On Differential Privacy and Adaptive Data Analysis with Bounded Space

Itai Dinur; Uri Stemmer; David P. Woodruff; Samson Zhou

doi:https://doi.org/10.1007/978-3-031-30620-4_2

On Differential Privacy and Adaptive Data Analysis with Bounded Space

Itai Dinur, Uri Stemmer, David P. Woodruff, Samson Zhou

Ben-Gurion University of the Negev, Tel Aviv University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

We study the space complexity of the two related fields of differential privacy and adaptive data analysis. Specifically, 1.Under standard cryptographic assumptions, we show that there exists a problem P that requires exponentially more space to be solved efficiently with differential privacy, compared to the space needed without privacy. To the best of our knowledge, this is the first separation between the space complexity of private and non-private algorithms.2.The line of work on adaptive data analysis focuses on understanding the number of samples needed for answering a sequence of adaptive queries. We revisit previous lower bounds at a foundational level, and show that they are a consequence of a space bottleneck rather than a sampling bottleneck. To obtain our results, we define and construct an encryption scheme with multiple keys that is built to withstand a limited amount of key leakage in a very particular way.

Original language	American English
Title of host publication	Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings
Editors	Carmit Hazay, Martijn Stam
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	35-65
Number of pages	31
ISBN (Print)	9783031306198
DOIs	https://doi.org/10.1007/978-3-031-30620-4_2
State	Published - 1 Jan 2023
Event	42nd Annual International Conference on Theory and Applications of Cryptographic Techniques, EUROCRYPT 2023 - Lyon, France Duration: 23 Apr 2023 → 27 Apr 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14006 LNCS

Conference

Conference	42nd Annual International Conference on Theory and Applications of Cryptographic Techniques, EUROCRYPT 2023
Country/Territory	France
City	Lyon
Period	23/04/23 → 27/04/23

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Access to Document

https://doi.org/10.1007/978-3-031-30620-4_2

Cite this

Dinur, I., Stemmer, U., Woodruff, D. P., & Zhou, S. (2023). On Differential Privacy and Adaptive Data Analysis with Bounded Space. In C. Hazay, & M. Stam (Eds.), Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings (pp. 35-65). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14006 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-30620-4_2

Dinur, Itai ; Stemmer, Uri ; Woodruff, David P. et al. / On Differential Privacy and Adaptive Data Analysis with Bounded Space. Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings. editor / Carmit Hazay ; Martijn Stam. Springer Science and Business Media Deutschland GmbH, 2023. pp. 35-65 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{c2880643ae0f4f0e9d59d2f8d1fd9bd6,

title = "On Differential Privacy and Adaptive Data Analysis with Bounded Space",

abstract = "We study the space complexity of the two related fields of differential privacy and adaptive data analysis. Specifically, 1.Under standard cryptographic assumptions, we show that there exists a problem P that requires exponentially more space to be solved efficiently with differential privacy, compared to the space needed without privacy. To the best of our knowledge, this is the first separation between the space complexity of private and non-private algorithms.2.The line of work on adaptive data analysis focuses on understanding the number of samples needed for answering a sequence of adaptive queries. We revisit previous lower bounds at a foundational level, and show that they are a consequence of a space bottleneck rather than a sampling bottleneck. To obtain our results, we define and construct an encryption scheme with multiple keys that is built to withstand a limited amount of key leakage in a very particular way.",

author = "Itai Dinur and Uri Stemmer and Woodruff, {David P.} and Samson Zhou",

note = "Publisher Copyright: {\textcopyright} 2023, International Association for Cryptologic Research.; 42nd Annual International Conference on Theory and Applications of Cryptographic Techniques, EUROCRYPT 2023 ; Conference date: 23-04-2023 Through 27-04-2023",

year = "2023",

month = jan,

day = "1",

doi = "https://doi.org/10.1007/978-3-031-30620-4_2",

language = "American English",

isbn = "9783031306198",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "35--65",

editor = "Carmit Hazay and Martijn Stam",

booktitle = "Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings",

address = "Germany",

}

Dinur, I , Stemmer, U, Woodruff, DP & Zhou, S 2023, On Differential Privacy and Adaptive Data Analysis with Bounded Space. in C Hazay & M Stam (eds), Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14006 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 35-65, 42nd Annual International Conference on Theory and Applications of Cryptographic Techniques, EUROCRYPT 2023, Lyon, France, 23/04/23. https://doi.org/10.1007/978-3-031-30620-4_2

On Differential Privacy and Adaptive Data Analysis with Bounded Space. / Dinur, Itai ; Stemmer, Uri; Woodruff, David P. et al.
Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings. ed. / Carmit Hazay; Martijn Stam. Springer Science and Business Media Deutschland GmbH, 2023. p. 35-65 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14006 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - On Differential Privacy and Adaptive Data Analysis with Bounded Space

AU - Dinur, Itai

AU - Stemmer, Uri

AU - Woodruff, David P.

AU - Zhou, Samson

PY - 2023/1/1

Y1 - 2023/1/1

N2 - We study the space complexity of the two related fields of differential privacy and adaptive data analysis. Specifically, 1.Under standard cryptographic assumptions, we show that there exists a problem P that requires exponentially more space to be solved efficiently with differential privacy, compared to the space needed without privacy. To the best of our knowledge, this is the first separation between the space complexity of private and non-private algorithms.2.The line of work on adaptive data analysis focuses on understanding the number of samples needed for answering a sequence of adaptive queries. We revisit previous lower bounds at a foundational level, and show that they are a consequence of a space bottleneck rather than a sampling bottleneck. To obtain our results, we define and construct an encryption scheme with multiple keys that is built to withstand a limited amount of key leakage in a very particular way.

AB - We study the space complexity of the two related fields of differential privacy and adaptive data analysis. Specifically, 1.Under standard cryptographic assumptions, we show that there exists a problem P that requires exponentially more space to be solved efficiently with differential privacy, compared to the space needed without privacy. To the best of our knowledge, this is the first separation between the space complexity of private and non-private algorithms.2.The line of work on adaptive data analysis focuses on understanding the number of samples needed for answering a sequence of adaptive queries. We revisit previous lower bounds at a foundational level, and show that they are a consequence of a space bottleneck rather than a sampling bottleneck. To obtain our results, we define and construct an encryption scheme with multiple keys that is built to withstand a limited amount of key leakage in a very particular way.

UR - http://www.scopus.com/inward/record.url?scp=85161635916&partnerID=8YFLogxK

U2 - https://doi.org/10.1007/978-3-031-30620-4_2

DO - https://doi.org/10.1007/978-3-031-30620-4_2

M3 - Conference contribution

SN - 9783031306198

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 35

EP - 65

BT - Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings

A2 - Hazay, Carmit

A2 - Stam, Martijn

PB - Springer Science and Business Media Deutschland GmbH

T2 - 42nd Annual International Conference on Theory and Applications of Cryptographic Techniques, EUROCRYPT 2023

Y2 - 23 April 2023 through 27 April 2023

ER -

Dinur I , Stemmer U, Woodruff DP, Zhou S. On Differential Privacy and Adaptive Data Analysis with Bounded Space. In Hazay C, Stam M, editors, Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 35-65. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: https://doi.org/10.1007/978-3-031-30620-4_2

On Differential Privacy and Adaptive Data Analysis with Bounded Space

Abstract

Publication series

Conference

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this