Differentially Private Post-Processing for Fair Regression

Authors: Ruicheng Xian, Qiaobo Li, Gautam Kamath, Han Zhao

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In Section 4, we empirically explore the trade-offs achieved by our post-processing algorithm on Law School and Communities & Crime datasets.
Researcher Affiliation Academia 1University of Illinois Urbana-Champaign 2University of Waterloo and Vector Institute.
Pseudocode Yes Algorithm 1 Fair and Private Post-Processing (Attribute-Aware)
Open Source Code Yes Code is available at https://github.com/rxian/fair-regression.
Open Datasets Yes Communities & Crime (Redmond & Baveja, 2002). ... Law School (Wightman, 1998).
Dataset Splits No The paper states, 'The datasets are randomly split 70-30 for training (i.e., post-processing) and testing,' but does not provide details for a separate validation split used in their specific experimental setup.
Hardware Specification No The paper does not provide any specific details about the hardware used to run the experiments, such as CPU or GPU models, memory, or cloud instance types.
Software Dependencies No The paper does not explicitly list any software dependencies with specific version numbers required for reproducibility.
Experiment Setup Yes Algorithm 1... requires specifying an interval [s, t], number of bins k, fairness tolerance α, privacy budget ε.