Differentially private permutation tests: Applications to kernel methods

I Kim, A Schrab - arxiv preprint arxiv:2310.19043, 2023‏ - arxiv.org
Recent years have witnessed growing concerns about the privacy of sensitive data. In
response to these concerns, differential privacy has emerged as a rigorous framework for …

Monk outlier-robust mean embedding estimation by median-of-means

M Lerasle, Z Szabó, T Mathieu… - … conference on machine …, 2019‏ - proceedings.mlr.press
Mean embeddings provide an extremely flexible and powerful tool in machine learning and
statistics to represent probability distributions and define a semi-metric (MMD, maximum …

Robust kernel hypothesis testing under data corruption

A Schrab, I Kim - arxiv preprint arxiv:2405.19912, 2024‏ - arxiv.org
We propose two general methods for constructing robust permutation tests under data
corruption. The proposed tests effectively control the non-asymptotic type I error under data …

Federated experiment design under distributed differential privacy

WN Chen, G Cormode, A Bharadwaj… - International …, 2024‏ - proceedings.mlr.press
Experiment design has a rich history dating back over a century and has found many critical
applications across various fields since then. The use and collection of users' data in …

Multivariate mean comparison under differential privacy

M Dunsche, T Kutta, H Dette - International Conference on Privacy in …, 2022‏ - Springer
The comparison of multivariate population means is a central task of statistical inference.
While statistical theory provides a variety of analysis tools, they usually do not protect …

Minimax Optimal Two-Sample Testing under Local Differential Privacy

J Mun, S Kwak, I Kim - arxiv preprint arxiv:2411.09064, 2024‏ - arxiv.org
We explore the trade-off between privacy and statistical utility in private two-sample testing
under local differential privacy (LDP) for both multinomial and continuous data. We begin by …

M-estimation and Median of Means applied to statistical learning

T Mathieu - 2021‏ - theses.hal.science
The main objective of this thesis is to study methods for robust statistical learning.
Traditionally, in statistics we use models or simplifying assumptions that allow us to …

Application of kernel hypothesis testing on set-valued data

A Bellot, M van der Schaar - Uncertainty in Artificial …, 2021‏ - proceedings.mlr.press
We present a general framework for kernel hypothesis testing on distributions of sets of
individual examples. Sets may represent many common data sources such as groups of …

Kernel Hypothesis Testing with Set-valued Data

A Bellot, M van der Schaar - arxiv preprint arxiv:1907.04081, 2019‏ - arxiv.org
We present a general framework for hypothesis testing on distributions of sets of individual
examples. Sets may represent many common data sources such as groups of observations …

Hypothesis testing and causal inference with heterogeneous medical data

A Bellot - 2021‏ - repository.cam.ac.uk
Learning from data which associations hold and are likely to hold in the future is a
fundamental part of scientific discovery. With increasingly heterogeneous data collection …