Learning discrete distributions: user vs item-level privacy

Authors: Yuhan Liu, Ananda Theertha Suresh, Felix Xinnan X. Yu, Sanjiv Kumar, Michael Riley

NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show empirically that the bounds in Theorem 5 should hold by estimating the ℓ1 distance between Bin(m, 0.01) and Bin(m, 0.011).
Researcher Affiliation Collaboration Yuhan Liu Cornell University yl2976@cornell.edu Ananda Theertha Suresh Google Research theertha@google.com Felix Yu Google Research felixyu@google.com Sanjiv Kumar Google Research sanjivk@google.com Michael Riley Google Research riley@google.com
Pseudocode Yes Algorithm 1 Private hypothesis selection: PHS(H, D, α, ε); Algorithm 2 Learning binomial distributions: Binom(D, ε, α); Algorithm 3 Dense regime: Dense(D, ε, δ, α); Algorithm 4 Estimation of binomial with small p: Small Binom(D, ε); Algorithm 5 Sparse regime: Sparse(D, ε, δ, α)
Open Source Code No The paper does not contain any explicit statement about releasing source code, nor does it provide a link to a code repository.
Open Datasets No The paper does not specify the use of any publicly available or open datasets for training or evaluation. The empirical demonstration in Figure 1 is based on synthetic data for illustrating a mathematical bound.
Dataset Splits No The paper does not describe any specific training, validation, or test dataset splits. The empirical evaluation shown is for a mathematical bound, not a model trained on a dataset.
Hardware Specification No The paper does not specify any hardware used for the empirical approximation shown in Figure 1, such as GPU/CPU models or cloud resources.
Software Dependencies No The paper states that the ℓ1 distance is approximated "by samples" for Figure 1, but it does not specify any software names or version numbers used for this approximation.
Experiment Setup No The paper does not provide specific details about any experimental setup, such as hyperparameter values, training configurations, or system-level settings. The empirical part is a simple approximation for a mathematical bound without detailed setup.