Reproduce “Creation begins with understanding: LLMs as strategy designers for privacy-preserving tabular data synthesis” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Jinmeng Li, Quan Zhang, Hangting Ye, He Zhao, Firas Laakom, Dandan Guo, Jürgen Schmidhuber
TabSSD uses compact tree-derived summaries of variable relationships to have an LLM write executable synthesis procedures, while records remain local. Across six metrics and ten methods, it has the best average rank, 2.17, balancing fidelity, downstream utility, and empirical privacy measures. Its mean runtime is 199.6 seconds versus 332.6 for ARF; this does not establish universal speed or formal privacy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.