Roman Vershynin

Roman Vershynin

University of California, Irvine

H-index: 52

North America-United States

About Roman Vershynin

Roman Vershynin, With an exceptional h-index of 52 and a recent h-index of 39 (since 2020), a distinguished researcher at University of California, Irvine, specializes in the field of Probability, Data Science.

His recent articles reflect a diverse array of research interests and contributions to the field:

PrivStream: An Algorithm for Streaming Differentially Private Data

Metric geometry of the privacy-utility tradeoff

Are most Boolean functions determined by low frequencies?

Private measures, random walks, and synthetic data

An algorithm for streaming differentially private data

Online Differentially Private Synthetic Data Generation

Hamiltonicity of Sparse Pseudorandom Graphs

Covariance’s loss is privacy’s gain: Computationally efficient, private and accurate synthetic data

Roman Vershynin Information

University

Position

Professor of Mathematics

Citations(all)

20704

Citations(since 2020)

10649

Cited By

14758

hIndex(all)

52

hIndex(since 2020)

39

i10Index(all)

86

i10Index(since 2020)

71

Email

University Profile Page

University of California, Irvine

Google Scholar

View Google Scholar Profile

Roman Vershynin Skills & Research Interests

Probability

Data Science

Top articles of Roman Vershynin

Title

Journal

Author(s)

Publication Date

PrivStream: An Algorithm for Streaming Differentially Private Data

arXiv preprint arXiv:2401.14577

Girish Kumar

Thomas Strohmer

Roman Vershynin

2024/1/26

Metric geometry of the privacy-utility tradeoff

arXiv preprint arXiv:2405.00329

March Boedihardjo

Thomas Strohmer

Roman Vershynin

2024/5/1

Are most Boolean functions determined by low frequencies?

arXiv preprint arXiv:2401.13143

Roman Vershynin

2024/1/23

Private measures, random walks, and synthetic data

Probability Theory and Related Fields

March Boedihardjo

Thomas Strohmer

Roman Vershynin

2024/4/20

An algorithm for streaming differentially private data

arXiv preprint arXiv:2401.14577

Girish Kumar

Thomas Strohmer

Roman Vershynin

2024

Online Differentially Private Synthetic Data Generation

arXiv preprint arXiv:2402.08012

Yiyun He

Roman Vershynin

Yizhe Zhu

2024/2/12

Hamiltonicity of Sparse Pseudorandom Graphs

arXiv preprint arXiv:2402.06177

Asaf Ferber

Jie Han

Dingjia Mao

Roman Vershynin

2024/2/9

Covariance’s loss is privacy’s gain: Computationally efficient, private and accurate synthetic data

Foundations of Computational Mathematics

March Boedihardjo

Thomas Strohmer

Roman Vershynin

2024/2

Online stochastic gradient descent with arbitrary initialization solves non-smooth, non-convex phase retrieval

Journal of Machine Learning Research

Yan Shuo Tan

Roman Vershynin

2023

Algorithmically effective differentially private synthetic data

Yiyun He

Roman Vershynin

Yizhe Zhu

2023

The quarks of attention: Structure and capacity of neural attention building blocks

Artificial Intelligence

Pierre Baldi

Roman Vershynin

2023/6/1

Differentially private low-dimensional representation of high-dimensional data

arXiv preprint arXiv:2305.17148

Yiyun He

Thomas Strohmer

Roman Vershynin

Yizhe Zhu

2023/5/26

AVIDA: An alternating method for visualizing and integrating data

Journal of Computational Science

Kathryn Dover

Zixuan Cang

Anna Ma

Qing Nie

Roman Vershynin

2023/4/1

Covering the hypercube, the uncertainty principle, and an interpolation formula

arXiv preprint arXiv:2310.13277

Paata Ivanisvili

Ohad Klein

Roman Vershynin

2023/10/20

Covariance loss, Szemeredi regularity, and differential privacy

arXiv preprint arXiv:2301.02705

March Boedihardjo

Thomas Strohmer

Roman Vershynin

2023/1/6

Differentially Private Low-dimensional Synthetic Data from High-dimensional Datasets

Yiyun He

Thomas Strohmer

Roman Vershynin

Yizhe Zhu

2023/10/13

Privacy of synthetic data: A statistical framework

IEEE Transactions on information theory

March Boedihardjo

Thomas Strohmer

Roman Vershynin

2022/10/25

Foundations of Attention Mechanisms in Deep Neural Network Architectures

Pierre Baldi

Roman Vershynin

2022/10/21

The quarks of attention

arXiv preprint arXiv:2202.08371

Pierre Baldi

Roman Vershynin

2022/2/15

Private sampling: a noiseless approach for generating differentially private synthetic data

SIAM Journal on Mathematics of Data Science

March Boedihardjo

Thomas Strohmer

Roman Vershynin

2022

See List of Professors in Roman Vershynin University(University of California, Irvine)

Co-Authors

H-index: 132
Pierre Baldi

Pierre Baldi

University of California, Irvine

H-index: 108
Terence Tao

Terence Tao

University of California, Los Angeles

H-index: 104
Emmanuel Candes

Emmanuel Candes

Stanford University

H-index: 61
Joel A. Tropp

Joel A. Tropp

California Institute of Technology

H-index: 42
Elizaveta (Liza) Levina

Elizaveta (Liza) Levina

University of Michigan-Dearborn

academic-engine