Private Distribution Learning with Public Data: The View from Sample Compression

Authors: Shai Ben-David, Alex Bie, Clément L Canonne, Gautam Kamath, Vikrant Singhal

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We study the problem of private distribution learning with access to public data. We show that the public-private learnability of a class Q is connected to the existence of a sample compression scheme for Q, as well as to an intermediate notion we refer to as list learning. Our work investigates the sample complexity of public-private learning, and does not give computationally efficient learners, or in some cases, algorithmic learners that run in finite time.
Researcher Affiliation Academia Shai Ben-David University of Waterloo Vector Institute shai@uwaterloo.ca Alex Bie University of Waterloo yabie@uwaterloo.ca Cl ement L. Canonne University of Sydney clement.canonne@sydney.edu.au Gautam Kamath University of Waterloo Vector Institute g@csail.mit.edu Vikrant Singhal University of Waterloo vikrant.singhal@uwaterloo.ca
Pseudocode No The paper does not include any structured pseudocode or algorithm blocks. Algorithmic steps are described in prose.
Open Source Code No The paper does not contain any explicit statements about releasing open-source code for the described methodology, nor does it provide any links to a code repository.
Open Datasets No The paper is theoretical and does not conduct empirical experiments that would involve using a publicly available dataset. No specific dataset is mentioned as being used for training or evaluation.
Dataset Splits No The paper is theoretical and does not describe empirical experiments. Therefore, there are no mentions of training, validation, or test dataset splits.
Hardware Specification No The paper is theoretical and does not conduct empirical experiments, therefore, no specific hardware specifications are mentioned.
Software Dependencies No The paper is theoretical and does not describe any computational experiments, thus it does not list any specific software dependencies with version numbers.
Experiment Setup No The paper is theoretical and does not describe any empirical experimental setup, including hyperparameters or system-level training settings.