Reproduce “$α$-leakage Interpretation of Rényi Capacity” 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

Ni Ding, Farhad Farokhi, Tao Guo, Yinfei Xu, Xiang Zhang

This work unifies various information leakage measures, including Sibson mutual information and a generalized Pointwise Maximal Leakage, under a consistent f-mean framework based on Rényi measures. It reinterprets Rényi capacity as maximal information leakage, providing quantifiable privacy guarantees and suggesting a corresponding computational algorithm.

Key insights

  • Diverse information leakage measures, including Sibson mutual information and a generalized Pointwise Maximal Leakage, can be unified and intuitively understood through a consistent f-mean interpretation based on Rényi measures.
  • Rényi capacity, a fundamental communication theory concept, can be recontextualized as the maximal possible information leakage of a channel, offering a privacy-aware design principle.
  • Quantifiable privacy guarantees for α-leakage can be derived, connecting information-theoretic measures to approximation parameters and providing a concrete basis for privacy-preserving system design.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.