as general testers: they require only O(ϵ - 1 ) samples. In contrast, we prove that convexity can be tested with O(ϵ - 1 ) queries by testers that can make queries of their choice while uniform testers for this property require Ω(ϵ - 5 / 4 ) samples. Previously, the fastest known tester for convexity needed Θ(ϵ - 4 / 3 ) queries.

Original languageEnglish (US)
Pages (from-to)1247-1266
Number of pages20
JournalAlgorithmica
Volume81
Issue number3
DOIs
StatePublished - Mar 15 2019

Fingerprint

Property Testing
Figure
Testing
Query
Color
Convexity
Half-plane

All Science Journal Classification (ASJC) codes

  • Computer Science(all)
  • Computer Science Applications
  • Applied Mathematics

Cite this

Berman, Piotr ; Murzabulatov, Meiram ; Raskhodnikova, Sofya. / The Power and Limitations of Uniform Samples in Testing Properties of Figures. In: Algorithmica. 2019 ; Vol. 81, No. 3. pp. 1247-1266.
@article{b9e8c960acfa411aaac8d1cf5b5249dd,
title = "The Power and Limitations of Uniform Samples in Testing Properties of Figures",
abstract = "We investigate testing of properties of 2-dimensional figures that consist of a black object on a white background. Given a parameter ϵ∈ (0 , 1 / 2) , a tester for a specified property has to accept with probability at least 2/3 if the input figure satisfies the property and reject with probability at least 2/3 if it is ϵ-far from satisfying the property. In general, property testers can query the color of any point in the input figure. We study the power of testers that get access only to uniform samples from the input figure. We show that for the property of being a half-plane, the uniform testers are as powerful as general testers: they require only O(ϵ - 1 ) samples. In contrast, we prove that convexity can be tested with O(ϵ - 1 ) queries by testers that can make queries of their choice while uniform testers for this property require Ω(ϵ - 5 / 4 ) samples. Previously, the fastest known tester for convexity needed Θ(ϵ - 4 / 3 ) queries.",
author = "Piotr Berman and Meiram Murzabulatov and Sofya Raskhodnikova",
year = "2019",
month = "3",
day = "15",
doi = "10.1007/s00453-018-0467-9",
language = "English (US)",
volume = "81",
pages = "1247--1266",
journal = "Algorithmica",
issn = "0178-4617",
publisher = "Springer New York",
number = "3",

}

Berman, P, Murzabulatov, M & Raskhodnikova, S 2019, 'The Power and Limitations of Uniform Samples in Testing Properties of Figures', Algorithmica, vol. 81, no. 3, pp. 1247-1266. https://doi.org/10.1007/s00453-018-0467-9

The Power and Limitations of Uniform Samples in Testing Properties of Figures. / Berman, Piotr; Murzabulatov, Meiram; Raskhodnikova, Sofya.

In: Algorithmica, Vol. 81, No. 3, 15.03.2019, p. 1247-1266.

Research output: Contribution to journalArticle

TY - JOUR

T1 - The Power and Limitations of Uniform Samples in Testing Properties of Figures

AU - Berman, Piotr

AU - Murzabulatov, Meiram

AU - Raskhodnikova, Sofya

PY - 2019/3/15

Y1 - 2019/3/15

N2 - We investigate testing of properties of 2-dimensional figures that consist of a black object on a white background. Given a parameter ϵ∈ (0 , 1 / 2) , a tester for a specified property has to accept with probability at least 2/3 if the input figure satisfies the property and reject with probability at least 2/3 if it is ϵ-far from satisfying the property. In general, property testers can query the color of any point in the input figure. We study the power of testers that get access only to uniform samples from the input figure. We show that for the property of being a half-plane, the uniform testers are as powerful as general testers: they require only O(ϵ - 1 ) samples. In contrast, we prove that convexity can be tested with O(ϵ - 1 ) queries by testers that can make queries of their choice while uniform testers for this property require Ω(ϵ - 5 / 4 ) samples. Previously, the fastest known tester for convexity needed Θ(ϵ - 4 / 3 ) queries.

AB - We investigate testing of properties of 2-dimensional figures that consist of a black object on a white background. Given a parameter ϵ∈ (0 , 1 / 2) , a tester for a specified property has to accept with probability at least 2/3 if the input figure satisfies the property and reject with probability at least 2/3 if it is ϵ-far from satisfying the property. In general, property testers can query the color of any point in the input figure. We study the power of testers that get access only to uniform samples from the input figure. We show that for the property of being a half-plane, the uniform testers are as powerful as general testers: they require only O(ϵ - 1 ) samples. In contrast, we prove that convexity can be tested with O(ϵ - 1 ) queries by testers that can make queries of their choice while uniform testers for this property require Ω(ϵ - 5 / 4 ) samples. Previously, the fastest known tester for convexity needed Θ(ϵ - 4 / 3 ) queries.

UR - http://www.scopus.com/inward/record.url?scp=85052064646&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85052064646&partnerID=8YFLogxK

U2 - 10.1007/s00453-018-0467-9

DO - 10.1007/s00453-018-0467-9

M3 - Article

AN - SCOPUS:85052064646

VL - 81

SP - 1247

EP - 1266

JO - Algorithmica

JF - Algorithmica

SN - 0178-4617

IS - 3

ER -