Privacy Amplification by Iteration for ADMM with (Strongly) Convex Objective Functions

Authors: T-H. Hubert Chan, Hao Xie, Mengshi Zhao

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

Reproducibility Variable Result LLM Response
Research Type Experimental Despite being primarily theoretical, we conduct experiments on a general Lasso problem. Specifically, we empirically examine the effects of strong convexity and privacy noise magnitude on convergence rates.
Researcher Affiliation Academia T-H. Hubert Chan*, Hao Xie*, Mengshi Zhao* Department of Computer Science, The University of Hong Kong hubert@cs.hku.hk, hxie@connect.hku.hk, zmsxsl@connect.hku.hk
Pseudocode Yes Algorithm 1: One ADMM Iteration. Algorithm 2: Mechanism M1. Algorithm 3: Mechanism M2.
Open Source Code No The paper does not include an unambiguous statement that the authors are releasing the code for the work described in this paper, nor does it provide a direct link to a source-code repository.
Open Datasets No The paper mentions conducting 'experiments on a general Lasso problem' and a 'numerical illustration of our algorithms on a general Lasso problem'. This describes the problem type but does not specify a concrete, publicly available dataset with a link, DOI, or citation.
Dataset Splits No The paper does not provide specific details on training, validation, or test dataset splits (e.g., percentages, sample counts, or references to predefined splits).
Hardware Specification No The paper does not explicitly describe any specific hardware specifications (e.g., GPU/CPU models, memory amounts, or detailed cloud/cluster configurations) used for running experiments.
Software Dependencies No The paper does not list specific software components with their version numbers (e.g., Python 3.x, PyTorch 1.x, CUDA x.x, or specific solver versions).
Experiment Setup No The paper does not provide specific hyperparameter values (e.g., learning rate, batch size, number of epochs) or detailed system-level training settings. It only generally refers to 'details are given in the full version' for experimental results.