Practical locally private heavy hitters

Raef Bassily; Kobbi Nissim; Uri Stemmer; Abhradeep Thakurta

Practical locally private heavy hitters

Raef Bassily, Kobbi Nissim, Uri Stemmer, Abhradeep Thakurta

Research output: Contribution to journal › Conference article › peer-review

111 Scopus citations

Abstract

We present new practical local differentially private heavy hitters algorithms achieving optimal or near-optimal worst-case error - TreeHist and Bitstogram. In both algorithms, server running time is Õ(n) and user running time is Õ(1), hence improving on the prior state-of-the-art result of Bassily and Smith [STOC 2015] requiring Õ(n⁵/²) server time and Õ(n³/²) user time. With a typically large number of participants in local algorithms (n in the millions), this reduction in time complexity, in particular at the user side, is crucial for the use of such algorithms in practice. We implemented Algorithm TreeHist to verify our theoretical analysis and compared its performance with the performance of Google's RAPPOR code.

Original language	English
Pages (from-to)	2289-2297
Number of pages	9
Journal	Advances in Neural Information Processing Systems
Volume	2017-December
State	Published - 1 Jan 2017
Externally published	Yes
Event	31st Annual Conference on Neural Information Processing Systems, NIPS 2017 - Long Beach, United States Duration: 4 Dec 2017 → 9 Dec 2017

ASJC Scopus subject areas

Computer Networks and Communications
Information Systems
Signal Processing

Cite this

@article{8b9c001db033462aaf0c838200810755,

title = "Practical locally private heavy hitters",

abstract = "We present new practical local differentially private heavy hitters algorithms achieving optimal or near-optimal worst-case error - TreeHist and Bitstogram. In both algorithms, server running time is {\~O}(n) and user running time is {\~O}(1), hence improving on the prior state-of-the-art result of Bassily and Smith [STOC 2015] requiring {\~O}(n5/2) server time and {\~O}(n3/2) user time. With a typically large number of participants in local algorithms (n in the millions), this reduction in time complexity, in particular at the user side, is crucial for the use of such algorithms in practice. We implemented Algorithm TreeHist to verify our theoretical analysis and compared its performance with the performance of Google's RAPPOR code.",

author = "Raef Bassily and Kobbi Nissim and Uri Stemmer and Abhradeep Thakurta",

note = "Publisher Copyright: {\textcopyright} 2017 Neural information processing systems foundation. All rights reserved.; 31st Annual Conference on Neural Information Processing Systems, NIPS 2017 ; Conference date: 04-12-2017 Through 09-12-2017",

year = "2017",

month = jan,

day = "1",

language = "English",

volume = "2017-December",

pages = "2289--2297",

journal = "Advances in Neural Information Processing Systems",

issn = "1049-5258",

}

TY - JOUR

T1 - Practical locally private heavy hitters

AU - Bassily, Raef

AU - Nissim, Kobbi

AU - Stemmer, Uri

AU - Thakurta, Abhradeep

PY - 2017/1/1

Y1 - 2017/1/1

N2 - We present new practical local differentially private heavy hitters algorithms achieving optimal or near-optimal worst-case error - TreeHist and Bitstogram. In both algorithms, server running time is Õ(n) and user running time is Õ(1), hence improving on the prior state-of-the-art result of Bassily and Smith [STOC 2015] requiring Õ(n5/2) server time and Õ(n3/2) user time. With a typically large number of participants in local algorithms (n in the millions), this reduction in time complexity, in particular at the user side, is crucial for the use of such algorithms in practice. We implemented Algorithm TreeHist to verify our theoretical analysis and compared its performance with the performance of Google's RAPPOR code.

AB - We present new practical local differentially private heavy hitters algorithms achieving optimal or near-optimal worst-case error - TreeHist and Bitstogram. In both algorithms, server running time is Õ(n) and user running time is Õ(1), hence improving on the prior state-of-the-art result of Bassily and Smith [STOC 2015] requiring Õ(n5/2) server time and Õ(n3/2) user time. With a typically large number of participants in local algorithms (n in the millions), this reduction in time complexity, in particular at the user side, is crucial for the use of such algorithms in practice. We implemented Algorithm TreeHist to verify our theoretical analysis and compared its performance with the performance of Google's RAPPOR code.

UR - http://www.scopus.com/inward/record.url?scp=85047004408&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85047004408

SN - 1049-5258

VL - 2017-December

SP - 2289

EP - 2297

JO - Advances in Neural Information Processing Systems

JF - Advances in Neural Information Processing Systems

T2 - 31st Annual Conference on Neural Information Processing Systems, NIPS 2017

Y2 - 4 December 2017 through 9 December 2017

ER -

Practical locally private heavy hitters

Abstract

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this