Practical Differentially Private Top-k Selection with Pay-what-you-get Composition

Authors: David Durfee, Ryan M. Rogers

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical We design algorithms that ensures (approximate) (ε, δ > 0)-differential privacy and only needs access to the true topk elements from the data for any chosen k k. We design different algorithms for either setting so that the privacy parameter ε needs to scale with either in the restricted sensitivity setting or k in the unrestricted setting. Additionally, we provide a pay-what-you-get privacy composition bound for our algorithms. That is, our algorithms might return fewer than k elements when the top-k elements are queried, but the overall privacy budget only decreases by the size of the outcome.
Researcher Affiliation Industry David Durfee1 and Ryan Rogers1 1Data Science Applied Research, Linked In
Pseudocode Yes Algorithm 1 Limit Domk, k; Top-k from the k k limited domain Algorithm 2 multi Limit Domk ,ℓ ; Multiple queries to limited domain
Open Source Code No The paper does not provide any explicit statements about releasing source code or links to a code repository for the methodology described.
Open Datasets No The paper is theoretical and primarily focuses on algorithm design, privacy guarantees, and composition theorems. It does not conduct empirical experiments that involve training models on datasets.
Dataset Splits No The paper is theoretical and focuses on algorithm design and privacy guarantees. It does not conduct empirical experiments that would require validation splits.
Hardware Specification No The paper is theoretical and does not describe empirical experiments. Therefore, no hardware specifications for running experiments are provided.
Software Dependencies No The paper is theoretical and does not describe empirical experiments with specific software implementations. Therefore, no specific software dependencies with version numbers are mentioned.
Experiment Setup No The paper is theoretical and focuses on algorithm design and privacy guarantees. It does not detail an experimental setup with hyperparameters or training configurations, as no empirical experiments are conducted.