Privately generating tabular data using language models

Alexandre Sablayrolles, Yue Wang, Brian Karrer
6/7/2023
cs.LGcs.CLcs.CR

No official code found. Best implementation: facebookresearch/synlm (15 stars, last commit 2023-06-13).

Abstract

Privately generating synthetic data from a table is an important brick of a privacy-first world. We propose and investigate a simple approach of treating each row in a table as a sentence and training a language model with differential privacy. We show this approach obtains competitive results in modelling tabular data across multiple datasets, even at small scales that favor alternative methods based on marginal distributions.

View on arXivView PDF

Code Implementations(1)

facebookresearch/synlmCommunity100%

Code for paper: "Privately generating tabular data using language models".

150PythonJan 27, 20233 years agoNOASSERTION

Similar Papers

Differentially Private Tabular Data Synthesis using Large Language Models

90% match2024

DP-2Stage: Adapting Language Models as Differentially Private Tabular Data Generators

87% match1 repo12024

Deep Generative Models, Synthetic Tabular Data, and Differential Privacy: An Overview and Synthesis

84% match2023

Initial Exploration of Zero-Shot Privacy Utility Tradeoffs in Tabular Data Using GPT-4

84% match2024

Synthesizing Privacy-Preserving Text Data via Finetuning without Finetuning Billion-Scale LLMs

84% match2026

Private prediction for large-scale synthetic text generation

84% match2024

Cite this paper

@article{sablayrolles2023privately,
  title  = {Privately generating tabular data using language models},
  author = {Alexandre Sablayrolles and Yue Wang and Brian Karrer},
  year   = {2023},
  eprint = {2306.04803},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2306.04803v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2306.04803.svg)](https://codeofpaper.com/paper/2306.04803)

Discussion