THE SAMPLE COMPLEXITY OF DISTRIBUTION-FREE PARITY LEARNING IN THE ROBUST SHUFFLE MODEL

Kobbi Nissim; Chao Yan

doi:10.29012/jpc.805

THE SAMPLE COMPLEXITY OF DISTRIBUTION-FREE PARITY LEARNING IN THE ROBUST SHUFFLE MODEL

Kobbi Nissim, Chao Yan

Research output: Contribution to journal › Article › peer-review

Abstract

We provide a lower bound on the sample complexity of distribution-free parity learning in the realizable case in the shuffle model of differential privacy. Namely, we show that the sample complexity of learning d-bit parity functions is Ω(2^d/2). Our result extends a recent similar lower bound on the sample complexity of private agnostic learning of parity functions in the shuffle model by Cheu and Ullman (12). We also sketch a simple shuffle model protocol demonstrating that our results are tight up to poly(d) factors.

Original language	English
Pages (from-to)	1-14
Number of pages	14
Journal	Journal of Privacy and Confidentiality
Volume	12
Issue number	2
DOIs	https://doi.org/10.29012/jpc.805
State	Published - 1 Jan 2022
Externally published	Yes

Keywords

Differential privacy
parity learning
private learning

ASJC Scopus subject areas

Computer Science (miscellaneous)
Statistics and Probability
Computer Science Applications

Access to Document

10.29012/jpc.805

Cite this

@article{c19bdcac3de8453b901dbe5f3df721ca,

title = "THE SAMPLE COMPLEXITY OF DISTRIBUTION-FREE PARITY LEARNING IN THE ROBUST SHUFFLE MODEL",

abstract = "We provide a lower bound on the sample complexity of distribution-free parity learning in the realizable case in the shuffle model of differential privacy. Namely, we show that the sample complexity of learning d-bit parity functions is Ω(2d/2). Our result extends a recent similar lower bound on the sample complexity of private agnostic learning of parity functions in the shuffle model by Cheu and Ullman (12). We also sketch a simple shuffle model protocol demonstrating that our results are tight up to poly(d) factors.",

keywords = "Differential privacy, parity learning, private learning",

author = "Kobbi Nissim and Chao Yan",

year = "2022",

month = jan,

day = "1",

doi = "10.29012/jpc.805",

language = "English",

volume = "12",

pages = "1--14",

journal = "Journal of Privacy and Confidentiality",

issn = "2575-8527",

publisher = "Cornell University",

number = "2",

}

TY - JOUR

T1 - THE SAMPLE COMPLEXITY OF DISTRIBUTION-FREE PARITY LEARNING IN THE ROBUST SHUFFLE MODEL

AU - Nissim, Kobbi

AU - Yan, Chao

PY - 2022/1/1

Y1 - 2022/1/1

N2 - We provide a lower bound on the sample complexity of distribution-free parity learning in the realizable case in the shuffle model of differential privacy. Namely, we show that the sample complexity of learning d-bit parity functions is Ω(2d/2). Our result extends a recent similar lower bound on the sample complexity of private agnostic learning of parity functions in the shuffle model by Cheu and Ullman (12). We also sketch a simple shuffle model protocol demonstrating that our results are tight up to poly(d) factors.

AB - We provide a lower bound on the sample complexity of distribution-free parity learning in the realizable case in the shuffle model of differential privacy. Namely, we show that the sample complexity of learning d-bit parity functions is Ω(2d/2). Our result extends a recent similar lower bound on the sample complexity of private agnostic learning of parity functions in the shuffle model by Cheu and Ullman (12). We also sketch a simple shuffle model protocol demonstrating that our results are tight up to poly(d) factors.

KW - Differential privacy

KW - parity learning

KW - private learning

UR - http://www.scopus.com/inward/record.url?scp=85140993667&partnerID=8YFLogxK

U2 - 10.29012/jpc.805

DO - 10.29012/jpc.805

M3 - Article

AN - SCOPUS:85140993667

SN - 2575-8527

VL - 12

SP - 1

EP - 14

JO - Journal of Privacy and Confidentiality

JF - Journal of Privacy and Confidentiality

IS - 2

ER -

THE SAMPLE COMPLEXITY OF DISTRIBUTION-FREE PARITY LEARNING IN THE ROBUST SHUFFLE MODEL

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this