An end-to-end Differentially Private Latent Dirichlet Allocation Using a Spectral Algorithm

Authors: Chris Decarolis, Mukul Ram, Seyed Esmaeili, Yu-Xiang Wang, Furong Huang

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

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct empirical studies with synthetic and real-life datasets, which confirm that the DP spectral algorithm systematically outperforms DP variational inference.
Researcher Affiliation Academia 1Department of Computer Science, University of Maryland 2Department of Computer Science, UC Santa Barbara. Correspondence to: Furong Huang <furongh@cs.umd.edu>.
Pseudocode Yes Procedure 1 ( 1 + 0 1 + , δ1 + δ0 1 + δ)-Differential Privacy (DP) Noise Calibration
Open Source Code No The paper does not provide any explicit statements about releasing source code or links to a code repository.
Open Datasets No The paper mentions using "Wikipedia Dataset" but does not provide a specific link, DOI, repository name, or formal citation with authors/year for accessing it publicly.
Dataset Splits No The paper does not explicitly provide specific train/validation/test dataset splits (e.g., exact percentages, sample counts, or citations to predefined splits).
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes To evaluate Configuration 1, we set the vocabulary size and the number of topics to be small (d = 50, k = 5) in our synthetic settings. The vocabulary size is truncated to be d = 8000. Number of words d = 8000, number of documents N = 50000, 0 = 0.01.